Remote Sensing and GIS 2 (scrap)

Question 2

Explain the following figure.

Answer 2

Interactive piecewise linear stretch (PLS) uses several different linear functions to stretch different DN ranges of an input image.

PLS is a very versatile point operation function. It can be used to simulate a non-linear function that cannot be easily defined by a mathematical function.

(a) Original image.
(b) The PSL function for contrast enhancement.
(c) Enhanced image.
(d) The PSL function for thresholding.
(e) The binary image produced by thresholding.

Question 3

Given a DEM, how to calculate the slope and aspect of topography using gradient filters?

Answer 3

Question 4

Why the histogram of a Laplacian filtered image is symmetrical to a high peak at zero with both positive and negative values?

Answer 4

Question 5

Why the histogram of a Laplacian filtered image is symmetrical to a high peak at zero with both positive and negative values?

Answer 5

Question 6

Why the histogram of a Laplacian filtered image is symmetrical to a high peak at zero with both positive and negative values?

Answer 6

Question 7

Why the histogram of a Laplacian filtered image is symmetrical to a high peak at zero with both positive and negative values?

Answer 7

Question 8

Why do we have the feature orientated PC selection (FPCS) method?

Answer 8

• We can display and analyse individual PC images or display three PCs as a colour composite.
• As PCs are condensed image information independent of each other, more colourful (i.e. informative) colour composities can be produced from these PC images.
• However, a PC as a combination of the original spectral bands, its relationship to the original spectral signatures of image features corresponding to various ground objects are not apparent.
• To solve this problem, a FPCS method for colour composition was proposed.

Question 9

Describe the feature orientated PC selection (FPCS) method and discuss its application of PC colour composition.

Answer 9

• The technique provides a simple way to select PCs based on the spectral signatures of interested spectral targets (e.g. minerals) so as to enhance the spectral information of these targets by desired colours in the colour composite of the selected PCs.
• The technique involves examination of the eigenvectors to decide the contributions from original bands (either negative or positive) to each PC.
• Specific PCs can then be selected based on the major contributors, which are likely to display the desired targets (spectral features).

Question 10

Describe the combined approach of the FPCS and SPCA for spectral enhancement.

Answer 10

• The outcome of the spectral contrast mapping largely depends on the spectral band groupings.
• Knowing the spectral signatures of intended targets, we can use the FPCS method to decide the grouping of bands and then the selection of PCs for the final RGB display.
• The resultant FPCS spectral contrast mapping colour composite resembles a simple SPCA spectral contrast mapping colour composite, but the signatures of red soils/regoliths, vegetation and clay minerals are more distinctively displayed in red, green and blue.

Question 11

Comment on the combined approach of the FPCS and SPCA for spectral enhancement.

Answer 11

• After all the effort of SPCA, both spectral contrast mapping and FPCS spectral contrast mapping images are less colourful than a simple colour composite of PCs.
• One of the reasons for this is that the selected PCs from the three different bands groups are not independent.
• They may be well correlated even though the PCs within each group are independent from each other.
• The way the image bands are grouped will control the effectiveness of spectral contrast mapping.

Question 12

Discuss the data characteristics of PC images and their applications.

Question 13

In the context of PCA, explain the covarience matrix, Σ_x?

Answer 13

• When X rerpresents an m band MS image, its covariance matrix, Σ_x, is a a full representation of the m dimensional ellipsoid cluster of the image data.
• The elements on the major diagonal of the covariance matrix are the varience of each image bands, while the symmetrical elements off the major diagonal (either side of the major diagonal) are the covarience between two different bands.

Question 15

In the context of PCA, how do eigenvectors and eigenvalues relate to the covarience’ and diagonal covarience matrix, Σ_x and Σ_y?

Define eigenvalue.

Answer 15

• According to the rules of matrix operations we can prove that the transformation G is the n x m transposed matrix of the eigenvectors of Σ_x.
• Σ_y is a diagonal matrix with eigenvalues of Σ_x as non-zero elements along the major diagonal (see image).
• The eigenvalue, λ_i, is the varience of PC_i image and it is proportional to the information contained in PC_i.
• The information content decreases with the increment of the PC rank.

Question 16

Define eigenvalue.

Answer 16

• The eigenvalue, λ_i, is the varience of PC_i image and it is proportional to the information contained in PC_i.
• The information content decreases with the increment of the PC rank.

Question 17

What useful information can you decipher from the table?

Answer 17

• The elements of g1are all positiveand therefore PC1 is a weighted average of all the original image bands.
• PC1 image concentrates features in common for all the six bands. For Earth observation satellite images, this common information is usually topography.
• The elements of g_i (i>1) are usually a mixture of positive and negative values and thus a PC image of higher rank (>1) is a linear combination of positively and negatively weighted images of the original bands.
• The higher rank PCs are lack of topographic features showing more contrast of spectral variation. They all have significantly smaller eigenvalues (PC variances) than the PC1. The eigenvalues decrease rapidly with the increment of the PC rank and thus lower and lower SNR as demonstrated by increasingly noisy appearance of high rank PC images.
• The PC6 image is nearly entirely noise containing little information as indicated by very small variance 1.012. In this sense, PC6 can be disregarded from the dataset and thus the effective dimensionality is reduced to 5 from the original 6 with ignorable information loss of 0.02%.

Question 18

In the context of PCA, explain the diagonal covarience matrix, Σ_y?

Answer 18

• The covarience matrix is a non-negative definite matrix symmetrical along its major diagonal.
• Such a matrix can be converted into a diagonal matrix via basic matrix operations.
• For independent variables in a multi-dimensional space, σ_ij = σ_ji = 0, and thus they have a diagonal covarience matrix.
• In math., the PCA is simply to find a transformation G that diagonalizes the covarience matrix, Σ_x, of the m bands image X to produce an n PC image Y with a diagonal covarience matrix, Σ_y.

Question 19

In the context of PCA, explain the covarience matrix, Σ_x?

Answer 19

• When X rerpresents an m band MS image, its covariance matrix, Σ_x, is a a full representation of the m dimensional ellipsoid cluster of the image data.
• The elements on the major diagonal of the covariance matrix are the varience of each image bands, while the symmetrical elements off the major diagonal (either side of the major diagonal) are the covarience between two different bands.

Question 21

In the context of PCA, how do eigenvectors and eigenvalues relate to the covarience’ and diagonal covarience matrix, Σ_x and Σ_y?

Define eigenvalue.

Answer 21

• According to the rules of matrix operations we can prove that the transformation G is the n x m transposed matrix of the eigenvectors of Σ_x.
• Σ_y is a diagonal matrix with eigenvalues of Σ_x as non-zero elements along the major diagonal (see image).
• The eigenvalue, λ_i, is the varience of PC_i image and it is proportional to the information contained in PC_i.
• The information content decreases with the increment of the PC rank.

Question 22

Define eigenvalue.

Answer 22

• The eigenvalue, λ_i, is the varience of PC_i image and it is proportional to the information contained in PC_i.
• The information content decreases with the increment of the PC rank.

Question 23

What useful information can you decipher from the table?

Answer 23

• The elements of g1are all positiveand therefore PC1 is a weighted average of all the original image bands.
• PC1 image concentrates features in common for all the six bands. For Earth observation satellite images, this common information is usually topography.
• The elements of g_i (i>1) are usually a mixture of positive and negative values and thus a PC image of higher rank (>1) is a linear combination of positively and negatively weighted images of the original bands.
• The higher rank PCs are lack of topographic features showing more contrast of spectral variation. They all have significantly smaller eigenvalues (PC variances) than the PC1. The eigenvalues decrease rapidly with the increment of the PC rank and thus lower and lower SNR as demonstrated by increasingly noisy appearance of high rank PC images.
• The PC6 image is nearly entirely noise containing little information as indicated by very small variance 1.012. In this sense, PC6 can be disregarded from the dataset and thus the effective dimensionality is reduced to 5 from the original 6 with ignorable information loss of 0.02%.

Question 24

In the context of PCA, explain the diagonal covarience matrix, Σ_y?

Answer 24

• The covarience matrix is a non-negative definite matrix symmetrical along its major diagonal.
• Such a matrix can be converted into a diagonal matrix via basic matrix operations.
• For independent variables in a multi-dimensional space, σ_ij = σ_ji = 0, and thus they have a diagonal covarience matrix.
• In math., the PCA is simply to find a transformation G that diagonalizes the covarience matrix, Σ_x, of the m bands image X to produce an n PC image Y with a diagonal covarience matrix, Σ_y.

Question 25

How is histogram equalization (HE) acheived?

Answer 25

• By transforming an input image to an output image with a uniform (equalised) histogram.
  *

Question 26

What is histogram matching?

Answer 26

• Histogram matching is a point operation that transforms an input image to make its histogram match a given shape defined by either a math. function or a histogram of another image.
• It is particularly useful for image comparison and differencing (If two images being compared are modifed to have similar histograms, the comparison will be fairer.

Question 27

How is histogram equalization (HE) used to acheive histogram matching (HM)?

Answer 27

• HM can be implemented by applying HE twice.
• An equalized histogram is only decided by image size, N, and the output DN range, L.
• Images of the same size always have the same equalized histogram for a fixed output DN range and thus HE can act as a bridge to link images of the same size but different histograms.

Question 29

What is a digital image?

Answer 29

• A digital image is a two dimensional (2D) array of numbers in lines and columns or a raster dataset.
• It can also have a 3rd dimension layers (e.g. image bands).
• Each cell of a digital image is called a pixel and the number representing the brightness of the pixel is called a digital number (DN).
• A digital image is nota matrix!

Question 30

Explain the relationship between primary colours and complimentary primary colours with a diagram.

Answer 30

• A light of non-primary colour (C) stimulates (combines) different portion of each group (primary colour) to form the perception of this colour.
• The mixture of the lights of the three primary colours can t.f. produce any colours.
• This is the principle of RGB additive colour composition.

C = rR + gG + bB

• Mixtures of equal amount of 3 primary colours (r=g=b) are white or grey.
• Equal amount of any 2 primary colours generates a complementary colour: yellow, cyan and magenta.
• These three complementary colours can also be used as primaries to generate various colours as does in colour printing.

Question 31

Using a diagram to illustrate colour cube. Give the definition of the grey line in the colour cube.

Answer 31

The line from the origin of the colour cube to the opposite convex corner is known as the grey linebecause pixel vectors that lie on this line have equal components in red, green and blue (i.e. r=g=b).

Question 32

Using a diagram to illustrate colour cube. How is a colour composed of RGB components?

Answer 32

• Consider the components of an RGB display as the orthogonal axes of a 3D colour space; the maximum possible DN level in each component of the display defines the RGB colour cube.
• Any an image pixel in this system is represented by a vector from the origin to somewhere within the colour cube.
• Most standard RGB display system can display 8 bits/pixel/channel, up to 24 bits (~16.8 million) different colours.

Question 33

What is clipping and why is it often essential for image display?

Answer 33

• In digital images, a few pixels may occupy wide value range at the low and high ends of histograms (often represent noise).
• In this case, setting a proper cut-off to clip the both ends of the histogram is necessary in contrast enhancement to make effective usege of the dynamic range of a display device.
• Clipping is often given as a % of total number of pixels in an image e.g. set 1% and 99% as the cut-off limits at the low and high ends of the histogram.
• The image is then stretched as to set the DN levels where H_i(x_l)=1% to 0, and DN levels where H_i(x_h)=99% to 255 (for 8 bit display) in the output image.

Question 34

How is histogram equalization (HE) acheived?

Answer 34

• By transforming an input image to an output image with a uniform (equalised) histogram.
  *

Question 35

What is histogram matching?

Answer 35

• Histogram matching is a point operation that transforms an input image to make its histogram match a given shape defined by either a math. function or a histogram of another image.
• It is particularly useful for image comparison and differencing (If two images being compared are modifed to have similar histograms, the comparison will be fairer.

Question 36

How is histogram equalization (HE) used to acheive histogram matching (HM)?

Answer 36

• HM can be implemented by applying HE twice.
• An equalized histogram is only decided by image size, N, and the output DN range, L.
• Images of the same size always have the same equalized histogram for a fixed output DN range and thus HE can act as a bridge to link images of the same size but different histograms.

Question 37

What is histogram matching?

Answer 37

• Histogram matching is a point operation that transforms an input image to make its histogram match a given shape defined by either a math. function or a histogram of another image.
• It is particularly useful for image comparison and differencing (If two images being compared are modifed to have similar histograms, the comparison will be fairer.

Question 38

What is a digital image?

Answer 38

• A digital image is a two dimensional (2D) array of numbers in lines and columns or a raster dataset.
• It can also have a 3rd dimension layers (e.g. image bands).
• Each cell of a digital image is called a pixel and the number representing the brightness of the pixel is called a digital number (DN).
• A digital image is nota matrix!

Question 39

Explain the relationship between primary colours and complimentary primary colours with a diagram.

Answer 39

• A light of non-primary colour (C) stimulates (combines) different portion of each group (primary colour) to form the perception of this colour.
• The mixture of the lights of the three primary colours can t.f. produce any colours.
• This is the principle of RGB additive colour composition.

C = rR + gG + bB

• Mixtures of equal amount of 3 primary colours (r=g=b) are white or grey.
• Equal amount of any 2 primary colours generates a complementary colour: yellow, cyan and magenta.
• These three complementary colours can also be used as primaries to generate various colours as does in colour printing.

Question 40

Using a diagram to illustrate colour cube. Give the definition of the grey line in the colour cube.

Answer 40

The line from the origin of the colour cube to the opposite convex corner is known as the grey linebecause pixel vectors that lie on this line have equal components in red, green and blue (i.e. r=g=b).

Question 41

Using a diagram to illustrate colour cube. How is a colour composed of RGB components?

Answer 41

• Consider the components of an RGB display as the orthogonal axes of a 3D colour space; the maximum possible DN level in each component of the display defines the RGB colour cube.
• Any an image pixel in this system is represented by a vector from the origin to somewhere within the colour cube.
• Most standard RGB display system can display 8 bits/pixel/channel, up to 24 bits (~16.8 million) different colours.

Question 42

What is clipping and why is it often essential for image display?

Answer 42

• In digital images, a few pixels may occupy wide value range at the low and high ends of histograms (often represent noise).
• In this case, setting a proper cut-off to clip the both ends of the histogram is necessary in contrast enhancement to make effective usege of the dynamic range of a display device.
• Clipping is often given as a % of total number of pixels in an image e.g. set 1% and 99% as the cut-off limits at the low and high ends of the histogram.
• The image is then stretched as to set the DN levels where H_i(x_l)=1% to 0, and DN levels where H_i(x_h)=99% to 255 (for 8 bit display) in the output image.

Question 43

How is histogram equalization (HE) used to acheive histogram matching (HM)?

Answer 43

• HM can be implemented by applying HE twice.
• An equalized histogram is only decided by image size, N, and the output DN range, L.
• Images of the same size always have the same equalized histogram for a fixed output DN range and thus HE can act as a bridge to link images of the same size but different histograms.

Question 44

How is histogram equalization (HE) acheived?

Answer 44

• By transforming an input image to an output image with a uniform (equalised) histogram.
  *

Question 45

What is a digital image?

Answer 45

• A digital image is a two dimensional (2D) array of numbers in lines and columns or a raster dataset.
• It can also have a 3rd dimension layers (e.g. image bands).
• Each cell of a digital image is called a pixel and the number representing the brightness of the pixel is called a digital number (DN).
• A digital image is nota matrix!

Question 46

Explain the relationship between primary colours and complimentary primary colours with a diagram.

Answer 46

• A light of non-primary colour (C) stimulates (combines) different portion of each group (primary colour) to form the perception of this colour.
• The mixture of the lights of the three primary colours can t.f. produce any colours.
• This is the principle of RGB additive colour composition.

C = rR + gG + bB

• Mixtures of equal amount of 3 primary colours (r=g=b) are white or grey.
• Equal amount of any 2 primary colours generates a complementary colour: yellow, cyan and magenta.
• These three complementary colours can also be used as primaries to generate various colours as does in colour printing.

Question 47

Using a diagram to illustrate colour cube. Give the definition of the grey line in the colour cube.

Answer 47

The line from the origin of the colour cube to the opposite convex corner is known as the grey linebecause pixel vectors that lie on this line have equal components in red, green and blue (i.e. r=g=b).

Question 48

Using a diagram to illustrate colour cube. How is a colour composed of RGB components?

Answer 48

• Consider the components of an RGB display as the orthogonal axes of a 3D colour space; the maximum possible DN level in each component of the display defines the RGB colour cube.
• Any an image pixel in this system is represented by a vector from the origin to somewhere within the colour cube.
• Most standard RGB display system can display 8 bits/pixel/channel, up to 24 bits (~16.8 million) different colours.

Question 49

What is clipping and why is it often essential for image display?

Answer 49

• In digital images, a few pixels may occupy wide value range at the low and high ends of histograms (often represent noise).
• In this case, setting a proper cut-off to clip the both ends of the histogram is necessary in contrast enhancement to make effective usege of the dynamic range of a display device.
• Clipping is often given as a % of total number of pixels in an image e.g. set 1% and 99% as the cut-off limits at the low and high ends of the histogram.
• The image is then stretched as to set the DN levels where H_i(x_l)=1% to 0, and DN levels where H_i(x_h)=99% to 255 (for 8 bit display) in the output image.

Question 50

Explain the principle of using colours as a tool to visualize spectral information of (a) multi-spectral image(s).

Answer 50

• Though colours are light of the visible spectral range 380-750nm, they are used as a tool for information visualization in colour display of digital images.
• Thus, for digital image display, the assignment of each primary colour for a spectral band or layer can be arbitrarily depending on the requirements of applications, which is not necessarily the colour corresponding to the spectral range of the band.
• If we display three image bands in red, green and blue spectral ranges in RGB, then a true colour compositeimage is generated.

Question 51

Use a diagram to illustrate the 4f optical image filtering system and explain the principle of image filtering based on Fourier Transform.

Answer 51

• Fourier Transform (FT) to transfer an image into frequency domain.
• Remove or alter the data of particular frequencies by a filter.
• Inverse Fourier Transform (IFT) to transfer the filtered frequency spectrum back to the spatial domain to produce a filtered image.

Question 52

Give e.g.’s of constraint criterion for MCE

Question 53

Give e.g.’s of factor criterion for MCE

Question 54

Describe how you would assess relative significance between factors in a multi-criteria spatial analysis problem, such as suitability for a site selection.

Answer 54

• All criteria in the multi-criteria evaluation do not necessarily have equal significance on the outcome.
• So we use weights (should always sum to 1) to assess and quantify the relative significance of criteria.
• There are several methods of calculating weights, commonly:
  
  • Rating
  • Ranking (most common)
  • Pairwise Comparison Matrix (involves subjective decision about significance and weight calculation method)

Question 55

Mention how you would derive and apply weights within MCE.

Answer 55

• PCM Factor weight derivation:
• The factor weights are produced from principal Eigenvectors of the Pairwise comparison matrix.
• All the Factor Weights generated from this matrix sum to 1.0 and these are then multiplied by their respective factor images.

Question 56

Mention how you would combine scaled and weighted factors within MCE.

Answer 56

1. Boolean combination
2. Index Overylay
3. Arithmetic combination
4. Analytical Hierarchy Approach (AHP) and Weighted factors in Linear Combination (WLC)
5. Vectorial Fuzzy Modelling (VFM)
6. Ordered Weighted Average (OWA)
7. Weights of evidence modeling
8. Dempster-Shafer Theory

Question 57

There are several methods for factor combination within MCE.

Describe the boolean combination method.

Answer 57

• Simplest method
• Produce a series of factor maps where..
• A a location, every factor has two possible states (suitable or unsuitable, 1 or 0)
• Spatial relationship factors combined into a single map/index.
• e.g. mineral prospectivity mapping

Question 58

During boolean combination, spatial relationship factors are combined into a single map/index.

How?

Answer 58

• AND combinatorial operator
  
  • Retains only those areas suitable in all factors.
  • ‘Risk averse’ or conservative
• OR (conceptual opposite of AND)
  
  • Retains area where any suitable factor exists.
  • More or larger areas categorized as suitable.
  • ‘Risk taking’ or liberal

Question 60

There are several methods for factor combination within MCE.

What are the pros of Boolean combination?

Answer 60

• Simple, fast and intuitive - useful if knowledge & data lacking
• Use of AND & OR provide some variability - risk averse or risk taking outcomes can be very different - represent extreme opposite results

Question 61

There are several methods for factor combination within MCE.

What are the cons of Boolean combination?

Answer 61

• Too simplistic - in both description of the criterion information and in criteria combination
• Boolean anything represents ‘Hard’ decisions using sharp boundaries, discrete classes and crisp thresholds
• No allowance for uncertainty - simple, rigid classes of criterion memberiship and result, i.e. suitable or not suitable
• Unbounded and qualitative
  
  • No measure of quantity in the result - result is favorouble or not
  • No measure of confidence or limitation

Question 62

There are several methods for factor combination within MCE.

Describe Index Overlay

Answer 62

• Each spatial relationship factor has two or more discrete levels of suitability, represented as ordinal scale numbers.
  
  • A location with a value of 2 is more suitable than a location with a value of 1 but it is not necessarily twice as suitable.
• The resultant suitability map is constructed by summation, arithmetic mean or geometric mean
  
  • The higher the output number, the more suitable the location.
• e.g. lithological instability map

Question 63

There are several methods for factor combination within MCE.

What are the drawbacks of Index Overlay?

Answer 63

• More levels of suitability used gives a layer greater influence
• More input factors means a greater range in output suitability values
• Input factors are not scaled or the range/number of input criteria values must be controlled
• Final result is again unbounded and qualitative

Question 64

There are several methods for factor combination within MCE.

Describe Arithmetic Combination

Answer 64

• Modification of Index Overlay
• Ordinal scale replaced by ratio scale (real numbers) - factor value of 2 is now twice as suitable as 1 - removes need for input scaling.
• Can be useful in minex prospectivity mapping where particular features statistically related to size of deposit
  
  • Different metrics from input factors
  • e.g. mean number of discoveries per sq km, or mean weight of gold discovered per sq km
• Produces more conservative result (risk averse) than the Index Overlay method

Question 65

There are several methods for factor combination within MCE.

Describe Analytical Hierarchy Process (AHP)

Answer 65

• Structured technique for organising and structuring complex decisions, based on mathematics and psychology.
• Rather than prescribing a “correct” decision, the AHP helps decision makers find one that best suits their goal and their understanding of the problem.
• Makes use of the Pairwise Comparison matrix technque.
• A series of criteria are produced, assessed, scaled, weighted and combined by summation.

Question 66

There are several methods for factor combination within MCE.

Describe Weighted factors in Linear Combination (WLC)

Answer 66

• An example of Analytical Hierarchy Process (AHP)
• WLC involves evaluation, relative weighting and combination of selected criteria
• Allows consideration of both qualitative and quantitative aspects of inputs and decisions
  
  • Reduces complex decisions to a series of one-to-one comparisons, then synthesizes the results.
• Accepts that certain criteria are more important than others, and that criteria have intermediate values of suitability (i.e. they are not simply classed ‘suitable’ or ‘unsuitable’).
• Factors (continuous variables) and constraints (Boolean) can be used

Question 67

What are two important issues re: Weighted factors in Linear Combination (WLC)?

Answer 67

1. How is the relative importance of each criterion determined (to derive the weights)?
   
   • Assessed using a Pairwise comparison matrix
2. How should each factor be standardised? (All must contribute positively toward the outcome and be scaled)
   
   • Scaled using Fuzzy membership functions

Question 68

Weighted factors in Linear Combination (WLC) also allows a lack of suitability in one factor to be compensated for by higher suitability in another factor(s).

• Via the full and equal ‘trade-off’ between factors
• Weighted summation - any zeros encountered are not removed

Describe this procedure.

Answer 68

1. Identify the criteria (decide which criteria are factors and which are constraints)
2. Produce an image or coverage of each
3. Standardise each one, using a function, to a Real number scale (0 to 1) or byte scale (0 to 255) or % scale (0-100)
4. Derive weighting coefficients which convey the relative importance of each factor (using Pairwise Comparison Matrix)
5. Linearly combine the factor weights with the standardised factors and the constraints to produce the Suitability Map.

Question 69

There are several methods for factor combination within MCE.

Describe the Vectoral Fuzzy Modeling method.

Answer 69

• To improve on the conventional fuzzy logic issues, a novel, vectoral fuzzy technique was developed producing 2 values for each factor
  
  • Calculated prospectivity
  • Confidence (the similarity between input factors)
• Combination of the two values involves calculating of a vector for each spatial relationship factor
  
  • The combined lengths (lc) and directions of each vector provide the aggregate suitability.
  • The more similar the input values, the longer the resultant vector and the larger the value of lc (the higher the suitability)

Question 70

Re: Vectoral Fuzzy Modeling

What’s shown in image b?

Answer 70

• Two vectors of equal prospectivity (i.e. direction) but different confidence (length)
• The lower vector has the larger confidence (i.e. longer vector) and thus has a greater influence on the output when combined using the vectorial fuzzy logic method.
• When combining two fuzzy vectors using vectorial fuzzy logic, conflicting suitability amongst the input factors reduces the confidence

Question 71

Briefly discuss the relative pros and cons of weighted linear factors in combination versus Boolean method.

Answer 71

• WLC is a very commonly used, intuitive method but maybe too liberal bc of the full trade-off, which may not be desirable
• WLC method allows every location to pass through to the end
• Often a good first attempt and useful where data are plentiful but decision rules are not well understood and need to be handled carefully.
• Control over the amount of trade-off would be advantageous.
• In contrast, a simple Boolean method (or geometric mean) is very harsh in this respect but it ensures that known unfavourable conditions can be removed from the analysis where appropriate.
• Useful if decision rules are well understood but data are incomplete or coarse.
• The relationships between factors are not necc. linear
• It is common for a combination of these methods to be adopted

Question 74

Give an example of an application of aggregation of fuzzy scaled inputs and state the associated constraints and factors

Answer 74

Suitable areas for urban development

• Constraints
  
  • Water areas
  • Underdeveloped land vs developed land
• Factors
  
  • Distance from protected watersheds
  • Distance from protected parks
  • Distance from open water bodies
  • Slope of the land
  • Elevation of the land

Question 75

Name some applications of hyperspectral imaging

Answer 75

1. Agriculture (precision farming)
2. Food processing - impurity detection, quality control
3. Medical - tissue physiology, disease detection
4. Surveillance & security - face recognition, airport security
5. Minerals - O&G, and Min exploration (alteration minerals & REEs)
6. Civil engineering - mapping small-scale chemical changes in materials
7. Physics - material behaviour
8. Astronomy - planets, comets, asteroids, meteorite trails
9. Environment & hazards - pollution monitoring etc
10. Archaeology - non-invsive chemical mapping of paints and pigments

Question 76

What are the physical causes of the various types of spectral behaviour?

Answer 76

Scattering effects

• Diffuse and/or specular reflection
• Volume and/or surface scattering
• Single and/or multiple scattering
• Wavelength, particle & surface - dependent

Question 77

What are the chemical causes of the various types of spectral behaviour?

Answer 77

• Refractive index (ratio of the speed of light through one material to that through another)
  
  • Reflectance, transmission
  • Wavelength-dependent
• ** Absorption coefficient ** (creates most significant signature)
  
  • Chemical composition & crystallography
  • Electronic and/or vibrational processes
  • Wavelength-dependent

Question 78

What processes cause absorption (re: spectral analysis)?

Answer 78

• Electronic transitions (Vis & NIR)
  
  • Crystal field effects
  • Charge transfer
• Vibrational transitions (SWIR & TIR)
  
  • Bond stretch, bending & rotation

In all cases, absorbed energy is later emitted at longer wavelength than it is absorbed

Question 79

Re: spectral behaviour; explain the causes of electronic transitions (ET)

Answer 79

• Isolated atoms and ions have discrete energy states.
• Absorption of photons of a specific wavelength causes a change from one energy state to a higher one (excitation).
• On relaxation, emission of a photon occurs, with a drop in energy state to a lower one.
• When a photon is absorbed it is usually not emitted at the same wavelength (e.g. causes heating, on relaxation T drops)
• Atoms can rotate and vibrate with respect to each other
• Electrons can migrate from one energy level to another
• All require high energy (effects confined to VNIR)

Question 80

What are the diagnostic effects of Electronic Transitions upon absorption?

Answer 80

Ordinarily there are no truly diagnostic effects (broad) but there are several subtle effects which are useful.

Mainly, Crystal Field (CF) effects and
Charge Transfer (CT) effects

Question 81

Explain the cause of Crystal Field (CF) effects.

Answer 81

• The most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).
• Absorptions are affected by a crystal’s symmetry and degree of lattice disortion.
• CF effects produce subtle changes to the spectral signatures
• Excited states affected by the electrostatic field around the atom.
• The crystal field varies with crystal structure from mineral to mineral (the amount of splitting varies for different molecular geometries) - diagnostic within mineral species.
• Different spectra can be produced by the same ion depending on oxidation state.
• In complete spectra (i.e. in lab. conditions) mineral identification is possible.
• Causes different colours in transition metal complexes (the compound will appear as the complimentary colour from the wavelength absorped).

Question 83

CF is the most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).

Why the transition metals?

Answer 83

• Outer electron shells partially filled
• d orbital energy states are split when an atom lies in a crystal field - the splitting allows an electron to move to a higher level (Crystal Field Theory)

Question 84

Explain the cause of Charge Transfer (CT) absorptions.

Answer 84

• The absorption of a photon causes an electron to move between ions or between ligands.
• For any substance, one of its components must have electron donating properties and another component must be able to accept electrons.
• Absorption of radiation then causes the transfer (jump) of an electron from the donor (ligand orbital) to an orbital associated with the acceptor (metal orbital). Usually when metal is in oxidation (oxidised?) state.

Question 85

What are the benefits of Charge Transfer (CT) absorptions?

Answer 85

• Usually diagnostic at mineralogical level
• Strengths of these effects are 100-1000s of times stronger than CF effects
• Dominantly in the UV but also extend into the VNIR

Question 86

What are the electronic effects on spectra of rock-forming minerals and rocks?

Answer 86

• CF and CT effects observed in spectra of many minerals and rocks
• CT effects not particularly helpful for species-level identification - too broad
• CF effects cause absorptions at specific wavelengths and so are very useful for identification
• All minerals containing transition metals will display both effects in their spectra

Question 87

Explain the causes of Vibrational Transitions (VT) spectral behaviour.

Answer 87

• In general, for a molecule with N atoms, there are 3N-6 normal vibration modes called fundamentals.
• Each vibration can occur at multiples of the original fundamental frequency.
• Additional vibrations are called overtones (involving multiples of a single fundamental mode) and combinations (involving different modes).
• Fundamental vibrations are caused by small displacements of atoms within molecules from their state of equilibrium.
• Most important ones associated with OH- ions, water molecules or fluid inclusions, but
• Many silicates and alteration minerals (involving S-O bond) show features too.

Question 88

How can Charge Transfer (CT) absorptions cause the same mineral-type to be in different colours?

Answer 88

• Can occur between the same metal in different oxidation states, such as between Fe²⁺ and Fe³⁺,
• i.e. where ions of more than one form of element exist together, charge transfers occur between them - producing minerals of different colours
• CT absorptions are the main cause of red colour in iron oxides and hydroxides

Question 91

Vibrational Transitions (VT) take place where?

Answer 91

Most VT occur in the SWIR (some in TIR) and ‘IR vibrationally active’ minerals have dipole moments

Question 92

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 92

• Absorption minima shift to longer wavelengths within the iron oxide family, enabling species-level identification in lab. spectra.
• Reflectance minima at diagnostic wavelengths
• But in Landsat image spectra we can only detect the general presence of iron-oxides
• All iron oxides reflect strongly in band 3 and absorb in band 1
• Iron-oxides can be most effectively identified using a 3/1 ratio (Fe-oxide rich gossans are often associated with mineral deposits so this is an exploration tool)

Question 93

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 93

• Band widths are much narrower
• Extra coastal (new b1) added
• Allows better discrimination (potentially)
• (b1+b2+b3) / (b4 + b5) - haematite
• b3/b2 or b4/b2 - goethite & jarosite
• b4/b1 - jarosite (possibly)

Question 94

Describe how the reflective spectral properties of Iron oxides/hydroxides in Sentinel-2 can be used for interpretation of images and spectra.

Answer 94

• Similar to Landsat 8
• More VNIR bands and narrower bandwidths
• Allows potentially much better discrimination (and possibly identification)

Question 95

Describe how the reflective spectral properties of Hydrated minerals in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 95

• Hydrated (clay) mineral absorption features in the SWIR
• Shape, position and symmetry of absorption features enables species-level identification of clay minerals
  
  • Al-OH bond bending produces distinctive features near 2.25 um.
  • Absorption caused by HOH bonds at 1.4 & 1.9 um.
• But in Landsat TM we can only detect the general presence of clay minerals - b7 straddles the diagnostic absorptions 2.1 to 2.35 um.
• In general, clay minerals reflect strongly in mid-SWIR and absorb in far-SWIR (not muscovite)
• e.g. TM 5/7 ratio (reveals hydrothermal alteration minerals and t.f. is an exploration tool)

Question 96

Describe how the reflective spectral properties of Hydrated minerals in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 96

• There are still only 2 SWIR bands…
• Band widths narrower than L5 & 7
• Allows discrimination of white mica from kaolinite & smectites (potentially)
• b6/b7 - general clays, white mica (?)
• b6/(b5+b7) - kaolinite
• Not easy and NB vegetation

Question 97

What are the common rock forming minerals that have diagnostic spectral features?

Answer 97

• Quartz (silica & silicates), iron, carbonates
• Substances containing O, H, C, AL and other transition metals
• Mixtures of these account for most earth materials

Question 98

What substances have reflective spectral features in the Visible (VIS) spectral region?

Answer 98

• Ferrous iron, Fe-AL silicates
• e.g. pigeonite, olivine, staurolite

Question 99

What are the chemical causes of the various types of spectral behaviour?

Answer 99

• Refractive index (ratio of the speed of light through one material to that through another)
  
  • Reflectance, transmission
  • Wavelength-dependent
• ** Absorption coefficient ** (creates most significant signature)
  
  • Chemical composition & crystallography
  • Electronic and/or vibrational processes
  • Wavelength-dependent

Question 100

What processes cause absorption (re: spectral analysis)?

Answer 100

• Electronic transitions (Vis & NIR)
  
  • Crystal field effects
  • Charge transfer
• Vibrational transitions (SWIR & TIR)
  
  • Bond stretch, bending & rotation

In all cases, absorbed energy is later emitted at longer wavelength than it is absorbed

Question 101

Re: spectral behaviour; explain the causes of electronic transitions (ET)

Answer 101

• Isolated atoms and ions have discrete energy states.
• Absorption of photons of a specific wavelength causes a change from one energy state to a higher one (excitation).
• On relaxation, emission of a photon occurs, with a drop in energy state to a lower one.
• When a photon is absorbed it is usually not emitted at the same wavelength (e.g. causes heating, on relaxation T drops)
• Atoms can rotate and vibrate with respect to each other
• Electrons can migrate from one energy level to another
• All require high energy (effects confined to VNIR)

Question 102

What are the diagnostic effects of Electronic Transitions upon absorption?

Answer 102

Ordinarily there are no truly diagnostic effects (broad) but there are several subtle effects which are useful.

Mainly, Crystal Field (CF) effects and
Charge Transfer (CT) effects

Question 103

Explain the cause of Crystal Field (CF) effects.

Answer 103

• The most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).
• Absorptions are affected by a crystal’s symmetry and degree of lattice disortion.
• CF effects produce subtle changes to the spectral signatures
• Excited states affected by the electrostatic field around the atom.
• The crystal field varies with crystal structure from mineral to mineral (the amount of splitting varies for different molecular geometries) - diagnostic within mineral species.
• Different spectra can be produced by the same ion depending on oxidation state.
• In complete spectra (i.e. in lab. conditions) mineral identification is possible.
• Causes different colours in transition metal complexes (the compound will appear as the complimentary colour from the wavelength absorped).

Question 104

Explain the cause of Crystal Field (CF) effects.

Answer 104

• The most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).
• Absorptions are affected by a crystal’s symmetry and degree of lattice disortion.
• CF effects produce subtle changes to the spectral signatures
• Excited states affected by the electrostatic field around the atom.
• The crystal field varies with crystal structure from mineral to mineral (the amount of splitting varies for different molecular geometries) - diagnostic within mineral species.
• Different spectra can be produced by the same ion depending on oxidation state.
• In complete spectra (i.e. in lab. conditions) mineral identification is possible.
• Causes different colours in transition metal complexes (the compound will appear as the complimentary colour from the wavelength absorped).

Question 105

Explain the cause of Crystal Field (CF) effects.

Answer 105

• The most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).
• Absorptions are affected by a crystal’s symmetry and degree of lattice disortion.
• CF effects produce subtle changes to the spectral signatures
• Excited states affected by the electrostatic field around the atom.
• The crystal field varies with crystal structure from mineral to mineral (the amount of splitting varies for different molecular geometries) - diagnostic within mineral species.
• Different spectra can be produced by the same ion depending on oxidation state.
• In complete spectra (i.e. in lab. conditions) mineral identification is possible.
• Causes different colours in transition metal complexes (the compound will appear as the complimentary colour from the wavelength absorped).

Question 106

What are the physical causes of the various types of spectral behaviour?

Answer 106

Scattering effects

• Diffuse and/or specular reflection
• Volume and/or surface scattering
• Single and/or multiple scattering
• Wavelength, particle & surface - dependent

Question 107

CF is the most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).

Why the transition metals?

Answer 107

• Outer electron shells partially filled
• d orbital energy states are split when an atom lies in a crystal field - the splitting allows an electron to move to a higher level (Crystal Field Theory)

Question 108

Explain the cause of Charge Transfer (CT) absorptions.

Answer 108

• The absorption of a photon causes an electron to move between ions or between ligands.
• For any substance, one of its components must have electron donating properties and another component must be able to accept electrons.
• Absorption of radiation then causes the transfer (jump) of an electron from the donor (ligand orbital) to an orbital associated with the acceptor (metal orbital). Usually when metal is in oxidation (oxidised?) state.

Question 109

What are the electronic effects on spectra of rock-forming minerals and rocks?

Answer 109

• CF and CT effects observed in spectra of many minerals and rocks
• CT effects not particularly helpful for species-level identification - too broad
• CF effects cause absorptions at specific wavelengths and so are very useful for identification
• All minerals containing transition metals will display both effects in their spectra

Question 110

Explain the causes of Vibrational Transitions (VT) spectral behaviour.

Answer 110

• In general, for a molecule with N atoms, there are 3N-6 normal vibration modes called fundamentals.
• Each vibration can occur at multiples of the original fundamental frequency.
• Additional vibrations are called overtones (involving multiples of a single fundamental mode) and combinations (involving different modes).
• Fundamental vibrations are caused by small displacements of atoms within molecules from their state of equilibrium.
• Most important ones associated with OH- ions, water molecules or fluid inclusions, but
• Many silicates and alteration minerals (involving S-O bond) show features too.

Question 111

What are the electronic effects on spectra of rock-forming minerals and rocks?

Answer 111

• CF and CT effects observed in spectra of many minerals and rocks
• CT effects not particularly helpful for species-level identification - too broad
• CF effects cause absorptions at specific wavelengths and so are very useful for identification
• All minerals containing transition metals will display both effects in their spectra

Question 112

Explain the causes of Vibrational Transitions (VT) spectral behaviour.

Answer 112

• In general, for a molecule with N atoms, there are 3N-6 normal vibration modes called fundamentals.
• Each vibration can occur at multiples of the original fundamental frequency.
• Additional vibrations are called overtones (involving multiples of a single fundamental mode) and combinations (involving different modes).
• Fundamental vibrations are caused by small displacements of atoms within molecules from their state of equilibrium.
• Most important ones associated with OH- ions, water molecules or fluid inclusions, but
• Many silicates and alteration minerals (involving S-O bond) show features too.

Question 113

What are the benefits of Charge Transfer (CT) absorptions?

Answer 113

• Usually diagnostic at mineralogical level
• Strengths of these effects are 100-1000s of times stronger than CF effects
• Dominantly in the UV but also extend into the VNIR

Question 114

Explain the causes of Vibrational Transitions (VT) spectral behaviour.

Answer 114

• In general, for a molecule with N atoms, there are 3N-6 normal vibration modes called fundamentals.
• Each vibration can occur at multiples of the original fundamental frequency.
• Additional vibrations are called overtones (involving multiples of a single fundamental mode) and combinations (involving different modes).
• Fundamental vibrations are caused by small displacements of atoms within molecules from their state of equilibrium.
• Most important ones associated with OH- ions, water molecules or fluid inclusions, but
• Many silicates and alteration minerals (involving S-O bond) show features too.

Question 115

What are the electronic effects on spectra of rock-forming minerals and rocks?

Answer 115

• CF and CT effects observed in spectra of many minerals and rocks
• CT effects not particularly helpful for species-level identification - too broad
• CF effects cause absorptions at specific wavelengths and so are very useful for identification
• All minerals containing transition metals will display both effects in their spectra

Question 116

How can Charge Transfer (CT) absorptions cause the same mineral-type to be in different colours?

Answer 116

• Can occur between the same metal in different oxidation states, such as between Fe²⁺ and Fe³⁺,
• i.e. where ions of more than one form of element exist together, charge transfers occur between them - producing minerals of different colours
• CT absorptions are the main cause of red colour in iron oxides and hydroxides

Question 117

How can Charge Transfer (CT) absorptions cause the same mineral-type to be in different colours?

Answer 117

• Can occur between the same metal in different oxidation states, such as between Fe²⁺ and Fe³⁺,
• i.e. where ions of more than one form of element exist together, charge transfers occur between them - producing minerals of different colours
• CT absorptions are the main cause of red colour in iron oxides and hydroxides

Question 118

What are the benefits of Charge Transfer (CT) absorptions?

Answer 118

• Usually diagnostic at mineralogical level
• Strengths of these effects are 100-1000s of times stronger than CF effects
• Dominantly in the UV but also extend into the VNIR

Question 119

Vibrational Transitions (VT) take place where?

Answer 119

Most VT occur in the SWIR (some in TIR) and ‘IR vibrationally active’ minerals have dipole moments

Question 120

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 120

• Absorption minima shift to longer wavelengths within the iron oxide family, enabling species-level identification in lab. spectra.
• Reflectance minima at diagnostic wavelengths
• But in Landsat image spectra we can only detect the general presence of iron-oxides
• All iron oxides reflect strongly in band 3 and absorb in band 1
• Iron-oxides can be most effectively identified using a 3/1 ratio (Fe-oxide rich gossans are often associated with mineral deposits so this is an exploration tool)

Question 121

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 121

• Band widths are much narrower
• Extra coastal (new b1) added
• Allows better discrimination (potentially)
• (b1+b2+b3) / (b4 + b5) - haematite
• b3/b2 or b4/b2 - goethite & jarosite
• b4/b1 - jarosite (possibly)

Question 122

Describe how the reflective spectral properties of Iron oxides/hydroxides in Sentinel-2 can be used for interpretation of images and spectra.

Answer 122

• Similar to Landsat 8
• More VNIR bands and narrower bandwidths
• Allows potentially much better discrimination (and possibly identification)

Question 123

Describe how the reflective spectral properties of Hydrated minerals in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 123

• Hydrated (clay) mineral absorption features in the SWIR
• Shape, position and symmetry of absorption features enables species-level identification of clay minerals
  
  • Al-OH bond bending produces distinctive features near 2.25 um.
  • Absorption caused by HOH bonds at 1.4 & 1.9 um.
• But in Landsat TM we can only detect the general presence of clay minerals - b7 straddles the diagnostic absorptions 2.1 to 2.35 um.
• In general, clay minerals reflect strongly in mid-SWIR and absorb in far-SWIR (not muscovite)
• e.g. TM 5/7 ratio (reveals hydrothermal alteration minerals and t.f. is an exploration tool)

Question 124

Describe how the reflective spectral properties of Hydrated minerals in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 124

• There are still only 2 SWIR bands…
• Band widths narrower than L5 & 7
• Allows discrimination of white mica from kaolinite & smectites (potentially)
• b6/b7 - general clays, white mica (?)
• b6/(b5+b7) - kaolinite
• Not easy and NB vegetation

Question 125

What is it possible to find from spectra in exploration and mapping?

Answer 125

• Chemistry of mineralized environments i.e. alteration zonation (kaolinite, illite, smectite, mica, chlorite, pyrophyllite, sulphates etc)
• Primary rock types (felsic, mica-rich, amphibole & chlorite rich mafic etc)
• Weathering regimes and processes (kaolinite and illite, smectites, gibbsite, sulphates etc)
• Fluid composition, T/pressure
  
  • e.g. Al/Mg-Fe substitution; high T species e.g. pyrophyllite, topas, dickite; fsp and albite chemistry in porphyry systems
• Impossible w broad-band imagery, partially achieveable using Aster. Really needs hyperspectral imagery

Question 126

full spectra vs image spectra, briefly ?

Answer 126

• Diagnostic absorption features seen using spectroscopy are swamped by atmospheric effects and many are t.f. undetectable using satellite imagery
• Broad band imagery
  
  • Bands too broad for identification
  • Can detect some spectral absorptions (using band ratios etc) but not well enough to identify mineral species
  • Groups of species or mineral associations may be identified using Aster (and potentially WorldView3)
• Hyper-spectral imagery (image spectrsocopy) is required to achieve this

Question 127

What are the common rock forming minerals that have diagnostic spectral features?

Answer 127

• Quartz (silica & silicates), iron, carbonates
• Substances containing O, H, C, AL and other transition metals
• Mixtures of these account for most earth materials

Question 128

What substances have reflective spectral features in the Visible (VIS) spectral region?

Answer 128

• Ferrous iron, Fe-AL silicates
• e.g. pigeonite, olivine, staurolite

Question 129

What substances have reflective spectral features in the Near Infra Red (NIR) spectral region?

Answer 129

• Iron oxides and hydroxides, ferric iron
• e.g. haematite, jarosite, goethite, limonite
• e.g. basic igneous mineral assemblages (Fe bearing silicates)

Question 130

What substances have reflective spectral features in the Short-Wave Infra-Red (SWIR) spectral region?

Answer 130

• Silicates, alterations minerals and natural weathering products (aluminous micas and clay minerals - OH bearing minerals),
• H-O-H bond stretching e.g. gypsum
• Metal-hydroxyl bond bending e.g. montmorilonite, kaolinite, muscovite
• C-O bond bending e.g. caclite, dolomite and magnesite

Question 131

What substances have reflective spectral features in the Mid or Thermal Infra-Red (TIR) spectral region?

Answer 131

• Si-O bond stretching
• E.g. quartz (behaves similar to a black-body radiator), albite, orthoclase, anorthite, labradorite, muscovite, augite, hornblende, olivine, garnets & carbonates
• Shift of the absorption trough to longer wavelenghts with the transition from felsic to mafic

Question 132

full spectra vs image spectra, briefly ?

Answer 132

• Diagnostic absorption features seen using spectroscopy are swamped by atmospheric effects and many are t.f. undetectable using satellite imagery
• Broad band imagery
  
  • Bands too broad for identification
  • Can detect some spectral absorptions (using band ratios etc) but not well enough to identify mineral species
  • Groups of species or mineral associations may be identified using Aster (and potentially WorldView3)
• Hyper-spectral imagery (image spectrsocopy) is required to achieve this

Question 133

What is it possible to find from spectra in exploration and mapping?

Answer 133

• Chemistry of mineralized environments i.e. alteration zonation (kaolinite, illite, smectite, mica, chlorite, pyrophyllite, sulphates etc)
• Primary rock types (felsic, mica-rich, amphibole & chlorite rich mafic etc)
• Weathering regimes and processes (kaolinite and illite, smectites, gibbsite, sulphates etc)
• Fluid composition, T/pressure
  
  • e.g. Al/Mg-Fe substitution; high T species e.g. pyrophyllite, topas, dickite; fsp and albite chemistry in porphyry systems
• Impossible w broad-band imagery, partially achieveable using Aster. Really needs hyperspectral imagery

Question 134

What substances have reflective spectral features in the Visible (VIS) spectral region?

Answer 134

• Ferrous iron, Fe-AL silicates
• e.g. pigeonite, olivine, staurolite

Question 135

full spectra vs image spectra, briefly ?

Answer 135

• Diagnostic absorption features seen using spectroscopy are swamped by atmospheric effects and many are t.f. undetectable using satellite imagery
• Broad band imagery
  
  • Bands too broad for identification
  • Can detect some spectral absorptions (using band ratios etc) but not well enough to identify mineral species
  • Groups of species or mineral associations may be identified using Aster (and potentially WorldView3)
• Hyper-spectral imagery (image spectrsocopy) is required to achieve this

Question 136

What is it possible to find from spectra in exploration and mapping?

Answer 136

• Chemistry of mineralized environments i.e. alteration zonation (kaolinite, illite, smectite, mica, chlorite, pyrophyllite, sulphates etc)
• Primary rock types (felsic, mica-rich, amphibole & chlorite rich mafic etc)
• Weathering regimes and processes (kaolinite and illite, smectites, gibbsite, sulphates etc)
• Fluid composition, T/pressure
  
  • e.g. Al/Mg-Fe substitution; high T species e.g. pyrophyllite, topas, dickite; fsp and albite chemistry in porphyry systems
• Impossible w broad-band imagery, partially achieveable using Aster. Really needs hyperspectral imagery

Question 137

What substances have reflective spectral features in the Mid or Thermal Infra-Red (TIR) spectral region?

Answer 137

• Si-O bond stretching
• E.g. quartz (behaves similar to a black-body radiator), albite, orthoclase, anorthite, labradorite, muscovite, augite, hornblende, olivine, garnets & carbonates
• Shift of the absorption trough to longer wavelenghts with the transition from felsic to mafic

Question 138

What substances have reflective spectral features in the Short-Wave Infra-Red (SWIR) spectral region?

Answer 138

• Silicates, alterations minerals and natural weathering products (aluminous micas and clay minerals - OH bearing minerals),
• H-O-H bond stretching e.g. gypsum
• Metal-hydroxyl bond bending e.g. montmorilonite, kaolinite, muscovite
• C-O bond bending e.g. caclite, dolomite and magnesite

Question 139

What substances have reflective spectral features in the Near Infra Red (NIR) spectral region?

Answer 139

• Iron oxides and hydroxides, ferric iron
• e.g. haematite, jarosite, goethite, limonite
• e.g. basic igneous mineral assemblages (Fe bearing silicates)

Question 140

What substances have reflective spectral features in the Mid or Thermal Infra-Red (TIR) spectral region?

Answer 140

• Si-O bond stretching
• E.g. quartz (behaves similar to a black-body radiator), albite, orthoclase, anorthite, labradorite, muscovite, augite, hornblende, olivine, garnets & carbonates
• Shift of the absorption trough to longer wavelenghts with the transition from felsic to mafic

Question 141

What substances have reflective spectral features in the Short-Wave Infra-Red (SWIR) spectral region?

Answer 141

• Silicates, alterations minerals and natural weathering products (aluminous micas and clay minerals - OH bearing minerals),
• H-O-H bond stretching e.g. gypsum
• Metal-hydroxyl bond bending e.g. montmorilonite, kaolinite, muscovite
• C-O bond bending e.g. caclite, dolomite and magnesite

Question 142

What substances have reflective spectral features in the Near Infra Red (NIR) spectral region?

Answer 142

• Iron oxides and hydroxides, ferric iron
• e.g. haematite, jarosite, goethite, limonite
• e.g. basic igneous mineral assemblages (Fe bearing silicates)

Question 143

What substances have reflective spectral features in the Visible (VIS) spectral region?

Answer 143

• Ferrous iron, Fe-AL silicates
• e.g. pigeonite, olivine, staurolite

Question 144

What are the common rock forming minerals that have diagnostic spectral features?

Answer 144

• Quartz (silica & silicates), iron, carbonates
• Substances containing O, H, C, AL and other transition metals
• Mixtures of these account for most earth materials

Question 145

Describe how the reflective spectral properties of Hydrated minerals in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 145

• There are still only 2 SWIR bands…
• Band widths narrower than L5 & 7
• Allows discrimination of white mica from kaolinite & smectites (potentially)
• b6/b7 - general clays, white mica (?)
• b6/(b5+b7) - kaolinite
• Not easy and NB vegetation

Question 146

Describe how the reflective spectral properties of Hydrated minerals in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 146

• Hydrated (clay) mineral absorption features in the SWIR
• Shape, position and symmetry of absorption features enables species-level identification of clay minerals
  
  • Al-OH bond bending produces distinctive features near 2.25 um.
  • Absorption caused by HOH bonds at 1.4 & 1.9 um.
• But in Landsat TM we can only detect the general presence of clay minerals - b7 straddles the diagnostic absorptions 2.1 to 2.35 um.
• In general, clay minerals reflect strongly in mid-SWIR and absorb in far-SWIR (not muscovite)
• e.g. TM 5/7 ratio (reveals hydrothermal alteration minerals and t.f. is an exploration tool)

Question 147

Describe how the reflective spectral properties of Iron oxides/hydroxides in Sentinel-2 can be used for interpretation of images and spectra.

Answer 147

• Similar to Landsat 8
• More VNIR bands and narrower bandwidths
• Allows potentially much better discrimination (and possibly identification)

Question 148

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 148

• Band widths are much narrower
• Extra coastal (new b1) added
• Allows better discrimination (potentially)
• (b1+b2+b3) / (b4 + b5) - haematite
• b3/b2 or b4/b2 - goethite & jarosite
• b4/b1 - jarosite (possibly)

Question 149

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 149

• Absorption minima shift to longer wavelengths within the iron oxide family, enabling species-level identification in lab. spectra.
• Reflectance minima at diagnostic wavelengths
• But in Landsat image spectra we can only detect the general presence of iron-oxides
• All iron oxides reflect strongly in band 3 and absorb in band 1
• Iron-oxides can be most effectively identified using a 3/1 ratio (Fe-oxide rich gossans are often associated with mineral deposits so this is an exploration tool)

Question 150

Vibrational Transitions (VT) take place where?

Answer 150

Most VT occur in the SWIR (some in TIR) and ‘IR vibrationally active’ minerals have dipole moments

Question 151

What are the diagnostic effects of Electronic Transitions upon absorption?

Answer 151

Ordinarily there are no truly diagnostic effects (broad) but there are several subtle effects which are useful.

Mainly, Crystal Field (CF) effects and
Charge Transfer (CT) effects

Question 152

Explain the cause of Charge Transfer (CT) absorptions.

Answer 152

• The absorption of a photon causes an electron to move between ions or between ligands.
• For any substance, one of its components must have electron donating properties and another component must be able to accept electrons.
• Absorption of radiation then causes the transfer (jump) of an electron from the donor (ligand orbital) to an orbital associated with the acceptor (metal orbital). Usually when metal is in oxidation (oxidised?) state.

Question 153

CF is the most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).

Why the transition metals?

Answer 153

• Outer electron shells partially filled
• d orbital energy states are split when an atom lies in a crystal field - the splitting allows an electron to move to a higher level (Crystal Field Theory)

Question 154

Re: spectral behaviour; explain the causes of electronic transitions (ET)

Answer 154

• Isolated atoms and ions have discrete energy states.
• Absorption of photons of a specific wavelength causes a change from one energy state to a higher one (excitation).
• On relaxation, emission of a photon occurs, with a drop in energy state to a lower one.
• When a photon is absorbed it is usually not emitted at the same wavelength (e.g. causes heating, on relaxation T drops)
• Atoms can rotate and vibrate with respect to each other
• Electrons can migrate from one energy level to another
• All require high energy (effects confined to VNIR)

Question 155

What processes cause absorption (re: spectral analysis)?

Answer 155

• Electronic transitions (Vis & NIR)
  
  • Crystal field effects
  • Charge transfer
• Vibrational transitions (SWIR & TIR)
  
  • Bond stretch, bending & rotation

In all cases, absorbed energy is later emitted at longer wavelength than it is absorbed

Question 156

What are the chemical causes of the various types of spectral behaviour?

Answer 156

• Refractive index (ratio of the speed of light through one material to that through another)
  
  • Reflectance, transmission
  • Wavelength-dependent
• ** Absorption coefficient ** (creates most significant signature)
  
  • Chemical composition & crystallography
  • Electronic and/or vibrational processes
  • Wavelength-dependent

Question 157

What substances have reflective spectral features in the Near Infra Red (NIR) spectral region?

Answer 157

• Iron oxides and hydroxides, ferric iron
• e.g. haematite, jarosite, goethite, limonite
• e.g. basic igneous mineral assemblages (Fe bearing silicates)

Question 158

What substances have reflective spectral features in the Short-Wave Infra-Red (SWIR) spectral region?

Answer 158

• Silicates, alterations minerals and natural weathering products (aluminous micas and clay minerals - OH bearing minerals),
• H-O-H bond stretching e.g. gypsum
• Metal-hydroxyl bond bending e.g. montmorilonite, kaolinite, muscovite
• C-O bond bending e.g. caclite, dolomite and magnesite

Question 159

Re: spectral behaviour; explain the causes of electronic transitions (ET)

Answer 159

• Isolated atoms and ions have discrete energy states.
• Absorption of photons of a specific wavelength causes a change from one energy state to a higher one (excitation).
• On relaxation, emission of a photon occurs, with a drop in energy state to a lower one.
• When a photon is absorbed it is usually not emitted at the same wavelength (e.g. causes heating, on relaxation T drops)
• Atoms can rotate and vibrate with respect to each other
• Electrons can migrate from one energy level to another
• All require high energy (effects confined to VNIR)

Question 160

What are the diagnostic effects of Electronic Transitions upon absorption?

Answer 160

Ordinarily there are no truly diagnostic effects (broad) but there are several subtle effects which are useful.

Mainly, Crystal Field (CF) effects and
Charge Transfer (CT) effects

Question 161

What processes cause absorption (re: spectral analysis)?

Answer 161

• Electronic transitions (Vis & NIR)
  
  • Crystal field effects
  • Charge transfer
• Vibrational transitions (SWIR & TIR)
  
  • Bond stretch, bending & rotation

In all cases, absorbed energy is later emitted at longer wavelength than it is absorbed

Question 162

CF is the most common electronic process and is caused by excitation of unfilled electron shells particularly of transition elements (Ni, Cr, Co, Fe, etc.).

Why the transition metals?

Answer 162

• Outer electron shells partially filled
• d orbital energy states are split when an atom lies in a crystal field - the splitting allows an electron to move to a higher level (Crystal Field Theory)

Question 163

Explain the cause of Charge Transfer (CT) absorptions.

Answer 163

• The absorption of a photon causes an electron to move between ions or between ligands.
• For any substance, one of its components must have electron donating properties and another component must be able to accept electrons.
• Absorption of radiation then causes the transfer (jump) of an electron from the donor (ligand orbital) to an orbital associated with the acceptor (metal orbital). Usually when metal is in oxidation (oxidised?) state.

Question 164

What are the benefits of Charge Transfer (CT) absorptions?

Answer 164

• Usually diagnostic at mineralogical level
• Strengths of these effects are 100-1000s of times stronger than CF effects
• Dominantly in the UV but also extend into the VNIR

Question 165

How can Charge Transfer (CT) absorptions cause the same mineral-type to be in different colours?

Answer 165

• Can occur between the same metal in different oxidation states, such as between Fe²⁺ and Fe³⁺,
• i.e. where ions of more than one form of element exist together, charge transfers occur between them - producing minerals of different colours
• CT absorptions are the main cause of red colour in iron oxides and hydroxides

Question 166

What are the chemical causes of the various types of spectral behaviour?

Answer 166

• Refractive index (ratio of the speed of light through one material to that through another)
  
  • Reflectance, transmission
  • Wavelength-dependent
• ** Absorption coefficient ** (creates most significant signature)
  
  • Chemical composition & crystallography
  • Electronic and/or vibrational processes
  • Wavelength-dependent

Question 167

What are the physical causes of the various types of spectral behaviour?

Answer 167

Scattering effects

• Diffuse and/or specular reflection
• Volume and/or surface scattering
• Single and/or multiple scattering
• Wavelength, particle & surface - dependent

Question 168

Vibrational Transitions (VT) take place where?

Answer 168

Most VT occur in the SWIR (some in TIR) and ‘IR vibrationally active’ minerals have dipole moments

Question 169

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 169

• Absorption minima shift to longer wavelengths within the iron oxide family, enabling species-level identification in lab. spectra.
• Reflectance minima at diagnostic wavelengths
• But in Landsat image spectra we can only detect the general presence of iron-oxides
• All iron oxides reflect strongly in band 3 and absorb in band 1
• Iron-oxides can be most effectively identified using a 3/1 ratio (Fe-oxide rich gossans are often associated with mineral deposits so this is an exploration tool)

Question 170

What are the common rock forming minerals that have diagnostic spectral features?

Answer 170

• Quartz (silica & silicates), iron, carbonates
• Substances containing O, H, C, AL and other transition metals
• Mixtures of these account for most earth materials

Question 171

Describe how the reflective spectral properties of Hydrated minerals in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 171

• There are still only 2 SWIR bands…
• Band widths narrower than L5 & 7
• Allows discrimination of white mica from kaolinite & smectites (potentially)
• b6/b7 - general clays, white mica (?)
• b6/(b5+b7) - kaolinite
• Not easy and NB vegetation

Question 172

Describe how the reflective spectral properties of Hydrated minerals in Landsat 7 ETM+ can be used for interpretation of images and spectra.

Answer 172

• Hydrated (clay) mineral absorption features in the SWIR
• Shape, position and symmetry of absorption features enables species-level identification of clay minerals
  
  • Al-OH bond bending produces distinctive features near 2.25 um.
  • Absorption caused by HOH bonds at 1.4 & 1.9 um.
• But in Landsat TM we can only detect the general presence of clay minerals - b7 straddles the diagnostic absorptions 2.1 to 2.35 um.
• In general, clay minerals reflect strongly in mid-SWIR and absorb in far-SWIR (not muscovite)
• e.g. TM 5/7 ratio (reveals hydrothermal alteration minerals and t.f. is an exploration tool)

Question 173

Describe how the reflective spectral properties of Iron oxides/hydroxides in Sentinel-2 can be used for interpretation of images and spectra.

Answer 173

• Similar to Landsat 8
• More VNIR bands and narrower bandwidths
• Allows potentially much better discrimination (and possibly identification)

Question 174

Describe how the reflective spectral properties of Iron oxides/hydroxides in Landsat 8 OLI can be used for interpretation of images and spectra.

Answer 174

• Band widths are much narrower
• Extra coastal (new b1) added
• Allows better discrimination (potentially)
• (b1+b2+b3) / (b4 + b5) - haematite
• b3/b2 or b4/b2 - goethite & jarosite
• b4/b1 - jarosite (possibly)

Question 175

What are the physical causes of the various types of spectral behaviour?

Answer 175

Scattering effects

• Diffuse and/or specular reflection
• Volume and/or surface scattering
• Single and/or multiple scattering
• Wavelength, particle & surface - dependent

Question 176

What substances have reflective spectral features in the Mid or Thermal Infra-Red (TIR) spectral region?

Answer 176

• Si-O bond stretching
• E.g. quartz (behaves similar to a black-body radiator), albite, orthoclase, anorthite, labradorite, muscovite, augite, hornblende, olivine, garnets & carbonates
• Shift of the absorption trough to longer wavelenghts with the transition from felsic to mafic

Question 177

full spectra vs image spectra, briefly ?

Answer 177

• Diagnostic absorption features seen using spectroscopy are swamped by atmospheric effects and many are t.f. undetectable using satellite imagery
• Broad band imagery
  
  • Bands too broad for identification
  • Can detect some spectral absorptions (using band ratios etc) but not well enough to identify mineral species
  • Groups of species or mineral associations may be identified using Aster (and potentially WorldView3)
• Hyper-spectral imagery (image spectrsocopy) is required to achieve this

Question 178

What is it possible to find from spectra in exploration and mapping?

Answer 178

• Chemistry of mineralized environments i.e. alteration zonation (kaolinite, illite, smectite, mica, chlorite, pyrophyllite, sulphates etc)
• Primary rock types (felsic, mica-rich, amphibole & chlorite rich mafic etc)
• Weathering regimes and processes (kaolinite and illite, smectites, gibbsite, sulphates etc)
• Fluid composition, T/pressure
  
  • e.g. Al/Mg-Fe substitution; high T species e.g. pyrophyllite, topas, dickite; fsp and albite chemistry in porphyry systems
• Impossible w broad-band imagery, partially achieveable using Aster. Really needs hyperspectral imagery

Question 179

Discuss the principle of SPCA for colour composition.

(Not actually in q’s, might be useful for understanding)

Answer 179

• As PCs are independent without information redundancy, colour composites of PCs are often very effective to highlight particular ground objects and minerals that are not distinguishable in colour composites of the original bands.
• However, in PC colour composites, the noise may be exaggerated because the high rank PCs contain significantly less information than lower rank PCs and have very low SNR.
• When PC images are stretched and displayed in the same value range, the noise in higher rank PCs is improperly enhanced.
• We would like to use three PCs with comparable information levels for colour composite generation.
• Selective Principal Component Analysis (SPCA) techniques can produce PC colour composites condensing maximum information of either topographic or spectral features and meantime makes the information content in each PC displayed in RGB better balanced.
• There are two types of SPCA: dimensionality and colour confusion reduction and spectral contrast mapping.