Section 5 Flashcards
What are the two types of geometric correction?
Image to image
Image to map/vectors
What is an internal geometric error?
Errors introduced by the remote sensing scanner or with earths rotation.
What is another term for internal geometric error?
Systematic distortions.
What are the six types of systematic distortions?
Scan skew Mirror scan velocity Panoramic distortion Platform velocity Earth rotation Perspective
What are external geometric errors?
Movements by the aircraft like altitude change
Name and define the two types of external geometric errors
Altitude changes, attitude changes (roll pitch yaw)
What does changing altitude effect?
The scale of an image
What is more likely to have an attitude change, satellites or airplanes?
Airplanes
What is a way to prevent attitude distortions?
Gyro-stabilization equipment
What is the roll of a plane? And how will this effect an image?
Side to side rotation of a plane (wings dipping) and it will cause compression and expansion in the pixels of left and right side of the image
What is the pitch of an airplane? And what does it do to images?
The nose or tail of a plane dipping and it causes expansion and contraction in the pixels on the top and bottom of an image
What is the yaw of an image? And how does this effect an image?
This is the rotation directionally and it changes the rotation of the pixels
What is image skew in a satellite?
As earth rotates as a satellite scans the earth the pixels won’t align properly
What is image to image geometric correction?
When you need to compare and image to another the master slave correction will be between two images taken in the same location
What is image to map geometric correction?
An image is tied to a map that has the same projection
What is rectification?
The linking of a slave image to a master image
What is resampling?
The transfer of data from the slave image to the new geometrically corrected images
What is geometric correction?
Assigning coordinates to an image to make it comparable to another image
What is a GCP and what are they used for?
Ground control points. They are used to correct the slave image to the master image
True or false
GCPs should be clustered in the important part o an image
False they should be uniformly distributed
What are the two rules of selecting GCPs?
Points should be permanent and distinctive
What makes correcting images from airborne sensors more difficult?
Altitude can be different scan line to scan line
What is a frame capture?
Aerial photo
What formula checks the accuracy of the GCPs?
Root mean square error
RMS
This tells how off your points are from eachother
When geometrically correcting an image what is the step after selecting GCPs?
Generating a new project based on these points
What are three ways to generate a new project.
Conformal transformation
Affine transform (first order polynomial)
Higher order polynomial transformation
What is a conformal transform?
A geometric correction used to translate an image in one coordinate system to another. Image must be geometrically correct
Why is a conformal transform used? How many GCPs does it use?
Scale changes, rotation of images, translation to 3D. 2 GCPs
What is a first order transformation/affine transformation?
Similar to conformal, but image can only be changed in the x and y directions.
How many GCPs are needed for a first order polynomial?
3
What is a high order polynomial transformation?
A geometric correction that does not preserve parallel lines or shape
How many GCPs does a high order polynomial require?
6 or more
What is another term for higher order polynomial?
Rubber sheeting. This allows for differing geometry
What is the downside of higher order polynomial?
Edges of the image become unstable.
What are the three types of image resampling?
Nearest neighbour
Bilinear interpolation
Cubic convolution
What does the nearest neighbour resampling method do?
It preserves the true values of the original data. It is referred to as a zero order interpolator
Describe a bilinear interpolation resampling method
It is referred to as a first order interpolator. This method changes the original values by averaging pixel values
What is the advantage of using bilinear over NN?
There is a degree of smoothing to the image
What is the disadvantage of bilinear?
The original data numbers are not preserved
Describe how cubic convolution works.
It is an improvement on the bilinear in that more pixels are averaged to create new values
How is cubic better than bilinear?
There is less blur to the image
When is it better to use an NN resampling method?
When you will be taking measurements from the data
