Automating Generalisatin

Question 1

Issues with manual representation

Answer 1

• time consuming

- problems with consistency - different people will produce different things - subjective

Question 2

Automate generalisation

Answer 2

Algorithms are used to perform tasks that were previously done manually. Research began in the 1960s and heavily invested into in the 1970s (Burrough et. al 1998)

Consistent

Question 3

Limitations to automated generalisation

Answer 3

Extremely difficult to capture what a cartographer does and for a computer to replicate = difficult to get computers to design things. Practice can differ from theory - experience and rule of thumb

Question 4

Independent processing

Answer 4

Often used in the early days to reduce data when storage was an issue. Random choice in retaining points = this is unlikely to give a good generalisation as some points are more significant than others.

Nth point routine = every nth point is retained. This may result in critical points getting lost. It depends on the starting point

Question 5

Local processing- 3 aspects

Answer 5

Looks at small selection of line using distant or angle or both criteria to decide on the data that should be kept.

• distance
• angular
• angle and distance

Question 6

Distance

Answer 6

Weed out points that are close together. Points close together can be removed - often used to avoid many close points.

Mainly used when capturing data.

Question 7

Angular constraints

Answer 7

Points on or close to a straight line add little information and so can be removed = parallel corridors. First point outside the corridor becomes the new point.

Question 8

Angle and distance

Answer 8

Examine local relationships to remove non critical points. Distance = Doesn’t consider the curvature of a line Angular = doesn’t consider spacing of points. Combination takes both into account.

Most likely to remove minor(insignificant) elements but retain general characteristics

Question 9

Reuman - witkam algorithm

Answer 9

Similar to angular method. Retains last point in the corridor as a new anchor

Question 10

Opheim

Answer 10

Similar to R-W but the max length is specified. Removes points in minimum distance

Question 11

Angle + distance - Jenks

Answer 11

Works using 3 points to decide if the middle point should be eliminated.

• if p2 is closer to p1 than min 1, then eliminate p2
• if p3 is closer to p1 than min 2, then eliminate p2
• if angle at p2 is less then ang, then eliminate p2

If p2 is retained it becomes themes starting point if it is eliminated, p3 becomes p2

Question 12

Global processing routines

Answer 12

Most allied by Doug,as-Poiker although has received criticism. It divides the line into two,looking for more critical points. Use multiple tolerances and stores attributes.

Question 13

Issues with algorithms

Answer 13

No simple answerable which algorithm is the best it depends on the nature

Question 14

Linear smoothing

Answer 14

When generalizing a point it might be best to smooth it. This eliminates distinguishing errors and will improve aesthetic appearance

Question 15

Generalisation and spatial conflicts

Answer 15

Often can’t consider just a single class or feature I.e roads. Scale reductions can create problems between relationships - features must be retained.