Coordinate Geometry and Geometric Calculations


(from overheads, no links)

Objectives of this lecture
1 Review linkage of analytical geometry (trigonometry) for GIS
2 Examine simple calculations
3 Establish techniques for treating lines
4 Consider concerns of speed, efficiency, numerical stability


Basic Coordinate Geometry: A Thought-Experiment

1) How do you obtain coordinates if:

a) you are at a known point (such as a horizontal control marker) and observe an angle and a distance to some object.
b) you record the successive angles and distances of a "trip" (traverse) from a known point, out for some number of legs and then back to the point of beginning. [What do you expect about the "closure" of this traverse? What will you have to do if the "circle" is not 360.00000°?]
c) you record a lot of angles from two known points. What else do you need to know? [What would you do with angles recorded from many points (some known and some unknown?]

d) Would it be any better if you had distances instead of angles?


2) How would you determine bearing and distance from coordinates?

3) Take some sample parcel descriptions and figure out their coordinate representations.
(examples: 2209 Mass Ave, Lexington MA and 8687 Oddfellow Road Bainbridge Island WA)


Fundamental Geometric Algorithms

Wise starts off introducing equation for a line as:

the point (x,y) such that y = a + b*x.

This is a dubious choice. See snippet from Programmer's Geometry about ax +by +c =0. Even more useful if "normalized" so that a-squared plus b-squared =1. Then a and b are "direction cosines" of the angles that the normal to the line makes with the axes, c is - distance to the origin. Neat?

Distance:

equation is bone-crushingly simple (Pythagoraean Theorem), but still raises issues of calculation. For example, why take square root? (square your tolerance instead...)

Bearing of a line

Distance from a point to a line:

raises issues of how to represent a line; useful for generalization
Handout from Programmer's Geometry shows how simple it can be...

 

Topological algorithms

(for treatment later) Point-in-polygon, segment intersection, polygon overlay


Roundoff and numerical stability:

the view of Prof. Kahan
Paradoxes in concepts of accuracy (numerical representations)
example of area of triangle calculations




Index from here: Next lecture | Schedule | Questions |

Version of 4 January 2003