Testing Simple Polygons

Abstract We consider the problem of verifying a simple polygon in the plane using “test points”. A test point is a geometric probe that takes as input a point in Euclidean space, and returns “+” if the point is inside the object being probed or “−” if it is outside. A verification procedure takes as input a description of a target object, including its location and orientation, and it produces a set of test points that are used to verify whether a test object matches the description. We give a procedure for verifying an n -sided, non-degenerate, simple target polygon using 5 n test points. This testing strategy works even if the test polygon has n + 1 vertices, and we show a lower bound of 3 n + 1 test points for this case. We also give algorithms using O( n ) test points for simple polygons that may be degenerate and for test polygons that may have up to n + 2 vertices. All of these algorithms work for polygons with holes. We also discuss extensions of our results to higher dimensions.

[1]  Steven Salzberg,et al.  Testing Orthogonal Shapes , 1995, Comput. Geom..

[2]  W. Rudin Real and complex analysis , 1968 .

[3]  Steven Skiena,et al.  Geometric probing , 1988 .

[4]  Richard Cole,et al.  Shape from Probing , 1987, J. Algorithms.

[5]  Kathleen Ann Rebecca Romanik Approximate testing theory , 1992 .

[6]  W. Clem Karl,et al.  Reconstructing objects from projections , 1991 .

[7]  S. Skiena Interactive reconstruction via geometric probing , 1992, Proc. IEEE.

[8]  Thomas C. Shermer,et al.  Probing Polygons Minimally Is Hard , 1992, Comput. Geom..

[9]  Carl Smith,et al.  Testing Geometric Objects , 1994, Comput. Geom..