Polygon matching using centroid distance sequence in polar grid
Bitao Fu, Huang Liuyin · 2016
A silhouette polygon matching method is proposed based on centroid distance in polar grid. With the polygon centroid as the origin, all the vertexes of the polygon in an image are converted into the polar coordinates. The polygon is partitioned into many grids according to the polar angle. The average centroid distance in each grid is considered as the descriptor and the similarity measure of the polygon. In addition, normalized polygon area is applied to speed up the progress of similar polygons searching. Experiments show that its calculating speed is higher by an order of magnitude than traditional algorithm using polygon feature point sets extraction and matching.