Abstract:

In view of some shortcomings about currently shape recognition algorithms such as a large amount of calculation, long processing time, fewer species recognition, a fast geometry figure recognition algorithm based on chain code feature was presented. The algorithm combines the chain code histogram and chain code spatial distribution entropy, which considers both the statistical feature and the spatial feature of the chain code, having the advantages of being invariant to the position, rotation and scale of the image content and having nothing to do with the start point of the chain code. The simulation results show that the algorithm is able to identify many types of graphics with high recognition accuracy and shorter time consuming, and the overall performance is excellent.

Key words: geometry figure, chain code feature, information entropy, recognition