Polygon graphic recognition based on improved fast corner detection

Abstract

Abstract The accuracy of polygon graphic recognition based on chain code features and Hough transform is low, and the computation is limited. Therefore, this paper proposes a polygon graphic recognition method based on improved features from accelerated segment test (FAST) corner detection. First, hole filling and Freeman chain code are used to segment the image, and the regular geometric features are obtained. Second, in order to improve the performance of the algorithm, an improved FAST corner detection combined with DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is proposed. Through clustering, false corner in the image can be eliminated, and the feature points are quickly filtered by local NMS (non-maximum suppression). Finally, polygon graphic recognition is realized by feature points, experimental results illustrate that this method has high calculation efficiency and recognition rate.