Research On Extraction Of Straight Line And Circle By Random Hough Transform Based On Region Partition And Adaptive Clustering Algorithm

Linear and circular detection is one of the important problems in the process of digital image processing, linear and circular extraction and matching plays an important role in many fields, and has a broad application background. This article USES the method of edge detection to extract the edge points, and through the accurate extraction of linear and circular Hough transform, and is applied to extract information of linear and circular matching, this paper main research work is embodied in the following aspects:(1) This paper discusses the basic theory of the edge detection, summarized and several main methods of edge detection are reviewed and a few Kind of edge detection algorithms and their respective advantages and disadvantages. Hough transform(2) The research of basic theory, aiming at the defects of the large amount of calculation, the traditional Hough transform is proposed using a new.more of a. parameter mapping to replace the traditional Hough transform of.one to many, way of mapping, and combining the ideas of graph partition data greatly simplifies the computation.(3) In view of the Hough transform precision is difficult to ensure that the shortage of the peak detection process, put forward new algorithm based on the idea of two voting mechanism, at the expense of the smaller computation time, ruled out for the precision of the larger increase.(4) In the traditional Hough transform, the author of this paper for the shortage of the radius of circular extraction must be the default, is presented based on the feature analysis and extraction of circular circle string extraction algorithm, the simulation analysis of the algorithm has high computing efficiency and accuracy.