摘要
本文提出一种基于自适应非最大抑制策略的Harris角点检测算法;算法从角点响应函数的局部极大值入手,计算图像中每个像素点的局部极大值通过设置抑制半径与角点响应函数的局部最大值关联最终获取角点。实验表明,该算法检测出的角点均匀分布,提高了检测角点的准确性,算法能够应用于图像特征匹配、运动估计等方面。
The traditional algorithm＇s detection result is not ideal, so Harris corner detection algorithm based on self-adapting non-maximal suppression strategy is proposed. The algorithm detects the pixels which corner response function is local maximum, corners are extracted by the suppression radius continuously decreasing. The problem of corner clustering and the difficulty of threshold selection are effectively avoided. Experimental result shows that the algorithm is able to detect corner evenly distributed, and the accuracy of the detection is also improved. This algorithm could be used for feature matching and motion estimation, etc quite well.
出处
《价值工程》
2015年第8期267-269,共3页
Value Engineering
