基于改进Harris算法的图像角点检测

针对Harris角点检测算法运行速度慢且单一阚值选取不当容易造成漏检正确角点或提取出较多伪角点的问题,提出了一种快速预筛选方法以提高检测效率,并结合多阈值角点提取及邻近角点剔除策略来改进算法.实验结果表明:单独使用该预筛选方式的改进算法仅用32.71％的时间就可以检测出原算法94.97％的角点;配合多阈值角点检测方法及临近点剔除策略时,改进算法的运行时间仅为原算法的61.94％,且检测出的角点分布更均匀,既能充分代表图像信息又有效地避免了角点簇拥现象.

Image Corner Detection Based on Improved Harris Algorithm

Owning to inappropriate selection of the single threshold,Harris corner detection algorithm presents the phenomenon of missing significant corners or detecting false corners.Thus a pre-selection method is presented to improve the efficiency of the algorithm,and multi-threshold corner extraction and neighbor points elimination strategy was applied to remove the crowded incorrect corners.The experimental results show that the algorithm adopting the proposed pre-selection method can detect as many as 94.97% of the corners of the original algorithm by using only 32.71% of computation time of the original one;by combining the multi-threshold corner extration with neighbor points elimination strategy,the computation time of the improved algorithm is only 61.94% that of the original one,meanwhile the detected corners are distributed more reasonable and can reflect the information of the image more efficiently.