基于细节增强的自适应Harris角点检测算法

摘要

针对传统Harris角点检测法计算速度较慢、聚簇现象较严重和在纹理信息少的区域提取角点数较少等不足,提出一种自适应Harris角点检测法,该算法利用巴特沃斯滤波器增强小波细节系数,通过计算角点响应函数的二阶值设定自适应阈值,从而提高Harris角点检测的精度和效率.实验表明,与传统的Harris角点检测法比较,该方法检测速度较快、角点的分布较均匀,且伪角点较少.%To improve the weakness of Harris comer detection algorithm such as slow computing speed, serious cluster phenomenon and less comers extracted in weak texture information area, an adaptive Harris corner detection method is proposed in the paper. It enhances the wavelet detailed coefficients by using Butterworth filter and sets an adaptive threshold by calculating the second order of corner response function, therefore the accuracy and efficiency of Harris comer detection method are all improved. Experimental results show that, compared with classical Harris comer detection method, the proposed one has rapid detection speed, uniform distribution of comer and less pseudo comers.