针对Harris角点检测算法在应用中实时性较差和运算量较大, 同时抗噪能力较差等问题, 提出一种基于Harris算法的改进算法, 利用图像局部加权熵与最小亮度变化(MIC)算法相结合的方法进行角点检测。首先, 运用图像局部加权熵算法思想, 初步得出候选角点集;然后计算Harris算法的角点响应函数(CRF)值, 将候选角点按CRF值大小差分为三类; 最后使用自适应模板和阈值的MIC算法进行角点检测, 得出最佳匹配点。实验结果表明, 该方法提高了原算法的实时性, 增加了角点提取数量和准确性, 并且能够有效去除大多数伪角点。

In terms of the application of Harris algorithm, since the less real-time and a large amount of computation together with poor anti-noise ability and other issues, an improved corner detection algorithm is proposed based on the Harris, combing the local weighted entropy with minimum intensity change (MIC) algorithm. First of all, the candidate corner point set is computed through the local weighted entropy algorithm. Then the candidate corner is divided into three categories according to the corner response function (CRF) value of Harris algorithm. Finally, the best matching corners are obtained through the adaptive template and MIC algorithm of the threshold. Experimental results show that the proposed algorithm can improve the real-time of the original algorithm, increase the quantity of the corner extraction with better accuracy, and remove the most false corners effectively.