一种改进的Hausdorff距离的图像匹配算法仿真分析

在农作物病虫害的图像诊断中,应用基于Hausdorff距离的边缘图像匹配算法可简便计算。但Haus-dorff距离缺点是不能处理目标被遮掩和外部点存在的情形,鲁棒性不足。针对传统Hausdorff距离匹配算法的缺点,引入一种改进的Hausdorff距离算法。改进的Hausdorff距离匹配算法可以对有噪声、遮掩、分辨率降低和水平角度旋转的图像进行很好的匹配。通过仿真,证明了改进的Hausdorff算法具有更强的鲁棒性和适应性,能在有畸变的图像中估计出最佳的匹配位置。

A simulating analysis about image matching algorithm based on an improved Hausdorff distance

In the image diagnosis on crop pests and diseases,the image edge matching based on the Hausdorff distance easily makes the calculation.This method,however,would be disabled when the target is covered or some points of the image are beyond of its matching field,and namely it is not robust.To reduce the disadvantage in the method,an improved Hausdorff distance algorithm has been invited.The new method can match well when the image is covered,stained,and rotated horizontally as well as in low resolution.Moreover,it has been proved by the simulation that the new method is more robust and adaptive matching the best position in an image with aberration.