利用外极线约束的图像匹配新算法

An Image Matching Algorithm Based on Epipolar Line Restriction

针对排爆机器人要求获取目标物的精确的三维坐标,在原来只用灰度匹配的基础上,提出了利用图像灰度局部特征约束的一种新算法。该算法首先对图像进行粗匹配找出可能出现的区域,再利用外极线约束进行精确匹配,这样就减少了搜索区间,提高了匹配精度。外极线约束算法对于有遮挡现象问题的处理也非常有效。

In order to improve the coordinate accuracy of target for the bomb-disposal robot, a new algorithm using local characteristic restriction of image grayscale has been proposed based on the original one, which uses grayscale matching only. Firstly, the new algorithm matches roughly to find the possible appearing area of characteristic points, and then matches accurately by epipolar line restriction. The algorithm significantly reduces the searching range, improves the accuracy of matching, and is very useful for shelter problems.