基于线性不变矩和角度向量的立体匹配算法  

A stereo matching method based on line moments and angle vector

作　　者：张令涛[1,2,3] 曲道奎[1,2] 徐方[1,2] 

机构地区：[1]中国科学院沈阳自动化研究所,沈阳110016 [2]沈阳新松机器人自动化股份有限公司,沈阳110168 [3]中国科学院研究生院,北京100049

出　　处：《高技术通讯》2009年第7期741-744,共4页Chinese High Technology Letters

基　　金：863计划(2007AA041701)资助项目。

摘　　要：针对室内场景双目立体匹配有别于一般场景立体匹配的特殊性,提出了一种计算简便、准确度高的立体图像匹配算法。该算法首先利用canny算子检测物体的边缘,根据边缘的线性不变矩寻找出目标物体,然后提取出目标物体轮廓的特征点,利用角度直方图计算出左右图像的旋转角度,最后利用角度向量实现左右图像的对应像素点的匹配。线性不变矩有效地将计算复杂度由二维降低到一维,大大降低了计算量。角度向量的提出降低了特征点匹配的复杂度,而且计算简便,准确率高。实验表明,该算法对图像的缩放、旋转、平移均免疫,具有较高的识别精度和良好的抗干扰性,计算效率高于传统方法,有着较高的应用价值。In view of the particularity of Indoor stereo matching, this paper presents an accurate and convenient stereo matching algorithm based on line moments and the angle vector. First, the algorithm finds the corresponding objects in stereo images by line moments, then calculates the rotation angle between matching contours by the angle histograms of the matching contours, and finally, finds corresponding matching points based on the angle vector. The complexity of calculation can be reduced from 2D to the 1D by using line moments, and the amount of calculation can be reduced greatly too. The angle vector is calculated simply and conveniently, making stereo matching accurate. It is proved by experiments that the stereo matching algorithm is immune from scale, rotation and translation.

关 键 词：线性不变矩 角度向量 立体匹配 角度直方图 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] O151.24[自动化与计算机技术—计算机科学与技术]