多源高分辨率遥感影像自动匹配算法

提出一种基于同名直线约束的多源高分辨率遥感影像自动匹配方法。对影像建立高斯滤波模型去噪,然后进行边缘检测,在边缘影像上进行Hough变换,通过端点检测的方法来获取直线;通过粗匹配参数进行同名直线粗匹配,应用梯度模型对同名直线进行精匹配;对影像提取特征点,特征点经过同名直线约束后,在对应影像上利用同名直线约束建立匹配范围。采用最小欧式距离准则在给定范围内提取初始同名点;利用RANSAC算法剔除错误同名点对,以获取最终的匹配结果。实验结果表明,与传统的SIFT匹配算法比较,方法具有可靠性好、提取同名点数量多的优点。

Automatically matching method for multi-source high resolution remote sensing images

It presents a matching method for multi-source high resolution remote sensing images based on corresponding lines. Firstly, Gaussian model is employed for image filtering, then line extracting is performed by Hough transform and Endpoint detection algorithm from the edge of the image; secondly, a coarse-to-fine line matching strategy is proposed based on coarse transformed parameters and gradient model; thirdly, feature points are extracted and restricted by corresponding lines, and the constraint model is built by corresponding lines; fourthly, the minimum Euclidean distance criterion is used to extract the initial homonymy points within the window constraint; at last, RANSAC is imbedded to eliminate mismatching points. Experimental results verify that the presented method has higher matching point than that resulted from SIFT matching method with the better reliability.