图像特征点提取及匹配技术研究

Image Feature Point Extractionand Matching Technology Research

图像特征点提取和图像匹配技术是图像处理领域中的重要技术,同时是多种图像处理及应用的基础。本文主要研究利用Forstner算子、Moravec算子、Harris算子、SIFT算子对近景影像和航空影像进行特征点提取;接着,采用基于SIFT的匹配来确定匹配同名像点,利用核线约束匹配的搜索范围,约束范围内的特征点;接着,采用灰度相似性约束和双向一致性约束对匹配结果进行检核;最后,得出特征提取和特征匹配的结果。基于本文研究的特征点影像提取及匹配方法,分别选取一组近景影像和数码航空影像进行实验,实验结果证明了本文方法的可靠性和高效性。

Image feature extraction and image matching techniques are important in the fields of image processing technology, and is the basis for a variety of image processing applications. This paper studies the use of Forstner operator, Moravec operator, Harris oper- ator and SIFFoperatorfor close - range images and aerial imagery for feature point extraction, and then, based on SIFT matching to de- termine a match with the same name as the point epipolar constraint match for your search range, within the constraints of the feature points, and then use the gray similarity constraints and matching bidirectional consistency constraints on nuclear inspection results, the conclusion that extraction and matching results. Image feature point extraction based on this study and the matching method to select one close - range image and digital aerial image were experiments. Experimental results show that the reliability and efficiency of this method.