光学技术, 2017, 43 (5): 450, 网络出版: 2017-11-07
利用共线性和对极约束实现匹配扩展
Matching extension based on colinearity and epipolar geometry constraint
图像处理 图像匹配 匹配扩展 共线性约束 对极约束 image processing image matching matching extension colinearity constraint epipolar geometry
摘要
结构化场景图像的纹理较为单一, 可提取的局部特征较少。在宽基线情况下, 基于图像局部特征的匹配方法很难获取正确匹配, 导致匹配精度降低。为此, 提出了一种以图像局部特征匹配为基础进行配扩展的方法。匹配扩展时, 将图像中线段的两端点作为匹配的基元, 利用线段两端点在不同视点间的共线性和对极约束限定候选匹配的搜索空间, 剔除错误匹配, 进行匹配关系的传播。实验结果表明: 当图像存在一定的尺度和视点变化以及明显的光照变化时, 该方法能够有效地增加匹配的数目, 为精确的极线几何估计和三维重建奠定了良好的基础。
Abstract
Structured scenes have uniform texture and the local features that can be detected are less. Matching method based on local features can hardly get correct matches in wide baseline case, resulting in low matching performance. A matching extension method based on colinearity and epipolar geometry constraint is presented. The two endpoints of each line existed in images of different views are used as candidate matching features simultaneously. Colinearity constraint and epipolar geometry of the two endpoints are used to reduce the searching space and reject outliers in the matching extension procedure. Experiment results show that the method can adapt to a certain degree of scale and viewpoints change and obviously illumination change. The initial matches are greatly enriched, which lay good foundation for accurate epipolar geometry estimation and three dimensional reconstructions.
翟优, 郭希维, 何鹏, 曾峦. 利用共线性和对极约束实现匹配扩展[J]. 光学技术, 2017, 43(5): 450. ZHAI You, GUO Xiwei, HE Peng, ZENG Luan. Matching extension based on colinearity and epipolar geometry constraint[J]. Optical Technique, 2017, 43(5): 450.