基于相位角点检测的特征匹配研究

Feature Matching Based on Phase Congruency Corner Detection

基于特征的匹配是立体匹配中最常用的方法,但是匹配结果受特征检测精度的影响较大.针对这一问题,提出一种基于相位一致性角点检测的匹配算法,该算法采用相位一致性模型对图像中的角点特征进行检测,检测结果不受亮度、对比度等因素影响,因此在不同光照环境下的多幅图像可以使用相同的固定阈值,避免了特征检测中阈值选取的困难.在此基础上,结合场景的深度信息采用图像的灰度局部区域相关系数进行特征匹配.实验结果表明,该算法获得的匹配结果具有很高的正确匹配率.

Feature point matching is most commonly adopted among all kinds of stereo image matching.The result of feature point matching is affected by many factors such as object occlusions,lighting conditions and noises,therefore it is important to find a robust algorithm of feature point detection.In this paper described a new corner detector developed from the phase congruency model of feature detection.The new operator uses the principal moments of the phase congruency information to determine the corners and results in reliable feature detection under varying illumination conditions with fixed thresholds.A new feature point matching algorithm is then proposed.It employs the condition that the depth of the scene is locally continuous as extra constraint,and uses the method for extended assignment problem for global optimization.Experiments showed that the results of the algorithm are satisfactory.