低光照环境下基于特征点补偿的图像匹配方法研究

Research on image matching method based on feature point compensation in low light environment

图像匹配主要通过匹配图像中的大量特征点来实现图像处理,它是目标识别跟踪和室内定位导航等应用领域的重要技术基础。在低光照环境下,采集的图像往往存在局部特征缺失的问题,使提取的特征点数目损耗较大,导致图像匹配精度降低。因此,提出了一种基于同态滤波和多尺度色彩保护(Multi-scale retinex with chromaticity preservation,MSRCP)的特征点补偿算法,使用随机采样一致性(Random sample consensus,RANSAC)算法进行图像匹配。研究结果表明:在低光照环境下,采用基于特征点补偿的图像匹配方法可有效地增加特征点的提取数目,提高了图像匹配准确率,准确率最高可达74.32%,在一定程度上解决了图像匹配困难的问题。

Image matching is mainly realized by matching a large number of feature points in the image,which is an important technical basis for object recognition tracking,indoor positioning navigation,and other application fields.Many images acquired in low light environments often have local features missing,which causes a large loss of the number of extracted feature points,resulting in a reduction of image matching accuracy.The feature point compensation algorithm is proposed based on homomorphic filtering and Multi-scale retinex with chromaticity preservation(MSRCP),and the Random sample consensus(RANSAC)algorithm is used for image matching.The experimental results show that in the low light environment,the image matching method based on feature point compensation can improve the accuracy of image matching by effectively increasing the number of feature points,and the accuracy rate can reach 74.32%,which largely solves the problem of image matching difficulty.