铁路视频监控中基于多算法结合的图像拼接

针对铁路视频监控系统中单个监控摄像头视角范围有限的问题,提出一种用于扩大监控视角的图像拼接方法。首先提出一种简单的输入图像排序算法,对输入的两张具有部分重叠的图像进行自动排序,可以提高监控系统的可靠性和图像的配准精度;然后采用尺度与旋转不变的SURF算法提取特征点,并用改进的RANSAC算法剔除误匹配点,进一步保证了图像配准的精度,使正确匹配的特征点达到总数的80%以上;最后以渐入渐出法进行图像平滑过渡融合,消除了由于光照引起的图像拼接缝隙。实验表明,本方法在图像拼接的速度和精度上取得了较好的效果,适用性较强。

Image mosaic method based on combination of some algorithms in railway video monitoring

Single camera in the railway video monitoring system has a finite visual angle.In order to solve this problem,a kind of image mosaic algorithm to enlarge the target view is proposed in the paper.First,a simple method is presented to sort two input images that have some region overlapped,which can improve the reliability of the system and the registration precision of two images automatically.And then one scale and rotation invariant algorithm named SURF is adopted to extract the feature points.After finishing the match of these points,an improved RANSAC algorithm is used to get rid of some wrong consequences,which can more effectively insure the registration precision and has the right match points more than 80%.At last,the paper chooses gradually in and out method as image fusion way,which removes the mosaic gap that caused by exposure difference.The experiment indicates that this image mosaic algorithm not only has a good efficiency in velocity and accuracy,but also is well of applicability.