倾斜影像匹配研究及应用

Research and Application of Tilt Image Matching

通过分析铁路沿线不同时期的倾斜影像数据,实现铁路沿线不同时期的设施和环境监测。为了解决当前铁路倾斜影像采用传统匹配算法匹配精度较低的问题,对SIFT算法和AKAZE算法进行了融合。首先采用AKAZE建立非线性尺度空间,实现特征区域的查找及边缘的保留。然后采用SIFT对关键点进行提取和描述,并采用RANSAC算法对关键点进行优化,提高匹配准确率。选取倾斜影像进行测试,融合算法的准确率、召回率以及加权平均值最优,证明了融合算法对于倾斜影像具有更好的匹配效果。

By analyzing the oblique image data of different periods along the railway,the monitoring of facilities and environment in different periods along the railway can be realized. In order to solve the problem of the low matching accuracy of the traditional matching algorithm,the SIFT algorithm and Akaze algorithm are fused in this paper. Firstly,AKAZE is used to establish nonlinear scale space to realize feature region searching and edge preserving. Then SIFT is used to extract and describe the key points;and RANSAC algorithm is used to optimize the key points to improve the matching accuracy. The fusion algorithm has the optimal accuracy,recall rate and weighted average value,which proves that the fusion algorithm has better matching effect for the tilted images.