摘要
针对侧扫声纳图像分辨率高测深精度低而多波束声纳图像分辨率低测深精度高的特点,提出了一种基于SUFR的声纳图像自动配准与融合方法。该算法检测同一区域内侧扫声纳图像和多波束图像的特征点,通过最近邻匹配获得匹配点后,计算图像间的变换矩阵,利用空间变换完成配准,采用加权融合法实现两者的融合。实验结果表明该算法具有很好的鲁棒性,配准精度达到像素级,可实现两者的高精度自动配准与融合,取得了理想的效果。
According to multi-beam sonar system,the high-resolution backscatter but poor horizontal position accuracy,and side-scan sonar system,the accurate bathymetry and horizontal position but low resolution,the study is concerned with an automatic registration and fusion method of sonar image based on SURF.To achieve the integration of multi-beam sonar system and side-scan sonar system with the weighted fusion method,it extracts feature points by using SURF,computes transformation matrix by using match points,and performs registration and fusion with a spatial transform.The results indicate that this method is robust and stable with registration accuracy up to pixel level realizing the quite precise automatic registration and fusion,and is more suitable for sonar image.
出处
《测绘与空间地理信息》
2013年第3期56-58,64,共4页
Geomatics & Spatial Information Technology
基金
广州海洋地质调查局天然气水合物专项数据库建设及战略研究项目(GHZ201100312)资助
