改进保形半投影的无人机遥感图像拼接算法  

UAV remote sensing image stitching algorithm with improved SPHP

作　　者：赵辉[1,2] 杜雅欣 王红君 岳有军[1] ZHAO Hui;DU Yaxin;WANG Hongjun;YUE Youjun(Tianjin University of Technology,Tianjin 300384,China;School of Engineering Technology,Tianjin Agricultural College,Tianjin 300384,China)

机构地区：[1]天津理工大学,天津300384 [2]天津农学院工程技术学院,天津300384

出　　处：《现代电子技术》2021年第20期133-136,共4页Modern Electronics Technique

基　　金：天津市科技支撑计划项目(17ZXYENC00080;18YFZCNC01120)

摘　　要：在无人机遥感图像拼接过程中,针对由于无人机在遥感图像拼接过程中视差所造成的鬼影影响配准的问题,文章提出一种改进的保形半投影算法(SPHP)。首先,根据TPS模型将目标图像进行变形,实现与源图像投影对齐;然后,采用SPHP算法,根据求得的单应矩阵进行坐标变换,平滑的外推至非重叠区域;最后,进行图像融合。实验结果表明,所提算法在拼接视差无人机遥感图像时,能够有效地改善视差问题,提高配准精度,解决非重叠区域的形变失真问题。In the process of splicing UAV remote sensing images,an improved SPHP(shape⁃preserving half⁃projecting)algorithm is proposed to deal with the problem of ghosting registration caused by parallax in the process of splicing UAV remote sensing images.First,the target image is deformed according to the TPS model to achieve projection alignment with the source image.Then,the SPHP algorithm is used to perform coordinate transformation according to the obtained homography matrix,and smoothly extrapolate it to non⁃overlapping regions.Finally,image fusion is performed.The experimental results show that the algorithm proposed in this paper can effectively improve the parallax and registering accuracy,and solve the problem of deformation distortion in non⁃overlapping regions when it is used to stitch the UAV remote sensing images with parallax.

关 键 词：图像拼接 无人机遥感图像 SPHP 图像变形 坐标变换 图像融合 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP751.1[电子电信—信息与通信工程]