基于约束逼近投影变换的全景图像拼接方法-《集美大学学报（自然科学版）》

文章信息/Info

Title:: Panoramic Image Stitching Based on ConstraintAs-Projective-As-Possible Transform

作者:: 余清洲1; 2; 陈水利1; 3; 蔡国榕1; 4; 苏松志5; 吴云东1; 4; （1.厦门市无人机遥感应用工程技术研究中心，福建厦门 361021；2.集美大学理学院，福建厦门 361021；3.集美大学诚毅学院，福建厦门 361021；4.集美大学计算机工程学院，福建厦门 361021；5.厦门大学信息科学与技术学院，福建厦门 361005）

Author(s):: YU Qing-zhou1; 2; CHEN Shui-li1; 3; CAI Guo-rong1; 4; SU Song-zhi5; WU Yun-dong1; 4; （1.Xiamen UAVRS Application Engineering Technology Research Center,Xiamen 361021,China; 2.School of Science,Jimei University,Xiamen 361021,China;3.Chengyi College,Jimei University, Xiamen 361021,China;4.School of Computer Engineering,Jimei University,Xiamen 361021,China; 5.School of Information Science and Technology,Xiamen University,Xiamen 361005,China）

Keywords:: Projective; Warp; Moving DLT; As-Projective-As-Possible Warps; Bundle Adjustment

摘要:: 全景图像拼接算法是遥感信息处理、图像处理等领域的研究热点.基于逼近投影变换的算法对低空遥感图像拼接有较好的效果，但对于远离参考图像的矩形块会出现较大的畸变并且拼接缝周边会出现大量的形变、鬼影等现象.为此，在逼近投影变换的基础上，提出一种基于光束平差与投影变换相结合的图像拼接策略，目的是通过约束变换总误差来减少畸变现象.针对低空航拍图像的实验结果表明，与传统拼接算法相比，新方法减少了图像中的鬼影和模糊，且避免了建筑物、道路等地物的畸变现象.

Abstract:: Panoramic image mosaic algorithm is a hot topic in the fields of remote sensing and image processing.State-of-the-art such as As-Projective-As-Possible (APAP) warp is efficient in low altitude remote sensing image mosaic.However,in APAP related algorithms,the rectangular block away from the reference image and seams suffer from large scale deformation and parallax,respectively.Therefore,in this paper,a novel algorithm that is based on the as-projective-as-possible warps and bundle adjustment is proposed.The purpose is to control the total transform error,thus to reduce image distortion.Experimental results conducted on lowaltitude aerial images show that the proposed method reduces the phenomenon of ghosting and blurring.Moreover,the distortions of buildings and roads have been alleviated.