利用缝合线拼接无人机影像的方法  

Unmanned Aerial Vehicle Image Stitching Method Based on Seam Line

作　　者：陈晓轩 徐家利 CHEN Xiaoxuan;XU Jiali(Shaanxi Provincial Transport Planning Design and Research Institute,Xi’an 710065,China;Twenty-seventh Research Institute of China Electronics Technology Group Corporation,Zhengzhou 450018,China)

机构地区：[1]陕西省交通规划设计研究院有限公司,陕西西安710065 [2]中国电子科技集团公司第二十七研究所,河南郑州450018

出　　处：《地理空间信息》2024年第5期49-52,共4页Geospatial Information

基　　金：陕西交通科技资助项目(18-06K)。

摘　　要：针对无人机海量影像拼接生产应用需求,提出了一种基于梯度和准则影像的动态规划缝合线改进拼接方法;重构了差异准则影像,将梯度差异加入到差值影像中,对差值影像做进一步改进。以无人机影像为实验数据,对缝合线提取和拼接结果进行了验证分析。结果表明,该方法生成的最佳缝合线可以较好地避免穿过建筑物,解决了影像几何错位和重影问题,具有一定的应用价值。In order to meet the demand for the production and application of large-scale image stitching using unmanned aerial vehicle(UAV)imagers,we proposed an improved seam line stitching algorithm based on gradient and criterion images.We incorporated gradient differences into the difference images to reconstruct difference criterion images,and further improved the difference images.Taking UAV images as experimental data,we validated and analyzed the seam line extraction and stitching results.The results show that the proposed method can effectively avoid crossing buildings,solve the problems of geometric misalignment and image overlap,and has some practical value for application in engineering projects.

关 键 词：无人机影像 影像拼接 梯度差异 动态规划 

分 类 号：P231[天文地球—摄影测量与遥感]