针对高分辨率遥感图像的融合拼接算法

Fusion and stitching algorithm for high-resolution remote sensing image

为给观测人员提供视角宽广、内容丰富、真实客观的全景图像,通过对无人机航拍相邻图像间的位置关系的研究结合改进的特征检测与描述算法,提出一种针对高分辨率遥感图像的特征查找与匹配算法,同时提出一种基于最佳缝合线的多分辨率融合算法,制定多幅高分辨率遥感图像的分层拼接策略。在实验阶段,特征点匹配数目相比传统SURF算法平均减少44%,运行时间减少37%,相比文献[10]算法匹配数目平均减少26%,运行时间减少22%,结合最终给出的50幅高分辨率遥感图像拼接结果,说明了优化算法在多幅高分辨率遥感图像拼接中的有效性,为进一步研究打下坚实基础。

To provide observers a wide-angle,rich,realistic and objective panoramic image,a high-resolution remote sensing algorithm was proposed by combining the research on the positional relationship between adjacent images of UAV aerial photography and the improved feature detection and description algorithm.The feature searching and matching algorithm of image was presented.At the same time,fusion and stitching algorithm for high-resolution remote sensing image was proposed,and a layered mosaic strategy of multiple high-resolution remote sensing images was developed.In the experimental stage,the number of feature point matching is reduced by 44%and the running time is reduced by 37%compared with the traditional SURF algorithm.Compared with the number of algorithms in the literature[10],the average number is reduced by 26%,and the running time is reduced by 22%.The results of the mosaic of 50 high-resolution remote sensing images fully demonstrate the effectiveness of the optimization algorithm in the mosaic of multiple high-resolution remote sensing images,laying a solid foundation for further research.