摘要:
随着无人机(unmanned aerial vehicle,UAV)平台技术的发展,越来越多的应用行业和研究领域开始使用UAV影像数据.不同于现有的摄影测量结合像控点的UAV影像匹配方法,提出一种新的UAV影像匹配方法.该方法采用彩色尺度不变特征转换(color scale-invariant feature transform,CSIFT)算法,利用彩色信息的空间不变特性提取基准影像与待匹配影像的特征匹配点对;并采用单应性矩阵与随机抽样一致性(random sample consensus,RANSAC)算法对匹配结果进行提纯,得到最终匹配结果.仿真实验表明,该方法可在保证实验过程鲁棒性的同时,与传统的尺度不变特征转换(color scale-invariant feature transform,SIFr)方法相比,将匹配准确率从70%提高到了88%,而且大大减少了特征点对的数量,缩短了处理时间,提高了UAV影像匹配效率.
Abstract:
With the development of the unmanned aerial vehicle (UAV) technique,currently UAV data have been gradually used in application and research fields.Differing from the current UAV matching method that integrates Photogrammetry and Image control points,the novel method presented in this paper is based on the color scale-invariant feature transform (CSIFT) feature to collect the key match points in the reference image and the image to be matched,followed by using the random sample consensus algorithm (RANSAC) method to extract the match points for the final matching result.The authors use an example to verify the feasibility and the validity of the method.Compared with the SIFI method,the accuracy could be increased from 70% to 88%,and the algorithm proposed in this paper can not only guarantee the matching result but also produce less matching points and use less working time.
下载: