可变阈值SIFT算法匹配无人机影像摘要.pdf

Geomatics Science and Technology 测绘科学技术, 2018, 6(1), 8-14 Published Online January 2018 in Hans. /journal/gst /10.12677/gst.2018.61002 Variable Threshold SIFT Algorithm for UAV Image Matching Sainan Hua, Xingsheng Deng* School of Traffic and Transportation Engineering, Changsha University of Science Technology, Changsha Hunan th th th Received: Dec. 26 , 2017; accepted: Jan . 9 , 2018; published: Jan . 16 , 2018 Abstract It is difficult for traditional image processing algorithm to deal with the Unmanned Aerial Vehicles (UAV) image, because the UAV have light weight, poor stability and the captured digital images are overlapped irregularly. Aiming at the problems, this paper adopts the SIFT algorithm with an adaptive threshold and expanded detection range, which has the characteristic of scale invariance, rotation invariance. Two methods are adopted to process two pairs of overlapping images. The comparison is made between the number of image feature points extracted and the matching ef- fect. The experimental results show that the SIFT algorithm with variable threshold and expanded detection range has more advantages than the SIFT algorithm as to process the rotating and unst- able UAV images, thus better matching effect is achieved. Keywords Photogrammetry, Digital Image Matching, Adaptive Threshold, SIFT Algorithm 可变阈值SIFT算法匹配无人机影像 华赛男，邓兴升* 长沙理工大学交通运输工程学院，湖南 长沙 收稿日期：2017年12月26 日；录用日期：2018年1月9 日；发布日期：2018年1月16 日 摘 要 针对无人机重量轻、稳定性差，所拍摄的数字影像存在重叠度不规则、旋转尺度大，造成影像匹配处理 *通讯作者。 文章引用: 华赛男, 邓兴升. 可变阈值SIFT 算法匹配无人机影像[J]. 测绘科学技术, 2018, 6(1): 8-14. DOI: 10.12677/gst.2018.61002