基于改进SIFT算法的山谷地形影像匹配

针对山谷地形的低空影像中地貌单一且特征不易提取的问题,本文对SIFT算法改进,融合Harris特征提取算法优势,得到一种可用于山谷地形下低空无人机影像特征提取与匹配的算法。算法首先利用Harris算法和SIFT算法分别提取特征点,对两种算法提取的特征点进行合并,然后运用SIFT算法对合并后的特征点进行描述,再利用特征点特征向量的欧氏距离进行粗匹配,最后利用RANSAC算法进行精匹配来提高匹配精度。为了验证该算法的有效性,选用一组山地影像数据进行实验并与SIFT算法进行比较,结果表明:算法有效地提升了山谷地形影像上特征点匹配精度。

Valley Terrain Image Matching Based on Improved SIFT Algorithm

Aiming at the problem of single landform and difficult to extract features in low-altitude images of valley terrain,this paper improves the SIFT algorithm and combines the advantages of Harris feature extraction algorithm to obtain an algorithm that can be used for feature extraction and matching of low-altitude UAV images under valley terrain..The algorithm first uses Harris and SIFT algorithms to extract feature points respectively,combines the feature points extracted by the two algorithms,and then uses SIFT algorithm to describe the combined feature points,and then uses the Euclidean distance of the feature point feature vectors for rough matching.Finally,the RANSAC algorithm is used for precise matching to improve the matching accuracy.In order to verify the effectiveness of the algorithm,one sets of mountain image data were selected for experiments and compared with the SIFT algorithm.The results show that the algorithm effectively improves the accuracy of feature point matching on valley terrain images.