多匹配策略融合的无人机影像匹配方法

UAV image matching method based on multiple matching strategy fusion

针对AKAZE算法在无人机影像匹配过程中存在的匹配精度低和稳定性较差问题,本文提出一种基于多匹配策略融合的改进影像匹配方法。该方法首先对影像降采样并利用AKAZE算法检测多尺度特征。然后采用一种稳定的RootSIFT描述符进行特征描述。其次,融合最近邻距离比值、双向匹配和余弦相似度约束匹配策略进行特征匹配以降低误匹配率。最后,采用随机抽样一致性(RANSAC)算法确定最终的特征对应关系,并求得几何变换模型。实验结果表明,该方法在获得更多正确匹配点对的同时具有较高的匹配正确率和精度,能够更好适用于无人机影像匹配。

To solve the low matching accuracy and low stability problems of AKAZE algorithm in UAV image matching process,an image matching method based on multiple matching strategy fusion is proposed.Firstly,images are decreased resolution and the AKAZE algorithm is used to detect multi-scale feature point.Secondly,a stable RootSIFT descriptor is used to feature description.Then,multiple matching fusion strategy included the ratio of the distance,a bidirectional matching and cosine similarity constraint matching measure are merged to reduce the mismatch rate.Finally,Random Sampling Consistency(RANSAC)algorithm is used to determine the final feature correspondence,and the geometric transformation model is obtained.The experimental results show that this method has higher matching correct rate and accuracy while obtaining more correct matching pairs.So is more suitable for UAV image matching in different scenes.