基于ORB特征的无人机遥感图像拼接改进算法

Improved image stitching algorithm based on ORB features by UAV remote sensing

针对传统的SIFT算法运行速度较慢、不适合处理实时性要求高的无人机遥感图像的缺点,提出了一种基于ORB特征的快速遥感图像拼接改进算法。首先通过ORB算法快速得到特征点和特征描述,采用K最近邻算法(KNN)进行粗匹配,然后采用随机抽样一致性算法(RANSAC)进行精匹配,最后使用改进的加权平均方法对图像进行融合拼接。实验结果表明,该算法在保证匹配精度的基础上,处理速度较经典的SIFT算法提高了41倍。在图像融合时,该算法能有效地消除拼接重影错位现象。

Aiming to serve low process speed problems of traditional SIFT algorithms that unfitness in real-time-application for UAV remote sensing images, an improved remote sensing image stitching algorithm based on ORB feature points detecting is proposed. Firstly, the system quickly extracts the feature points and descriptions by ORB algorithm, and adopts KNN algorithm to rough match these points. Then it uses the Random Sample Consensus algorithm(RANSAC)to remove the fault matching and to realize fine match. Finally, it uses the improved weighted average method for image fusion and stitching. Experimental results show that the proceed speed of the algorithm is 41 times faster than the method based on SIFT algorithm. In image fusion, the algorithm can effectively eliminate the double dislocation.