基于ORB的无人机图像快速拼接算法  

Fast Image Mosaic of UAV Based on ORB Algorithm

作　　者：王慎谦 张荣国[1] 李晓波[1] 王晓[1] 王芳[1] WANG Shen-qian;ZHANG Rong-guo;LI Xiao-bo;WANG Xiao;WANG Fang(School of Computer Science and Technology,Taiyuan University of Science and Technology,Taiyuan 030024,China)

机构地区：[1]太原科技大学计算机与技术学院,太原030024

出　　处：《太原科技大学学报》2021年第5期374-379,共6页Journal of Taiyuan University of Science and Technology

基　　金：国家自然基金(51375132);山西省自然科学基金(201801D121134)。

摘　　要：为了更好满足无人机航拍图像拼接对实时性和稳定性的要求,提出一种四叉树局部熵自适应阈值的ORB(Oriented FAST and Rotated BRIEF)算法,首先对图像划分网格,通过计算每个网格内局部熵的最佳阈值提取FAST特征点,然后采用四叉树对提取的特征点进行最优筛选,最后采用KNN算法对特征点进行粗匹配,使用PROSAC算法对异常点的剔除。实验结果表明,在特征提取时间基本相同的情况下,本文算法比ORB算法匹配率提高3.8%.In order to better meet the real-time and stability requirements of UAV aerial image stitching,an improved ORB(Oriented Fast and Rotated BRIEF)algorithm based on local entropy adaptive threshold of quadtree is proposed.Firstly,the image is meshed,and FAST feature points are extracted by calculating the best threshold of local entropy in each grid.Then the quadtree is used to optimize the extracted feature points.Finally,KNN algorithm is used for rough matching of feature points.PROSAC algorithm is used to eliminate outliers.The experimental results show that the matching rate of the proposed algorithm is 3.8%higher than that of the ORB algorithm when the feature extraction time is basically the same.

关 键 词：ORB 四叉树 局部熵 PROSAC 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]