Real-Time UAV Imagery Stitching Based on Grid-Based Motion Statistics

Abstract

In order to receive and stitch sequential images from Unmanned Aerial Vehicle (UAV) synchronously, an improved approach based on feature points matching is proposed. Firstly, global images with overlapping regions are described by Oriented FAST and Rotated BRIEF (ORB) feature. Then Grid-based Motion Statistics (GMS) method is employed to obtain robust feature correspondence from primary matching pairs, these matching pairs are further validated by neighborhood support and used to get transformation matrix. The performance of the proposed algorithm is demonstrated through computer simulated experiments. Experimental results show that the improved method can efficiently solve the problem of smaller-overlapping and less-textured images with real time capability.