High-precision image splicing method under multi-vehicle cooperative constraint

Abstract

The invention discloses a high-precision image splicing method under multi-vehicle cooperative constraint. A multi-scale multi-view-angle image splicing model is constructed based on the spatial pose information of multiple unmanned aerial vehicles from the perspective of multi-unmanned-aerial-vehicle cooperative control; for the phenomena of ghosting, pixel breakage and the like of a single homography matrix, a self-adaptive homography matrix method is provided to improve the splicing effect; fr the problems that the same overlapping area is fuzzy and colors are inconsistent after splicing, a weighted smoothing algorithm is adopted, smooth transition of images of the overlapping part is achieved through weight distribution, and the color difference problem near splicing overlapping is effectively solved. Algorithm performance verification is carried out by using a self-contained aerial photography data set, and experimental results show that the method provided by the invention has good splicing performance, the registration precision is obviously improved, and the requirements of high real-time performance and high precision of multi-scale multi-view aerial photography image splicing in a multi-aircraft cooperative patrol flight process are met.

Description

technical field [0001] The invention relates to the technical field of image data processing, in particular to a high-precision image splicing method under the constraint of multi-machine cooperation. Background technique [0002] Image stitching is the process of combining multiple images of the same scene with overlapping areas into a single wide-field, high-resolution complete image. Currently commonly used image registration methods include image grayscale-based, transform domain-based and feature-based methods. Among them, feature-based matching methods are widely used, including basic HOG (Histogram of Oriented Gradients, histogram of directional gradient) features, SIFT (Scale-Invariant Feature Transform, scale-invariant feature transformation) algorithm, SURF (Speeded Up Robust Features, accelerated Robust feature) algorithm and ORB (Oriented FAST and Rotated BRIEF, fast feature point extraction and description algorithm) and other improved algorithms. The HOG feat...

AI Technical Summary

Problems solved by technology

[0005] In view of the above existing problems, the present invention provides a high-precision image stitching method under the constraints of multi-device collaboration, which can solve the problem of low splicing efficiency and poor quality caused by the large amount of data and complex types in the process of multi-UAV cooperative inspection

Images