Image matching method, image matching device and unmanned equipment

Abstract

The invention provides an image matching method, an image matching device and unmanned equipment. The image matching method comprises the following steps: acquiring a first image acquired by unmanned equipment at a first shooting position on a first track; acquiring a second image collected by the unmanned device at a second shooting position on the second track, and the second shooting position is determined by the first shooting position and absolute position information of the unmanned device; selecting one of the first image and the second image as a to-be-matched image and the other as a matched image; and performing feature matching on the first feature point on the first matching area of the to-be-matched image and the second feature point on the matching image, and the first matching area is a first overlapping area, which is overlapped with the matching image, on the to-be-matched image. According to the method of the invention, the feature points needing to be matched in the to-be-matched image in the feature matching process can be reduced, and meanwhile, the feature matching speed and accuracy are improved.

Description

technical field [0001] The present application relates to the field of unmanned technology, in particular to an image matching method, a matching device, and unmanned equipment. Background technique [0002] At present, image matching methods usually extract scale invariant feature transformation (Scale Invariant Feature Transform, SIFT) feature points from the image to be matched and the adjacent images of the image to be matched, and when matching the image to be matched, the image to be matched will be All SIFT feature points on the image are subjected to feature matching, and for any SIFT feature point on the image to be matched, the best matching point is searched among all SIFT feature points in adjacent images of the image to be matched. [0003] However, this method of performing feature matching on all SIFT feature points on the image to be matched, and searching for the best matching point among all SIFT feature points in the adjacent images of the image to be matc...

AI Technical Summary

Problems solved by technology

[0003]However, this method performs feature matching on all SIFT feature points on the image to be matched, and searches for all SIFT feature points in the adjacent images of the image to be matched The best matching point method has a large search range and takes a long time, which seriously affects the efficiency of real-time 3D reconstruction of unmanned equipment after feature matching

And for repetitive texture scenes such as agriculture, buildings, etc., the probability of mismatching caused by this method is relatively high, which is not conducive to the accuracy of subsequent 3D reconstruction

Images