A Stereo Matching Method for Extraterrestrial Orbiter

Abstract

The invention relates to a stereo matching method for extraterrestrial object inspectors, and belongs to a stereo vision matching technology. The method comprises the following steps of: obtaining a left image and a right image by a left camera and a right camera of an extraterrestrial object inspector, carrying out polar line correction on the left and right images, and respectively constructingL layers of left pyramids and right pyramids for the polar line corrected left and right images; carrying out bidirectional stereo matching on two images of the Lth layers of the left and right pyramids so as to obtain a left disparity map and a left disparity map; determining a parallax search range of each pixel of the (L-1)th layer by utilizing the disparity maps of the Lth layer, and carryingout bidirectional stereo matching on two images of the (L-1)th layer so as to obtain a left disparity map and a right disparity map of the (L-1)th layer; and circulating the steps until a parallax search range of each pixel of the first layer is determined by utilizing disparities of the second layer, and carrying out bidirectional stereo matching on two images of the first layer so as to obtain aleft disparity map and a right disparity map of the first layer.

Description

technical field [0001] The invention relates to a stereo matching method of an extraterrestrial celestial body patrol device, in particular to a multi-resolution fast stereo matching method of an extraterrestrial celestial body patrol device, and belongs to the technical field of computer vision and deep space detection. Background technique [0002] The detection of extraterrestrial celestial body rovers has become an important part of deep space exploration. Since extraterrestrial celestial bodies have the characteristics of unknown working environment, unstructured, complex terrain, large working range, and far away from the earth, it is necessary to focus on solving a series of problems such as autonomous environment perception and path planning of the rover. The rapid three-dimensional reconstruction of terrain on the surface of extraterrestrial celestial bodies is the key issue to ensure the smooth travel of rover on extraterrestrial celestial bodies. [0003] At pres...

AI Technical Summary

Problems solved by technology

[0005] Although the SGM and MGM algorithms have achieved good results, on the one hand, in the environment of unknown extraterrestrial celestial bodies, since we cannot predict the parallax search range of the scene, if the traditional SGM or MGM algorithm is used, the entire image All pixels of the

However, if the disparity search range is set too large, it will consume more computing time, occupy more memory space, and may cause wrong matching; if the disparity search range is set too small, the correct disparity will not be obtained.

At the same time, the imaging resolution of extraterrestrial celestial rovers is getting higher and higher, which will bring a huge amount of calculation

On the other hand, both SGM and MGM processing algorithms use the same set of penalty parameters P 1 and P 2 , leading to poor matching accuracy in disparity discontinuities and occlusions

Images