Autonomous underwater robot recovery docking foreground view field three-dimensional reconstruction stereo matching method

Abstract

The invention discloses an autonomous underwater robot recovery docking foreground view field three-dimensional reconstruction stereo matching method, which uses a minimum spanning tree algorithm to link all image pixels, so that pixel points which are far away from the point and have large color difference in an image transmit small support quantity, and pixel points which are close to the pointand have small color difference in the image transmit large support quantity. CENCUS change is combined with a tree structure, so that the problem of high matching error rate of a weak texture regionin an underwater picture by a traditional stereo matching algorithm is solved. CENCUS transformation is adopted to distinguish a weak texture region of the underwater image to realize cost calculation; then cost aggregation is performed by adopting a stereo matching algorithm based on a minimum spanning tree, then calculation is carried out by adopting a WTA algorithm to obtain a disparity value with the minimum cost, and finally smooth optimization is performed on disparity by adopting sub-pixel refinement to realize the output of the final disparity map.

Description

technical field [0001] The invention relates to a three-dimensional reconstruction of a foreground field of view for recovery and docking of an autonomous underwater robot, in particular to a stereo matching method for the three-dimensional reconstruction of the foreground field of view for the recovery and docking of an autonomous underwater robot, belonging to the field of robot technology. Background technique [0002] Autonomous Underwater Vehicle (AUV), because it carries energy and a controller that can make decisions on its own, can independently complete preset missions, adapt to the complex and harsh environment of the ocean, and replace the staff for diving operations. To ensure the safety of the staff, and the capital expenditure is small, it is widely used in the detection of marine mineral resources, the detection of submarine oil pipelines, underwater archaeology and other fields. However, due to the limited energy it carries, the voyage time and voyage distanc...

AI Technical Summary

Problems solved by technology

[0010] The purpose of the present invention is to provide an autonomous underwater robot recovery and docking method for three-dimensional reconstruction of the foreground field of view stereo matching method, to provide an improved stereo matching algorithm based on the minimum spanning tree structure, so as to solve the problem of weak underwater image texture, low matching accuracy and matching problems. The problem of slow speed, so as to realize underwater three-dimensional reconstruction, apply to the recovery control process of fully driven autonomous underwater robots, provide the environmental information required for path planning in the recovery process of underwater autonomous robots, and provide necessary information for subsequent path planning and obstacle avoidance. prior knowledge of

Images