A binocular stereo ranging method for bionic underwater robot fish

Abstract

The invention discloses a binocular three-dimensional ranging method for a bionic underwater robotic fish. The matching algorithm of SGBM can not only obtain more effective depth points, but also be superior to the algorithm of SIFT feature extraction and matching in terms of calculation speed; at the same time, in the water In the case of severe light attenuation in the lower scene, the image binocular distance measurement with higher precision in the short distance is used, combined with the long-distance sonar detection method; it not only makes up for the lack of interference of the sonar short-distance flow field, but also reduces the lack of light received by the long-distance binoculars to achieve high-precision underwater obstacle avoidance detection.

Description

technical field [0001] The invention relates to the technical field related to robot vision detection, in particular to a binocular three-dimensional ranging method for a bionic underwater robot fish. Background technique [0002] The bionic robot fish is used to complete underwater detection tasks. Sonar is generally used to detect obstacles. However, in the case of severe light attenuation in underwater scenes, sonar short-distance flow field interference is insufficient, and timely and effective judgments cannot be made; secondly, , there are also bionic robot fish that use binocular machine vision to judge obstacles in front of them during underwater movement, while the existing binocular ranging process uses methods such as SIFT feature extraction and matching, not only the obtained effective depth points are relatively less, and the complexity of the algorithm is high, which cannot meet the real-time performance of the main body in judging and avoiding obstacles during...

AI Technical Summary

Problems solved by technology

[0002] The bionic robot fish is used to complete underwater detection tasks. Sonar is generally used to detect obstacles. However, in the case of severe light attenuation in underwater scenes, sonar short-distance flow field interference is insufficient, and timely and effective judgments cannot be made; secondly, , there are also bionic robot fish that use binocular machine vision to judge obstacles in front of them during underwater movement, while the existing binocular ranging process uses methods such as SIFT feature extraction and matching, not only the obtained effective depth points are relatively less, and the complexity of the algorithm is high, which cannot meet the real-time performance of the main body in judging and avoiding obstacles during the movement process.

Images