Vision obstacle avoiding method and device

Abstract

The invention relates to the technical field of computer vision and specifically relates to a vision obstacle avoiding method and device. The method comprises the following steps: a first image and a second image of an obstacle object in a preset area are collected in real time; the first image and the corresponding second image are subjected to vision analyzing and processing operation, a processed depth image is obtained and is then subjected to automatic focusing optimization operation, an optimized depth image is obtained, obstacle information corresponding to the obstacle object in the optimized depth image is determined, and therefore the obstacle can be avoided according to the obstacle information; the depth image which is obtained via binocular vision calculation is subjected to optimization operation through automatic focusing, the depth image which is generated is high in accuracy, and excellent obstacle avoiding performance can be enabled according to the obstacle information corresponding to the depth image.

Description

technical field [0001] The present invention relates to the technical field of computer vision, in particular to a visual obstacle avoidance method and device. Background technique [0002] Binocular vision is an important branch of computer vision. Binocular vision can imitate the process of human eyes and human stereo vision perception. It is one of the core topics of computer vision research. In recent years, binocular vision technology has been widely used in obstacle detection, industrial automation production, intelligent security systems and other fields. [0003] With the rapid development of information technologies such as drone technology, autonomous driving technology, and robotics technology, the flying environment of drones, the driving environment of cars, and the surrounding environment of robots are all uncertain and complex, requiring constant Detecting the environment and avoiding collisions with obstacles will undoubtedly accelerate the further developme...

AI Technical Summary

Problems solved by technology

[0005] In the research, the inventors found that in the specific application environment of binocular vision in the prior art, especially under complex background and lighting conditions, the matching success rate of point pairs is low due to the existence of correlated noise interference, and the generated The accuracy of the depth information is low, resulting in poor obstacle avoidance performance

Images