Design of Robot Obstacle Recognition Location Ranging Algorithm Based on Computer Vision and Artificial Intelligence

摘要

Obstacle recognition in robot vision is closely related to the distribution and shape of obstacles in terrain environment. How to accurately identify obstacles in terrain environment in real time is the key to whether the robot can successfully pass through complex terrain. The traditional detection and maintenance methods of transmission lines require line workers to work in the field and high-pressure environment for a long time, which is labor-intensive and dangerous. In order to solve the shortcomings of the current positioning methods, this paper puts forward an obstacle positioning method based on computer vision technology, and optimizes the path according to the distance between the obstacle and the target area by using artificial potential energy field method to realize the autonomous movement of the inspection robot. In order to verify the feasibility of obstacle detection and obstacle avoidance scheme, and simulate the working environment and autonomous movement process of inspection robot, the obstacle avoidance behavior of inspection robot was tested. The experimental results show that the improved CNN robot obstacle recognition model is better than the artificial fish swarm algorithm (AFSA) in both accuracy and efficiency. Compared with traditional SVM and PSO, the obstacle recognition error of this method is obviously lower, which can satisfy the fast and accurate navigation of mobile robots.