Monocular Vision-based Patrol Robot Road Navigation Technology

Visual navigation technology is a part of driving assistance technology.It is a hot research direction.This article mainly aims at the research and development of autonomous navigation technology of industrial robots in industrial parks.The purpose is to develop a monocular visual navigation system that can assist or even replace the park inspection workers to autonomously inspect the park.The key technologies of the monocular visual navigation system are the camera calibration,feature line extraction and road navigation algorithm.Which extraction accuracy,real-time,determines whether the navigation system has engineering applications.Firstly,in the camera calibration technology,the calibration method of the original literature is analyzed.According to the actual navigation environment of the navigation system,a new camera calibration algorithm is designed.The algorithm is to build a mathematical mapping model based on the spatial geometry relationship between the camera,inspection robot and the ground.The data obtained by calibration is compared with the data obtained by actual mapping.The calibration accuracy of the algorithm meets the navigation requirements.Then,the traditional image processing algorithms are analyzed to establish a method for detecting road feature lines through image processing algorithms.Based on the detected characteristic lines,an algorithm for feature line extraction is designed.Through experimental verification,the algorithm can extract a series of points on the road feature line and fit these points out to the road feature line.A similar line.In the experimental test,the extracted road feature line has a very good accuracy and stability.Finally,the road navigation technology of the mobile robot is analyzed.According to the position change relationship between the vehicle body and the lane line,a road is designedbased on the angle of the extracted characteristic line relative to the central axis of the vehicle body and the distance parameter of the vehicle body front identification point.Navigation control algorithm.In the actual navigation experiment,the navigation algorithm can control the inspection robot very well,adjust its own posture relative to the lane line by means of left turn and right turn,and move along the road feature line.