Research Of Crawler Agriculture Robot Path Recognition And Tracking Based On OpenCV

The research of intelligent agricultural robot is more and more concerned,Vision navigation is the important research direction of intelligent agricultural robot autonomous navigation both in abroad and our country.To recognize the navigation path is the prime problem of the agricultural robot visual navigation.This paper takes the crawler electric tractor as a robot,focusing on the structured road environment,takes the road centerline as the robot’s navigation path to identify and trace,it mainly includes the research contents and conclusions as follows:(1)Based on Matlab plane grid calibration method to calibrate the camera,received internal parameters of the camera,and the calibration error of pixel within the scope of the permit,it shows our calibration result is valid and reliable.(2)Made a series of pretreatment to the structured road static image and dynamic video image which collected by camera,Including select Region of Interest,gray-scale transformation,image filtering,threshold segmentation,edge detection and so on.Compares and analysis the fixed threshold segmentation and adaptive threshold segmentation method,these two methods both can segment the road and background effectively through our experiments.And then compares and analysis Canny,Sobel,Laplace edge detection which are commonly used method,experiments results show that Canny edge detection can get the road edge clearly,it is ready to extract the center line of the road.(3)Used horizontal scanning method to extract the center line of the road,then get the navigation path of road static image and dynamic video image by using Hough transform linear detection method.Used pinhole imaging model,through the Conversion relationship of coordinates,we can obtain the navigation parameters of the world coordinate from the navigation path of computer plane coordinate,it laid a foundation for controlling the robot to track the actual navigation path.(4)Based on the MFC dialog application,made a road static image and a dynamic video image processing software,software test shows that it can get navigation path of road researched in this paper,this visualization of the software interface is convenient to use.(5)In order to make the robot track the navigation path,Based on the analysis of the working principles of electric control hydraulic steering system of crawler robot,determine the control strategy and methods of the navigation path which takes the fuzzy PID controller as the core of robot tracking control,designs the simulation model of control system in SIMULINK environment.Simulation results show that the fuzzy PID controller has the stable ability of path tracking and strong ability of anti-jamming,the control strategy and the method is feasible and effective for our control system.