Path Recognition And Obstacle Avoidance Navigation System Based On Machine Vision

Labor intensity of traditional agricultural practices is considerable. Being exposed in the environment of high temperature and high humidity will do harm to human health. At the same time, in the process of agricultural practices, there may be some toxic and harmful substances for humans. Therefore, intelligent agricultural machinery and equipment are of great importance. In order to reduce the labor intensity of personnel in agricultural practices, we take the high-clearance car which is independently developed by us as a platform and design the system of path identification and obstacle-avoiding navigation for facility agriculture in view of the problems of complex environment, large number of obstacles, narrow paths and so on. Finally, the stable and efficient control of obstacle-avoiding and navigation for the car is realized in the operational environment of the facility agriculture.In this paper, a system of path identification and obstacle-avoiding navigation based on facility agriculture is designed.In terms of the system of path identification and obstacle-avoiding navigation, this paper proposes an image processing algorithm for path detection, including using of the binarization with filters to identify the black and white paths, using RGB color space to identify colored paths, using 2GRB color space to identify green plants, using central line method to extract paths, using Hough transferring of single region to extract paths, using Hough transferring of double region to extract paths, using Hough transferring of multiple regions to extract paths so as to realize the autonomous navigation of the car. The system identifies and makes navigational controls of a series of paths, such as the identification lines of different colors, the dividing line between green plants and bare ground and so on in order to realize the path identification and navigational control in a variety of environments.In terms of the stereo visual obstacle-avoiding system, two kinds of sensors for extracting disparity map are introduced in this paper. The first one is Xtion sensor based on infrared distance measurement; the second one is a binocular camera; then, it introduces in detail the concrete implementation of the binocular calibration, binocular correction and stereo matching of binocular vision system; secondly, this paper presents a new interpolation algorithm which is specially applied to the system and uses two kinds of sensors to reconstruct 3D for objects; a method for controlling the car by hand gesture is applied; In this paper, the three-dimensional shape of obstacles in facility agriculture and the its distance with the car are obtained according to the extracted disparity map so as to design the optimal route for the car. In this way, the car can quickly and accurately identify and avoid obstacles; drive along the path to the specified location; finally, this paper also carry out related experiments-path identification, distance measurement of stereo vision, obstacle avoidance and so on to verify the rationality of the algorithm.The experimental results show that the developed path identification and obstacleavoiding navigational system can identify and navigate along a series of paths, such as the marking lines of different colors, the dividing lines between green plants and bare ground, and so on so as to accurately measure the three-dimensional shape of obstacles and avoid them. The system has high stability, good real-time performance and strong antiinterference ability. It is able to meet the requirements of practice of unmanned highclearance car in facility agriculture. What is more, it can also provide reference for the design of navigational control system of high-clearance car in facility agriculture.