| Since the 21st century, the science and technology, especially computer and electronics industry, has developed rapidly, which enhances the computer possessing power and image processing technology. Furthermore, the development of visual navigation technology based on image possessing is also promoted greatly. The key of visual navigation based on path recognition robot is to walk in accordance with the planned navigation line, detect obstacles on a real-time basis and avoid them in the process of walking.This paper is designed according to the visual navigation system of path-based recognition robot and gives a preliminary study of mobile robot path recognition and tracking based on computer vision. It includes:Firstly, the brief review is made on the research background of robot visual navigation system, the development and the status quo of the study area. The image possessing technology based on path recognition robot visual navigation is analyzed in this paper by neighborhood averaging, canny operator, histogram peak and valley method in order to deal with the camera image in denoising, edge extraction, and threshold value selection. Secondly, Hough transform theory is applied for ground path image to detect the navigation line, make the path tracking and realize the robots' avoidance of obstacles through obstacle avoidance path planning algorithm. In view of the technology of DirectShow, the robot visual navigation system video image acquisition is completed, as well as the realization of entire navigation system functions and methods.This paper aims to realize the function of the system in the development environment on account of DirectShow technology, OpenCV visual function library and Net programming. The experimental results show that the system can identify the navigation line, track the path, and detect the obstacles on a real-time basis during walking and then bypass obstacles, completing the visual navigation system based on path recognition. |
PDF Full Text Request