Research On Visual Navigation Of Forklift System Based On Feature Tracking

The visual navigation system has many advantages such as low cost, high reliability, high stability and strong situational awareness and so on. Because of these advantages, visual navigation has become the main research direction and development trends in the field of autonomous navigation. In this paper, the forklift visual navigation system takes the straight line of path edge and the straight line of shelf as tracking features. The following discusses the paper content and innovations:According to the actual needs and existing software and hardware conditions of the forklift visual navigation system,at the same time meet the real-time and reliability requirements, design the forklift visual navigation system.The forklift visual navigation system bases on feature tracking method. In image processing process, selects and improves algorithms in order to improve the real-time and stability of the system. Depending on the location of the features in the images, set up different processing ROI area, in order to reduce the amount of data and processing difficulty. In the image sequence, parameter information of previous frame estimate parameter of the next frame, in order to reduce navigation parameter error.According to the forklift work environment, design a straight line tracker. The edge straight line of path and shelf is the main tracking target. Edge line tracking includes establish feature matching collection, describe straight line feature.Considering the system processes, select C++ and OpenCV that suit the development and application of visual navigation as develop tools. OpenCV programmings not only call the high-level function to solve the complex problems of visual navigation, but also call the basic functions to create a system framework.