Research And Implementation Of Image Processing Algorithms In Intelligent Visual Navigation System

Intelligent navigation system is the key part of the intelligent transport system, which is applied widely in almost all the industries. Tourism as a prominent industry for Guizhou Province, developing the intelligent visit vehicle will have a promising perspective. The road image tracking system plays an important role in the intelligent navigation system, and how to detect accurately the lane and to track the obstacles will be the key point for the design of this system. Therefore, this paper will focus on the research of these two parts.Based on the main research contents, firstly, this paper introduces three kinds of methods for the image enhancement, including histogram equalization, homomorphism filtering and morphological processing. These three methods are used to enhance the road image and according to the final processing results, their advantages and disadvantages are analyzed respectively. Secondly, several operators for edge detection were illustrated detailed, such as Roberts, Sobel, Prewitt, Laplace, Canny and so on. And the new edge detection methods such as snake model, wavelet transform and morphological operators are also introduced. The edge detection results which are obtained by the above methods are given and compared.Based on the above research, combining the snake model and the Hough transform, the lane detection is realized and after that, using the morphological algorithm, the obstacle is located.The research results show that, Canny Operator is the best one for the edge detection, it can locate accurately the line and keep the good continuity of this line, and it has also the high contrast. The combination of the snake model and the Hough transform can recognize accurately the lane and the middle lines. The morphological algorithm can find the location of some simple obstacles.