The Research Of Barriers-Detection System Based On Image Processing

In recent years, with the development of computer and robot technology, the research of intelligent vehicles have achieved great accomplishments in limiting traffic accidents, vehicle safeguarding and fast running, etc. Intelligent vehicle technology is applied not only in transportation but also in scientific exploration, industry and military.Vision Accessorial System is important to intelligent vehicles, which can identify and cognize 3D world with images. Using CCD video camera as the visual sensor is the mainstream in the vision system of intelligent vehicles. Barriers-Detecting System based on image processing is to detect the barrier signals with the basic theory of image processing.Firstly, a summary of intelligent vehicle technology is presented; secondly, this article presents also the construction of hardware and software building environment, and a platform of vision system is constructed in order to provide original information for image processing and image recognition; thirdly, a method of neural network calibration based on flat template is raised and its principle is presented. Experiment shows that this method is easy for coding comparing with traditional calibration methods; fourthly, an appropriate image pretreatment algorithm is given, which is histogram equalization arithmetic and median filter algorithm. With the analysis of traditional edge detection operators and theory of morphology, an operator of noise immunity using multi-structure element is presented; lastly, the flow chart of barriers-detection is designed with the method of image pretreatment algorithm and mathematical morphology. Theory analysis and experiment show the effectiveness of this method.