Intelligent Vehicle Vision Systems Research Based On TMS320DM642

With the rapid development of computer technology and artificial intelligence, the unmanned vehicle technology has become widely studied in the field, and achieved good results, because the unmanned vehicle can effectively reduce the appearance of many problems of urban transport. As the eyes of unmanned vehicle vision systems, mainly in the camera-based, with a variety of sensor devices, intelligent information acquisition and processing environment around the car, then the path planning and control decisions based on the information in these environments, and finally intelligent autonomous vehicle navigation. Therefore, Vision Research is the key to the intelligent vehicle to achieve safe driving.Through the study of the background and significance of intelligent vehicle profiles vision research systems and unmanned vehicle vision systems analysis, leads to the importance of intelligent vehicle systems vision research. In this paper, based on DSP chips specifically for image/video development TMS320DM642 intelligent vehicle vision system, including the hardware environment introduces intelligent vehicle vision systems, image preprocessing methods lane edge detection and extraction, software environment and experimental verification and other parts, as follows:First, build an intelligent vehicle vision system hardware platform.In this paper, intelligent vehicle vision system platform constructed mainly a modified electric ATV for the body, using the DSP chip specifically for image/video development of the visual system TMS320DM642 as the main chip, which mainly monocular CCD camera-based ultrasonic, infrared sensors and other equipment, supplemented, building intelligent vehicle system hardware platform vision research, also describes the selection and installation of the camera.Second, the road image preprocessing algorithm to achieve.Image preprocessing procedure is the right features to extract road infrastructure, road image preprocessing methods include a gray processing of image, image of roads filtering, image of road enhancement and image of road segmentation. During pretreatment road image, first analyzes the reasons pretreatment, describes the more common image preprocessing algorithms work, and then be processed according to the actual characteristics of the road environment, the actual experimental results of the common pre-processing algorithm Compare and improved, to determine the appropriate paper image preprocessing algorithm for intelligent vehicles.Third, the intelligent vehicle’s lane edge detection and extraction algorithms and intelligent vehicle steering control method based on the extraction of lane.Intelligent vehicle lane detection and extraction is the key to intelligent vehicle autonomous navigation. This article first road image acquisition into blocks, remove excess interference-road information. Secondly, the use of edge detection method detects the lane and then extracted and fit, and finally according to the fitting out of the lane, intelligent vehicle steering control. Because steer motor use a stepping motor, so incremental PID control algorithm combined with slope deviation method to achieve intelligent vehicle steering control.Fourth, the experimental verification of intelligent vehicle visual presentation software environment and system processes as well as various algorithms.First describes the structure and characteristics of TMS320DM642 DSP chip software development environment -CCS, followed by intelligent vehicle vision systems development process, which describes the process of the entire visual system, and then on system initialization, image acquisition and processing procedures were introduced. Finally, experimental verification of various treatment algorithms employed herein practical effect, to ensure safe and accurate operation of the intelligent vehicle.Through intelligent vehicle run in the real campus environment, proof that the variety of correlation algorithm proposed in this paper about intelligent vehicle vision systems can make intelligent vehicle accurate and safe operation in the real campus environment.