Research And Design Of The A Column Blind Spot Elimination System Based On TMS320DM642

With the rapid development of automobile, people pay more attention to the car driving safety and comfort. The A-pillar blind spot affects the front view of drivers, may bring potential safety hazard. Therefore, to eliminate the A-pillar blind spot is very important. For this problem, an intelligent system based on DSP (digital signal processor) has been designed, it can completely eliminated the A column blind spot. The main work carried out as follows:1. A-pillar blind spot eliminate system based on TMS320DM642is research and designed, including the hardware platform and system software design. A new concept "electronic transparent" is brought up in this paper, that is use custom LCD screen affixed to the inside of the A-pillar, allows the driver to be able to see the A column blind spot real road images by observing the inside of the A-pillar display. Function of "the image changes as the vision point of the driver moves" is designed, to make sure LCD screen display the road image can fit into the natural vision field of the driver, providing comfortable and natural visual experience.2. Dual-channel video image real-time acquisition and processing based on TMS320DM642, one of the channels to capture driver’s perspective, after DSP processing the images generate a value to control another channel to capture the A-pillar blind spot traffic information. Achieve function of "the image changes as the vision point of the driver moves".3. Research the driver’s perspective positioning and A-pillar blind spot traffic image segmentation display. DSP has a rich resource and strong processing power, to make sure LCD screen display the road image can fit into the natural vision field of the driver.4. Face detection and recognition based on skin color model has been designed. Experimental results demonstrated that using color information to achieve the face recognition and positioning, have such characteristics:computing amount small, fast and strong robustness and practicality. Therefore, select the skin color model for detecting and positioning of the driver’s face.5. Build experiment platform and analysis the experimental results. The methods using the models based on the YCbCr skin color space for face detection and identification are effective, reliable. Further processing of the face detection and recognition image, can obtain the driver’s perspective information. Use driver’s information to control LCD screen display image corresponding to the traffic information, Eliminate A column blind spot Perspective obstacles. Meet the system application requirements.