Reasearch And Design On Traffic Information Video Detection Device Based On DSP

Transportation information collection system is an essential part of ITS (Intelligent Transportation System), it is the foundation of modern traffic management and control. At present, the traditional loop detector is widely used for traffic information collection, but disadvantages of loops caused by installation and maintenance restrict the detection precision and the detection range for collecting information accurately in real-time. With the rapid development of computer vision, image processing, pattern recognition and AI (Artificial Intelligence), video detection applied in the traffic area becomes a research focus. Video detection can obtain the state of moving objects and parameters of traffic flow for management and control of urban traffic. Compared with traditional information collection method, video detection has advantages in installation charge, detection range, function extending, useful life and so on.In this paper, designed by the traffic information video capture device is based on TI (Texas Instruments USA) TMS320DM642 DSP embedded video processing systems, which the chip is TMS320DM642 digital signal processor that specifical for digital media applications, have rich peripheral such as video interface, I2C director and Ethernet module, this can be achieved with the chip around the seamless connectivity, both to reduce the chip connection between the logic of complexity, and reduce the number of chips using the system, strengthen the system in working under the conditions of high-frequency electromagnetic compatibility. On the other hand, simplifying hardware design offer the possibility that reduce the development cycle, due to the entire hardware system with DSP as the core, the development board in the bottom of drivers operating system and embedded DSP / BIOS time Are easy to achieve, reduce development time. In addition, the video collection system in the video interface circuits used by the video decoder is SAA7121H that supplyed by Philips company, video encoder is TVP5150PBS that supplyed by TI company. Video codec from a standard analog video signals in decoding the digital image data and digital image data encoded as a standard analog video signals. This makes the system can be used with any standard analog video interface of image sensors and image monitors, greatly increasing the flexibility of the platform.In the thesis, the algorithm in the system design, including six parts: image preprocessing, the goal edge detection, the moving target of extraction, multi-target segmentation, target identification and target tracking. In the part of the pretreatment of the use of filtering and Gaussian smoothing algorithm to filter video image de-noising, prepare for the next work have the high-quality, the right to image detection. The target of some edge detection, the paper cited three edge detection operator: Sobel operator, Prewitt operator, Laplacian operator. Through the various operators of the marginal effect of plans, this paper is the Laplacian edge operator. In the part of pick-up moving objectives,we use the difference between frames from the use of the method, but because this is the subject of dealing with a smaller range of motion video, the two commonly used difference are often not enough information campaign, distilling the moving object is incomplete, considering the experiment data , An improved design of the three-difference and be a good test results. In the division of the multi-target, we design eight neighborhood of the border filled segmentation algorithm, the algorithm is based on the regional growth of a segmentation method, simple and practical, real-time system suitable for DSP processing. In the part of target Identification e, we design a arithmeticthat based on the eccentricity of the target image vector feature extraction algorithms, as the feature extraction algorithm has good stability, adaptability and steadiness characteristics, in different scenes of traffic can be better Applications. In the tracking of the target part,the paper use Kalman the forecast objective of tracking algorithm, the algorithm can make full use of Kalman filter function of the forecast to predict the next frame may be emerging regional goals, and then in smaller regional forecast in the relevant matches computing, then find the best matching point, making the objective of tracking more proactive. Built the hardware development platform, design the necessary software algorithms.The next work is needing to put the design of algorithms migrate to hardware systems. Software integrated development environment CCS provides a platform for achieving this process, in a familiar method of transplantation process, we can easily facilitate the hardware in the system to see algorithm results. The paper research provides the theory and algorithm for the intelligent transportation information collection system, it also supports the management and control of mixture traffic on technique. The research findings of the dissertation have certain theory significance and practical value to impel ITS modernization advancement of our country.