Research And Implementation Of Suspicious Vehicle Recognition Algorithm Based On DSP

In recent years, with the rapid development of the national economy, the amount of car’s use is also climbing, followed by more traffic problems and rising crime rates. In this context, the concept of intelligent transportation systems comes into being. It uses advanced technology, such as electronic technology, information technology, sensor technology and system engineering technology, to ease traffic congestion, protect traffic safety, and accurately detect suspicious vehicles. The better solution is to use video to detect vehicles, and the image signal can be received by the video and processed in a real-time. So, this can get more useful traffic information. Through video image processing method, this can not only save manpower and resource, but also more accurately and effectively detect suspicious vehicles.After reading lots of references at home and abroad and having a basic grasp of domestic and international research status, this article studies the suspicious vehicle recognition algorithm, and designs an algorithm, which is for dynamic and static images’ segmentation, position and matching, to be programmed in the DSP board. The major work completed in this article is as follows:In the research of algorithm, first of all, the RGB color space, HSV color space, Lab color space and YUV color space are used for color segmentation and the positioning of the target image; then, the algorithm of aspect ratio of object minimum bounding rectangle, the algorithm of histogram Euclidean distance and the algorithm of SIFT are used to match suspicious vehicles, Finally, the appropriate suspicious vehicle recognition algorithm is determined on DSP board.In terms of hardware implementation, firstly, according to the needs of the hardware system, the YUV color space is used to segment images; then, the algorithm of connected domain is used to precisely position the segmented images; finally, the feasible algorithm, which is determined among the algorithm research, is used to implement suspicious vehicle matching. The algorithm is completed in CCS integrated development environment, using the C language and in the ICETEK-DM6437-B evaluation board. The core processor is the chip of TMS320DM6437.Though testing, this system can effectively achieve the suspicious vehicle detection of dynamic and static image, have a certain flexibility and reliability, occupies a less resource, have a fast rate, and have a extensive application prospect.