Implementation And Optimization Research Of Algorithms For Image Processing On Vehicle Detection And Tracking Based On BWDSP1042

With the rapid economic development in today’s society, transportation accounted for a pivotal position in people’s daily life. Motor vehicles want to keep the amount of rapid growth. It is unable to solve the existing problems of a lot of traffic relying solely on the construction of roads and artificial management. To solve a series of problems, Intelligent Traffic System (ITS) appears. It has become a hotspot in recent years, and has brought enormous economic benefits to the development of the transport industry. Among them, the vehicle detection and tracking technology plays a very important role in the Intelligent Transportation System, which has become the research focus at home and abroad.A digital signal processor (Digital Signal Processor, DSP), is consists of a large-scale or VLSI chips. In this work, based on "Vehicle Detection and Tracking Technology", the image processing algorithms are studied. Its significance lies in improving the speed and accuracy of image processing. The main experimental platform of this work is BWDSP1042 developed by the thirty-eighth China Electronics Technology Group Institute. It is a high-performance DSP, suitable for the field of radar signal processing, electronic countermeasures, precision-guided, communication security and image processing and so on. The language it uses is assembly language.Meanwhile, this work carried out systematically and summarized for the spatial image enhancement method. Image enhancement technology is one of the most attractive areas in the digital image processing technology. Its purpose is to image processing, so that it is more suitable for a particular application compared to the original image, thereby obtaining better visual effect and more useful images for some particular applications. The methods of image enhancement are divided into two parts: spatial domain methods and frequency domain methods. In image processing, spatial domain image enhancement means enhanced the pixels of constituting the image. The basis of these methods is actually the process of directly dealing with the image pixel.The purpose of this article is to make vehicle detection and tracking imag es run on the C language platform to be implemented on BWDSP1042 and op timize running cycles and running time. In this work, the research content mai nly includes:the input and storage of the image; the computing of the average pixel of image, judgment day or night; background updating process and calc ulation on foreground image and the gradient image; vehicle detection and vehi cle tracking based on gradient image and foreground image. For the current im age processing algorithms, each step above in the actual arithmetic operation pr ocess is carried out repeatedly to debugging and eventually realized optimizatio n scheme on BWDSP1042. At last, implementation and optimization of the alg orithm is evaluated and the series of problems faced is summarized.