基于TMS320DM642和小波分析的红外图像压缩系统分析-控制理论与控制工程专业论文.docxVIP

ABSTRACT Aiming at the transfer and memorizer problem of mass image, this thesis designs a video system of compressed image based on DSP, and the system is used in traffic accidents of vehicle. The system is mainly based on the features that DSP is handy, cheap, stable and the developed EZW arithmetic is fit for the hardware. Digital signal processor, due to strong operation capability and its special soft and hard structure, has been widely used in image compressing. After studying the technologies and theory of DSP deeply, this thesis presents the selected principles of DSP. Through careful analysis and compare, we choose the DSP and other sections design, and then also introduce functions of each part. It expatiates the features of DSP and its designed process. Firstly, the video system of compressed image is presented that the TMS320DM642 works as the central processor to compress the images, the video input processor performs image acquisition and conversion, the CPLD is used for finishing scheduling, the interface of USB transmits the compressed images. Then circuit schematics and print circuit boards are designed. The acquired images of the system are stored in the SDRAM. If one frame is finished, an interrupt signal will be sent to DSP. Then DSP reads these data, compresses them rapidly using the developed EZW arithmetic and the compressed images are stored in FLASH. At last, the compressed images are sent to PC through the interface of USB. The system is small, low cost and credible. Then, introducing the wavelet theory and EZW arithmetic, the developed EZW arithmetic which makes use of the hardware of TMS320DM642 fully and makes the system become a real-time system is presented. At last, the system is applied to record the compressed image of traffic accidents of vehicle. The result of decoding image and coding speed is given. We test the relative quality of the system, which obtains the desired properties. Keywords：DSP, Infrared image compression, EZW ar