The Research Of The Real-time Image Processing System Based On ARM+DSP

With its rapid development, modern information technology receives a lot of attention. Relatively, more importance has been attached on researches and practices in image processing—especially in real-time image processing. During recent years, embedded system and DSP technology have been continually developed. They push forward the gradual transformation in the field of digital signal processing, from theoretical results to practical application. Thus new theory and application will be generated. They also produce powerful effects on the technological development of image processing and other relevant areas. It can be seen that a new method is worthy of study, by which ARM plus DSP, the dual core processor structure will be adopted in real-time image processing system, both theoretically and practically.This paper mainly focuses on the following aspects:1. The paper analyzes the latest development of real-time image processing system. It comes to a conclusion that ARM adopted as the core of the system has advantages and disadvantages and so does DSP. To combine real-time with high-efficiency and portability, the author designs a general real-time image processing system cored by ARM+DSP dual-core processor which by adding or cutting can be enormously used in image processing and image recognition.2. After the fingerprint recognition and iris recognition, palmprint recognition technology is the latest research directions in the field of biometric identification, which keeps in continuous research and exploratory stage. The paper introduces a specific example of combining palmprint recognition technology with a general real-time image processing system cored by ARM+DSP dual-core processor. The combination leads to a basic offline palmprint identification system. In the paper, systematic composition and basic running process are given, the most basic identification function is achieved, and the performance of palmprint identification system is enhanced.3. In the specific design, after comparing two kinds of system composition schemes, the TMS320VC5470 of TI (dual cores of ARM+DSP) are used as system core. According to the characteristics of the chip VC5470, to design the hardware principle of the system platform and expand the extended RAM, ROM(Flash): man-machine interface, peripheral circuit, JTAG interface; to choose the components demanded by the hardware principle; to design 8 layers PCB, with emphasis laid on the rules of routing and the signal integrity in high-speed digital circuit and high-speed PCB design which gives a better solution to the high density routing of BGA package and it's EMI problem. Meanwhile, discussing the method of TMS320VC5470 independent system Bootload, communications between two cores and downloading program into FLASH after separated from the host development environment.This paper's innovative work is as follows: by combining ARM with DSP effectively, they will play to their respective advantages and overcome their disadvantages in the course of real-time image processing system. The performance of the real-time image processing system is enhanced, the size of it is reduced, and the cost of it is saved. And based on these studies and research results, this general real-time image processing system cored by ARM+DSP dual-core processor is applied in the latest palmprint recognition principles, which lead to the handheld palmprint recognition system. It is significant to the real-time palmprint recognition technology research.Finally, the paper summarizes the achievement and deficiency in the research and proposes the prospect of the further research and application.