Research And Implementation Of Electronic Image Stabilization System Based On DSP

With the rapid development of intelligent system technology, a variety of researches and developments of video system in military and civilian fields has been widely concerned. Just because the electronic image stabilization technology is widely applied, it has also become an important research topic. The TMS320DM642 chip made by TI is recognized as the good performance of video processors, which is a specially designed for multimedia applications with 32-bit fixed-point DSP chip, whose good processing performance and on-chip peripherals can meet the requirements of real-time video processing.Real-time electronic image stabilization refers to stabilizing image sequences timely by using electronic image stabilization methods and making the camera system and tracking system operator not be subject to interference from the camera's movement to achieve real-time tracking and stability of the purpose of video images.In this paper,for real-time electronic image stabilization, as well as the unique characteristics of TMS320DM642, the use of DSP technology to achieve electronic image stabilization method is proposed.In this paper, electronic image stabilization algorithm and the basic principles of the process are elaborated. Feature point tracking algorithm based on Harris corner detection is mainly elaborated, which is transplanted and optimized on the DM642, then accomplished the real-time process of electronic image stabilization. First of all, the principle and mathematical model of electronic image stabilization are explicitly analyzed, the realization process of several classic methods are introduced, at the same time, the advantages and disadvantages of each methods are analyzed. Then this article introduced image stabilization algorithm——Feature point tracking algorithm based on Harris corner detection,which is firstly using the Harris detection algorithm to extract the image feature points, to match the feature points and estimate movement parameters, to use the motion vector obtained to compensate motion vector of image, and obtain the stable video sequences. Then this article combined with the characteristics of DSP to introduce the CPU core, the structure of on-chip memory and the external device interfaces of the TMS320DM642 chip and the instruction set of the DM642 and pipeline structure. It also summarized the development tools of DSP software, the development tool CCS and Software development process. Finally, with the combination of DM642 embedded hardware features, the DSP of the transplant process and method of PC-based electronic image stabilization is discussed. And then combining with the feature of electronic image stabilization algorithm, this article optimized from the level code and compilation level and memory from memory allocation and scheduling, cache optimization, EDMA, etc with the combination of the features of the DM642. Optimized to deal with increased speed, and achieved good results. After optimization, the speed increased and achieved good results.