Research Of Electronic Image Stabilization Theory And DSP Realization

With the progress of the technology, the video image processing technologyand the large scale integrated circuits are developed very fast, the applications of theelectronic image stabilization technology are increasingly widespread in the militaryfield and civil field. As the new image stabilization method, Electronic imagestabilization makes up the disadvantages of optical image stabilization method andmechanical image stabilization method. the motion estimation algorithm isparticularly important to the electronic image stabilization method, it determines thesystem’s speed and accuracy, effects the real-time performance of the system. Tomeets the performance requirements of the photoelectric radar, this paper hasimproved the typical gray projection algorithm, and achieved the real-timeelectronic image stabilization based on DSP technology.First, this paper introduces the basic principle and application of the commonimage stabilization methods, analysis of the reasons for the fuzzy video imagesequence, introduces the basic principles of electronic image stabilization and theevaluation criteria of electronic image stabilization system’s performance,investigates the basic principles of a variety of typical electronic image stabilizationalgorithm, analysis the advantages and disadvantages of them.Second, this paper improved the old gray projection algorithm to meet theperformance requirements of the photoelectric radar’s terminal video display system,first of all, preprocess the image, use the histogram equalization method to enhancethe contrast of the image, thereby enhancing the image stabilization accuracy.According to the characteristics of the gray projection curve, use the fast searchingmethod to develop the speed of the electronic image stabilization, then use themedian filtering method for the estimate of the vector of random mechanicalvibrations of the scanning motion camera. Finally use MATLAB to simulate theimproved algorithm, analysis and evaluate the effect of different electronic imagestabilization by the method of minus shadow, it proves that the improved grayprojection method is more accurate and faster than the typical gray projectionmethod.In the final part of this paper, use the video image processing system based onTMS320C6416chip, program with CCS to embed the improved gray projectionalgorithm into DSP, achieve Off-line real-time electronic image stabilization.Through playing the unprocessed and processed videos on the same display, itproves that the video which is processed by electronic image stabilization is clearerand more stable, and it has a good real-time performance. It meets the performancerequirements of the photoelectric radar.