Research On Key Technologies In Electronic Image Stabilizing Algorithm Based On Gray-scale Projection

With the rapid development of electronics, information, communications, computers and other high-tech, Electronic Image Stabilization (EIS) technology as the latest generation of image stabilization methods have been widely used in daily life. Range from military to civilian, the convenience of EIS technology can be seen everywhere. EIS appeared and was studied in the foreign countries firstly, and began to be recognized and in-depth studied by domestic researchers in recent years. At the same time, some new algorithms were derived and improved some better algorithms.This thesis was supported by the Scientific and technological project in Heilongjiang Province.On the basis of extensive data collection, it is briefly introduced the principle of EIS and some commonly used algorithms. Summed up the trends of EIS, and chose gray-scale projection algorithm as the main research according to the comparison between different algorithms. Because of the inadequacies of the algorithm, it needs to be improved. After the improved algorithm to be understood, it is studied and researched deeply. A problem was discovered by accident during the experiment. It is that processed results by the algorithm will be some errors when moving objects exist in the continuous image. According to this problem, a method, aims at moving objects'improvement based on gray projection of algorithm of the electronic image stabilization, was proposed in this paper. Through the relevant data and a large number of experiments, the cause of problem and method of solution has been found. This method can make algorithm better to process those images. And then, make a lot of experiments to validate it.With the above case, the improved algorithm was used in other environments, such as forest, and the issue still exists. The reason is that forest image is more complex to compare with others. Therefore, by the help form the innovative ways in TRIZ theory, a new method was proposed. It is that TRIZ theory is used to improve based on gray projection algorithm for electronic image stabilization. According to this theory, found a simple and effective method to preprocess the image, and make the image as much as possible to exclude invalid information, so it is can help to reduce errors. Through experimental verification, the method is effective.Further research in forest environment, in order to solve such issues, color model was studied, and proposed that based on the HSI seasonal adaptive method to Improve algorithm of the electronic image stabilization, and was applied to the forest fire early warning system. On the basis of long-term data collection, experimented with different trees in spring and autumn. It is proved that the seasonal adaptive method can automatically determine the seasonal changes. And this method satisfies the requirements of forest fire video early warning system.In the end, summarized the main studied contents of the full paper, and proposed some suggestion and direction to further study.