Research On Digital Image Fusion Algorithm For Traffic

With the rapid development of the national economy, the traffic has become the focus of attention of the public, the fluency of the urban traffic system, related to the life of the people happiness index. Domestic and foreign transportation departments and relevant scholars in the continuous upgrading of research to respective traffic control system, in order to improve the service index of intelligent transportation system(ITS), most of the intelligent transportation system development depends on the improvement of the hardware,such as raising the camera sensor acquisition quality, increase the sensing function, cost a lot of money and, on the basis of the above measures directly the traffic image digital processing, the lower cost and higher efficiency.The image of a single sensor processing can not meet the development of intelligent transportation system, the advantages of multi sensor intelligent traffic image information fusion is more obvious.In this thesis, the main research direction is traffic multi digital sensor signal fusion theory and algorithm,proposed the combination of wavelet transform and HSI transform for fusion algorithm of the traffic image, with needle for vehicle traffic objects put forward traffic images based on the approximate gray fusion algorithm, through MATLAB experiment platform for target simulation, the effect is obvious.The first chapter of this thesis first introduces the nowadays world urban intelligent traffic control system development, through understanding the development of intelligent transportation leads this to video module of intelligent transportation system in multi sensor fusion technology integration theory and the development situation. And a detailed explanation of the image evaluation standards.Next is the focus of this paper, paper in the third chapter the HSI transform fusion algorithms based on wavelet transform and, fusion algorithm is effective in improving the quality of traffic image, fusion results traffic image target feature prominent, improve the limitations of single traffic sensors. Finally, the paper puts forward the fusion algorithm based on approximate gray image of the traffic, the algorithm by constructing a new window function, the choice of the pixel through set the center point to select, close range image information of the reservation principle is through the pixel and the center point of the distance to determine, according to the distance of the pixels and center point to selectthe information lost. This can weaken the image fusion edge traces of, the effect is obvious.