Panoramic Visual Image Enhancement Technology Research

Since digital images to produce human description of the objective world is becoming more and more specific. The image of the visual carry a large amount of information on their own. According to the statistics, the information obtained 70% originate from the visual.Image visualization provided people to watch, as well as the subsequent information processing and utilization of provides a rich source of information. In order to make the image of the amount of the information needed to cover more big, more rich details, people tend to need more large and complex hardware devices. But when the development of the hardware is restricted by the objective conditions and factors such as cost, from the Angle of the software to make up for the inadequacy of the hardware is very important. Thus the image enhancement processing has always been the main direction of research on digital images.Panoramic vision system is a kind of horizontal Angle is 360°, approach or exceed 180 °in the view of the unilateral visual imaging system. Panoramic imaging technology is with the geometrical optics, imaging components, the development of image processing algorithms and emerging a imaging technology. It can image sensor carry more information than the average, intuitive is good, short sampling period. There are more and more applications to the panoramic visual images. This led to the people has higher requirement for the quality of panoramic image. Therefore, panoramic visual image enhancement research has the extremely vital significance. In this paper, the research for panoramic visual image enhancement technique. Specific research contents include:First, to define and analyz panoramic vision system of the imaging model. Including:single point of view, a single point of view of the light path imaging model, and their imaging process. Finally expounds the unification to turn it into a spherical model method.And, this paper presents a panoramic visual image algorithm, enrich the forms of panoramic visual images. This article will also be on the cylindrical panoramic vision system deployment form, the perspective of form, the polar coordinates of expansion and the expansion under rectangular coordinate system respectively, image enhancement method is proposed. At the same time this paper also gives a particularly effective means of image enhancement field in recent years, super-resolution (SR) strengthen the basic principle, and image quality assessment method.Secondly, explains the maximum a posteriori probability (MAP) super-resolution enhancement method. Analysis of the image degradation model. And from the perspective of the theory of mathematical model of optimization problems. And then the variance of the algorithm by grouping empowerment improvements to from the low resolution (LR) image of more redundant information. At the same time, also can better application in the panoramic perspective of visual image image and cylinder.Then, studies the issues related to learning super-resolution conical tower model.Conical tower model from the pyramid model development. The pyramid model initial application in the field of image multi-scale fusion. But using the multi-scale level comparison between each other, can pyramid structure of training and learning. Then in order to adapt to the panoramic visual image in polar coordinates of expansion, the pyramid model improvement for the conical tower model. To the panoramic visual images without expansion, better use of the original image. The experimental results show that the method has good effect for panoramic visual image enhancement.Finally, this paper established a gray surface model under the three dimensional space.The traditional regression models need one by one estimate deformation matrix, fuzzy matrix,sampling matrix, additive noise, such as matrix. Every estimate algorithm error. Integrated with multiple matrix, image degradation model will be gravely deviates from the actual situation. Obviously, according to the inverse enhancement get estimation of image degradation process, will exist great algorithm and computer rounding errors. The topological transformation of 3 d spatial gray level surface, can effectively simulate the image degradation process, the factors of the image degradation is united in together, to avoid the excessive error. At the same time, this paper discusses the image spatial gray level surface meet the topological properties of reason and the significance, for image enhancement technology provides a new direction. And through the experiments show that the method in the treatment of panoramic visual images under rectangular coordinate system of expansion is very effective. Also discussed the three-dimensional space model combined with a polar coordinate, build specifically for spherical panoramic visual image space topology model research.