Realization Of Aerial Image Enhancement Processing And Connection Technology

Digital image processing technology, which includes very large contents and has very broad application prospects,is an important research direction in the fields of computer vision. In this paper, an earnest study on image denoising and mosaicing of digital image processing technology,focusing on the aerial images processing, is carried on.Aerial images, taking aerial vehicles as a platform, is equipped with aerial digital cameras, gathering the scene video on the ground to obtain image information. Because of the vertical way, which is utilized by the imaging equipment during high-speed flight, to takes pictures of a scene on the ground, and the bad influence of rain and fog cover in the atmosphere, atmospheric flow and body vibration and other factors, great noise is brought in collected original image information. For the post- processing of these images, enhancing-and removing-noise pre-processing must be carried on so as to get valuable and accurate information.This paper,starting from the digital image processing technology, gives details about the content,principle and development of it. Next, combining with the characteristics of aerial images, the aerial images removing noise technologies are seriously studied, and a variety of de-noising methods are analyzed comparatively. Moreover, according to the actual situation of clouds noise of aerial images, a new partial differential de-noising idea is proposed. At last, simulation experiments about the new idea, based on MATLAB software, indicate the algorithm is of good performance.And therefore the algorithm lays the foundation for image mosaicing.Low-rank matrix technology, developing in recent years, is a new theory of the image processing domain. In this paper, through reading a lot of literature materials, the fundamental principles theoretical study and practical applications of low-rank matrix are summarized. Meanwhile, combined with the procedure of image mosaicing, the key technologies for the image registration are introduced. And thus, a new image registration method is proposed on the basis of using the low-rank matrix image processing. Experiments shows that the mosaicing of two or more images is realized.