A Study Of Image Clearness Technology For Remote Sensing Image

With the development of technology, the remote sensing technology is becoming more and more widely used in the areas such as national economy and the military defense. The specific applications are investigation of the resources, military reconnaissance, meteorological observation, cartographic drawing, city management, and so on. The development of digital remote sensing image processing and the development of remote sensing technology are closely connected. To get the property and status of the ground scene, operations on remote sensing images are inevitable, especially the images after degradation. The study about improving image clearness that we focus on in this paper is the basis of any operations on remote sensing images. One can get reliable operational data source only by making the sensing images contain clear and correct information about the scene. The main work and outcomes are as follows:(1) In the preprocessing stage of remote sensing image, we deduce transform relationship between the images from different coordinate systems on the basis of the geometric models of the imaging process. We get the geometric distortion corrected by operating with the different expressions of point, line and plane in different coordinate systems.(2) Given the problem of keeping tone of color images, we transform the images in RGB color space into the ones in HSV color space. On the basis of the above, we come up with a new technology about image enhancement with the better ability of the trilateral filter to extract the luminance image from the original image than the other image filters such as bilateral filter and Gaussian filter. Experiments prove that our method can emphasize on the crucial regions, and make the global image information clearer to the users.(3) In the process of image dehazing, we come up with two new methods to make the foggy images get rid of the influence of the bad weather efficiently. The two methods are single image haze removal using dark channel prior under the neighborhood-limited consideration and single image haze removal based on the guided image filter in the RGB color space. Experiments show that both the two new algorithms can improve execution efficiency to a great degree. We will make some comparison about the outcomes of the different algorithms later.(4) We sum up current technologies of image deblurring. and come up with a reasonable procedure in this paper, that is, selecting a sub-area in the original image for kernel estimating, getting the kernel recovered from the sub-area efficiently, getting the final clear image with the information of blur kernel and the original image. With the guidance of the global handling structure, we make concrete introduction separately about the above three stages, and show some comparison between the former experimental result and the modified experimental result. Experiments prove that our algorithm can produce images of ideal quality with a reasonable time.(5) In this paper, we design and implement an image clearness processing system based on remote sensing images. The clearness improvements on the remote sensing images make getting the further understandings about the actual scene easier for users. While the system contains all the techniques implied above, it includes some common operations on general images.