Researches On The Method And Application Of Cloud-removing For QuickBird Image

Cloud-removing of remote sensing image has always been a hot topic and problem both at home and abroad. With the rapid development of the Satellite remote sensing technology, Remote Sensing Digital Image Processing has became an emerging technical science. However, in our remote sensing image processing, the inevitable encounter of the cloud cover interference would reduce the clarity of the region of interest, thereby the utilization and quality of the image has been dropped. So Cloud-removing of Remote Sensing Image has great practical significance.In this paper it takes QuickBird image of Tibetan collapsing areas as an example , and introduces the Principle of cloud-removing of remote sensing image and how to deal with the related issues on the appearing of thin and thick cloud together. And bring about a new method to restore new image based on Pixel Fusion of improved Homomorphism filtering, in the first place the High-frequency Emphasize filter was introduced to the cloud-removing field, some experiments was made and it's effect is perfect.Main studies are as follows:(1)Introduction of the QuickBird images, remote sensing imaging process and the related content of remote sensing information;(2)Establish thin cloud imaging and analyze the principle of pixel-level fusion;(3)Based on traditional Homomorphism filtering method, facing the thin and thick cloud problem, the paper designs a new filter which is called emphasized high-pass filter, after studying and improving ,it proposes a new improved way on integrated cloud-removing Algorithm about Homomorphism filtering .(4)Take MATLAB software as a tool, through programming it achieves that removing cloud in the gray-level remote sensing image without assisted information and proceed restoring the background by using the ideas of pixel-level integration。(5)This paper introduces the mean, standard deviation, entropy and other indicators to assess the value of QuickBird images of the effect of cloud processing, from visual evaluation and statistical analysis of the two major aspects of the image ,it evaluate the impact of the cloud. Comparing the indicators proposed in this paper, we can see the new improved integration with the homomorphism filtering algorithm can not only remove the cloud in the meantime, but also make the features background information of the areas that have no cloud , a very good recovery.