Dodging Research For Remote Sensing Image

High spatial resolution is an important manifestation of high resolution remote sensing, remote sensing applications is currently the majority of workers in general concern indicators. Current spatial resolution satellite remote sensing image has reached the meter level, sub-meter, high-resolution satellite remote sensing of the enormous military value and economic benefits is attracting global military and civil applications are highly valued."Long-term Scientific and Technological Development Plan (2006-2020)"," space development" second five "plan" and so on the development of high-resolution Earth observation system to make the deployment, from the sky, near-space, aviation ground, multi-layered stereoscopic observation platform for high resolution remote sensing data.Remote Sensing get shorter and shorter intervals, the map updates faster and faster, which requires fast and accurate processing of remote sensing images dodging algorithm processing is one of the more important aspect of serving the brightness of the image can be removed, uneven contrast, protect image quality and visual effects for the production and interpretation of remote sensing image interpretation provide a scientific basis.Papers around dodging algorithm remote sensing conduct research, through analysis of the current strengths and weaknesses of existing algorithms and problems, put forward a corresponding improvement methods, the main innovations are as follows:1. For dodging Based on HIS transform algorithm to handle high dynamic range image image distortion occurs when the problems raised by using image dynamic sub-block HIS transform ideas and dodging the combination treatment. Dynamic of the image block is divided into blocks of high photon normal block and dark sub-block, and then use the sub HIS transform blocks of uniform light, and finally the treated linear stretch sub-blocks, ensure that the entire width consistent color images.2. Dodging for homomorphic filtering algorithms exist in the high frequency part of the process is not complete, image distortion and other issues, propose the use of wavelet packet on homomorphic filtering algorithm is improved idea, mainly using wavelet packet in the high frequency signal components to allow for more detailed the decomposition of this advantage, homomorphic filtering through images of wavelet packet decomposition and reconstruction, completed the image processing, effectively avoid distortion, meanwhile, in order to increase the image edge information of the original image with uniform edge detection results light fusion algorithm results, making the processed image edge information more abundant, experimental results show that the new algorithm for different types of uneven grayscale and color images have better processing power.