Study Of Method For Removing Cloud From Remote Sensing Images Based On Wavelet Transformation

With the rapid development of science and technology, remote sensing technologyhas achieved rapid development. It is widely used in many fields in the world, andprovides the possibility for human to understand the earth environment from the macroperspective. As the direct product of remote sensing technology, remote sensing imagescan provide a lot of valuable information for the people. But because of the complexityof atmospheric environment impact on spread spectrum and the limitation of remotesensing itself, optical remote sensing imaging process is easily affected by cloud andmist, resulting in remote sensing image blurring or loss of information, bring a lot ofinconvenience to the image analysis and processing. Research on how to eliminate theadverse effect is a hot topic in the filed of remote sensing image processing. Based onthe investigation of remote sensing technology and image processing technology,remove cloud method based on wavelet transform is studied. The research contents andachievements are as follows:1. Study the theory of wavelet transform comprehensively, analyses the differencesand contacts between wavelet analysis and Fourier analysis. Make a summary of thecourse of development of the wavelet transform, described several fundamentaltransformation in the wavelet transform, and focus on the multi-resolution analysis andMallat algorithm.Study on the application of wavelet analysis in the current situation,and puts forward its strong advantages in the field of image processing. On the basis ofsummary of study cloud imaging principle, imaging model of thin and thick cloudsimaging model are studied, established the degradation model of remote sensing imagecontaining thin, thick cloud information. These are the theoretical basis of this research.2. Studied the classical methods for removing thin cloud, such as multispectralimage method, histogram matching method, the homomorphic filtering method etc,analysis of these methods are summarized to cloud effect, advantages and disadvantages.On the basis of summarizing the principle and method for wavelet transform to removethin cloud, presents a thin cloud removal algorithm based on wavelet transfom, thisalgorithm can improve the distinguish inaccurate disadvantage of general wavelet algorithm in removing the thin cloud information, and as much as possible to keep thedetails of the information of the earth’surface details. It is conducive to further analysisand recognition after the image processing.3. In-depth studied the principle, method of wavelet fusion and method of clouddetection. Based on the principle of thick cloud in the low frequency part of the remotesensing image, put forward the wavelet fusion scheme to remove the thick clouds, theexperimental results show that, the fused image not only can effectively remove thethick cloud, but also can rich the image information, achieve the desired effect.