The Research Of Fast Image Haze Removal Algorithm For Earth Observation From Aviation

With the technical feature of flexibility,earth observation from aviation,as a new technology,is playing an increasingly important role in many fields.But the severe fog/haze weather makes it face new challenges.The existing algorithms can't be directly applied to the background of aerial observation because the image dehazing algorithms may be invalid due to the inaccurate estimation of key parameters,and the low processing speed makes these algorithms difficult to use in the actual application.Based on this,the research on restoration of color and gray-scale image under the fog/haze weather conditions that are two application aspects of earth observation from aviation is carried out and the fast implementation of algorithm is solved in this paper.In the aspect of color image dehazing,faced with the interference of complex artificial light and strong reflector,the mutual restricted problem of the estimation accuracy of global atmospheric light and transmission map existing in dark channel prior is solved.In addition,the guided image filtering is applied to the field of image dehazing.And then the theoretical analysis and optimization solve of light energy equation are given to improve the applicability of this filtering,and it can realize the fast color image dehazing.In the aspect of gray-scale image dehazing,faced with the current situation that the research of gray-scale image dehazing is relatively weak,the dehazing strategies of visual enhancement method and the atmospheric degradation model are combined,and then the concept of scene depth proportion image is proposed.Combined with the technology of wavelet decomposition,the gray-scale image can be decomposed to the vaious frequency domain subimages.Different recovery methods are used to the different frequency subimages of original gray-scale image.And then it can realize the fast gray-scale image dehazing.Experiment results demonstrate the power of the proposed algorithms in the indexes of subjective evaluation,objective evaluation and processing speed.The research can also provide some theoretical and technical reference for the future high resolution aerial observation in the fog/haze weather conditions.