语法修改.doc

Nayar and Narasimh ADDIN NE.Ref.{CB2E1CA3-9F77-4E49-AAA38BE}[1] expressed the complicated atmospheric scattering model into two parts. Attenuation model, in the process of transmission to the imaging device the light on the object surface is attenuated; Airlight model, atmospheric light shining on the suspension particle, after scattering and reflex reach imaging equipment take part in the imaging. Narrow-band camera obtained images’ gray values can be expressed as follow by the Lambert-Beer model:  MACROBUTTON AuroraSupport.NoMacro [Automatic section break] ADDIN  ,  MACROBUTTON AuroraSupport.PasteReferenceOrEditStyle ( SEQ Eq \* arabic \* MERGEFORMAT 1) ADDIN  where is the observed foggy image; is two-dimensional spatial coordinates expressed as , is atmospheric light, is the scene albedo, is the distance between the camera and the scene, is the atmospheric attenuation coefficient. The essence of the method based on the physical model is a use of atmospheric scattering model or it’s deformation to solve the scene albedo. This kind of method often needs to estimate the additional information of imaging system or scene, including the depth of the scene and the sky prior information, etc. Recently, single image haze removal ADDIN NE.Ref.BCA4-4FFD-94C4-595FA7F54EEB}[2-8] has achieved significant progresses as a result of its physically effects and flexibility. But they depend on the reasonable assumption or strong prior information and some shortcomings, so these methods all need to be improved. He’s ADDIN NE.Ref.{4E910AC8-39A1-4A2C-AD8F-9B7BB0BA97A7}[7] work is based on the statistics of the haze-free outdoor images and the key observation: most local patches in haze-free outdoor images contain some pixels which have very low intensities at least one color channel. This approach is physically valid to handle distant objects even in the heavy haze image and the result contains few halo effect. But He refined the transmis