The Color Constancy Algorithm Research Based On Computer Vision

Reflective spectral energy distribution on the surface of the object by the illumination of spectral energy distribution and surface spectral reflectance.Therefore,we look at a picture of color images of the surface color is not the real color of the object,but the color of the illumination and the intrinsic color of the object mixed the final color we see.To restore the real object surface color is required to remove color scene lighting lighting,this recovery method of true colors is called the color constancy algorithm.Considering the color processing in computer vision and computer graphics,the importance of estimation and representation in the scene lighting spectral reflectance plays an important role in improving the ability to apply a categories such as scene reconstruction,color reproduction,surface segmentation,target recognition and reflectivity estimates.For most non-homogeneous materials,the surface reflection is a linear combination of the diffuse and specular reflection component,the famous double color reflection model theory(DRM).While in computer vision and object recognition,many of the existing algorithms assume that only diffuse pixels in the scene,this is inaccurate for truth.The presence of specular reflection component image can cause many traditional computer vision algorithms produce incorrect results.Therefore,this paper monochromatic images of reflection component separation and puts forward the effective method.A novel segmentation technology based on double color reflection model,we call the mean deviation segmentation technology referred to as MSD,referenced to the mirror separated the light from the diffuse component.This technology provides an isolation by diffuse shading pixels directly access methods to access the surface shape information.And,more importantly,this method does not need any regional color segmentation,which is different from each color information gathered in the plane of the image and the traditional method of operation.And by combining the technology of Tan's theory of the iterative algorithm to extract the specular and the MSD technology,we put forward a kind of high efficient separation of pixel adaptive fusion algorithm,using the MSF image to reach the final pixel separation and the purpose of scene reduction.This paper puts forward an effective blending algorithm to separate the reflection components in the color images,this theory is based on the physics of chromaticity analysis and statistics of the appropriate choice of like blank color,mainly for the single light color images of real color reduction were studied.Through solving the least squares question of the dichromatic reflection model,reflection component separation method was used for a single pixel level,don't need to image segmentation and the local interaction of adjacent pixels.To explore the advantages of the technique on MSD,also design the scene lighting vision system is used to estimate the image of specular reflection properties.Target image feature subset provides a direct method to the information of the surface light,and recreates the each color channel of the dynamic distribution of specular reflection.This kind of image enhancement technology effectively recreate the reflectivity of the lighting scene,without transition the property of surface diffuse scenario represent the scene itself characteristics.This blending algorithm will be effective and have great significance in theory and application value.