Algorithm For Calculating Color Constancy Of Multiple Light Sources

The spectral energy distribution of the reflected light is the product of the spectral energy distribution of the scene light source and the spectral reflectance of the object surface.Therefore,the color in the image is not the true color of the object surface.In order to restore this color,this require that we have the ability to reduce the influence of the light source color.This method of calculating the true color of the object surface is called color constancy algorithm.Fortunately,the human eye has the ability to maintain this color constancy,if it can enable the computer to do so,then many problems in the field of computer vision can be easily solved.For the multi light source color constancy computation,author proposes a novel algorithm called the inverse intensity chromaticity space-color space transformation fusion algorithm based on grid division.The algorithm does not require a priori knowledge,human intervention,too much hypothesis of spectral energy distribution and surface reflectance characteristics of the light source,and does not require large image data sets for training.The main idea is to take the method to break up the whole into parts.First,using the grid-based sampling to divide a multilight source image into some patches,and assuming that each patch has uniform light source,which is equivalent to a single color light source.Then,we use inverse intensity chromaticity space to estimate illumination chromaticity.Finally,the light sources' estimation in all areas are fitted in the RGB color cube which called the illumination axis.Through the two color space transformation,the multi source image is corrected to the standard light source,which achieves color constancy image calculation.In order to test the advantages and disadvantages of the proposed algorithm,we compare it with the existing algorithms.In view of the proposed algorithm,this paper discusses the performance evaluation criteria for the subjective and objective,and evaluates the performance of the proposed algorithm in terms of two aspects,objective and subjective.Through the experimental research,the influence of grid size on the performance of the algorithm is discussed,and the optimal grid size of each algorithm is finally determined.The experimental results show that the fusion algorithm proposed in this paper has some correction effects on the multi light source outdoor images.In addition,it has some practical value in the field of printing quality on-line detection.