Research On New Color Parameterized Logarithmic Image Processing Model And Its Application

Linear image processing algorithms can lead to serious loss of image information.Logarithmic image processing models can better deal with details of images and outstand the image areas people interested in.Logarithmic image processing models are widely used in image processing because of its unique advantages,successfully solving many problems of image processing.Therefore logarithmic image processing models more and more become the research hot spots.However,the existing models of logarithmic image processing are mostly based on the gray images and the logarithmic models of color images remains to be development.How to apply logarithmic image processing models to the color images is the problem to be solved.First,this paper summarizes the existing logarithmic image processing model and analyzes the human visual system.Second,in view of the existing logarithmic bottleneck in the field of color image processing model,a new model of color parameterized logarithmic image processing(COPLIP)is proposed,extending the parameterized logarithmic image processing model(PLIP)to color images.Color tonal function of color images,operations of addition,multiplication and invertion and the concept of fundamental homomorphism function are defined.Then this paper compares the graphics of COPLIP fundamental homomorphism with PLIP model,and compares vector operations of COPLIP with basic linear model and the latest PSLIP(Parameterized Symmetric Logarithmic Image Processing),analyzing the characteristics and effectiveness of the new COPLIP model.In order to verify the validity of the new model,this paper designs an image enhancement and an edge detection algorithm based on COPLIP model.The image enhancement algorithm uses a-rg-yb space which is in accord with human visual model.Through the subjective evaluation and EGRNN-BIQA objective evaluation algorithm for blind image quality,the color of the enhanced images is more natural,with the better visual effect.COPLIP edge detection uses the idea of output component synthesis,and chooses the laplacian operator for edge detection with better effect.The images are all achieved rich edge information in COPLIP edge detection algorithm of different parameters.