Research Of Image Enhancement Based On Improved Retinex Algorithm

Image information technology development is closely related to the progress of human society.In the field of image processing,Retinex image enhancement algorithm is also a hot research topic as a important part in the research of image enhancement.However,for the image of different imaging conditions,Retinex algorithm can play its role when it is purposefully used to enhance processing.This paper firstly outlines the imaging characteristics of digital image,conventional image enhancement knowledge,common Retinex enhancement method,as well as existing problems of various noise information,poor image resolution in the enhance process of information.Then,two following kinds of image enhancement methods have been more comprehensively proposed to combine Retinex theory in the application field of the digital and medical image enhancement:1)The first enhancement method is for digital images.Because of digital image with indistinct imaging and other characteristics,an image enhancement algorithm based on the guided filtering of Multiscale Retinex with the color restoration(MSRCR)and Contourlet transform was proposed.The algorithm improves the limitations of traditional Retinex algorithm in the application,has good stability with edge preservation,iteration speed,clear imaging,noise suppression,thus it can provide a reference for actual production and application.2)The second method is to enhance medical images.Medical images has noise,uneven distribution of gray level and differences of image type characteristic,etc..An medical image enhancement algorithm based on the rapid bilateral filtering and quantum particle swarm optimization of Retinex is proposed.The algorithm could choose the suitable parameters to enhance processing,not only highlights from the whole image information,and improves the image resolution for different types and different gray medical images.