Research And Implementation Of Image Haze Removal Method Based On Dark Channel Prior

With the progress of the society,the development of industry,intelligent transportation systems,outdoor surveillance system,satellite remote sensing system,agricultural product quality detection system based on image information carrier for operation system has gradually entered people's life.In order to meet the job demands of these system,the hardware facilities of image acquisition are continuously improved,now have made great achievements.However,the haze caused by environmental pollution brings new challenges to people.In the haze weather,the image is not obvious,the color is darkened and the details of the scene are lost,so the quality of the image and its application value are greatly reduced.Therefore,the fast and effective haze removal method has the theoretical and practical urgent need for people.This paper solves the problem of long processing time of dark channel prior method in image restoration.First of all,the paper adopts multi-resolution method to down-sampling high-resolution haze images,at the same time,in order to get more edge information of haze image,this article joins twice down-sampling,and get two sub images,select one of the sub images to use the dark channel prior method,and extract the details of the other sub-image.Secondly,the two images are fused together and then up-sampled to get the final clear image.Finally,the results of the dark channel prior method and the improved processing method are all calculated by the operating system for Windows 7,the CPU is Intel(R)Core(TM)i7-37703.4 GHz processor,and the system memory is 8.00 GB of ordinary personal computer.In addition,the article evaluates and compares these two results.The experimental results show that both subjective and objective evaluation indexes are considered that the improved method proposed in this paper is not only improved greatly in running time,but also has a good visual perception.It can greatly reduce the amount of computation in the premise of ensuring the effect of haze removal,thus making the dark channel prior method more practical.