A Fast Video Haze Removal Algorithm Via Dark Channel Prior

In recent years,with the rapid development of video image processing technology,the carrier of information transmission is extended from simple words and sound modes to the current image and video mode.Because of the increasingly developed modern industry,the frequency of haze weather is becoming more and more high,and the weather conditions will lead to a sharp drop in the visibility of the weather,which seriously affects the collection and processing of outdoor video images,so that the video system can not work properly.Therefore,to improve the image quality under the fog and haze weather monitoring system and reduce the negative impact of the poor image quality under the haze weather,it is of great theoretical and practical significance to study the technology of video image demogging.At present,there are two main classification methods for fog degraded images: one is image defog based on image enhancement,and one is image defog based on image restoration of physical imaging model.At present,the most representative method of image restoration based on physical imaging model is the image de fogging algorithm based on dark primary priori.The algorithm has good fog removal effect and high clarity after removing the fog.The algorithm has attracted the attention of many scholars at home and abroad.However,the algorithm also has some shortcomings,for example,the image in which the brightness changes are relatively large will appear unfoggy.In addition,the complexity of the image demogging algorithm based on the dark original color prior is more complex,and the computational memory cost is large and the requirement is higher.Computing resources,for the digital webcam with limited outdoor computing resources,will reduce the real-time performance of video fogging.On the basis of the atmospheric scattering model,this paper analyzes the shortcomings of the dark original color prior fog algorithm in the video image removal,and proposes a fast fogging optimization algorithm based on dark original color prior.The main work of this paper is to do the following aspects:Firstly,according to the current haze removal algorithm base on dark channel prior need high-configuration device and halo effect could occur around the edge,this paper suing method of guided filter and median filter combination instead of the classic dark channel prior soft matting algorithm,improved memory consumption is relatively large and the halo effect is obvious;Secondly,at present in the video image haze removal algorithm based on dark channel prior with poor real-time performance,the transmission process for abandoning the traditional method,based on guided filter and median filter combination optimization method is adopted to further down sampling and interpolation of the transmittance basis,It reduces the computational complexity of the fog removal algorithm and shortens the processing time.Finally,in order to verify the real-time performance and effectiveness of the proposedalgorithm,this paper compares the image processing results with the classical MSRCR algorithm and the image processing results based on the dark original color prior algorithm.The experimental results show that the fogging algorithm in this paper satisfies the requirements of image recognition.The algorithm can reduce the time of fog removal and improve the real-time performance of video image de haze.The experimental results show that the algorithm can also effectively identify the target objects in the image,which greatly reduces the cost of the image video processing equipment.