Research On Restoration Algorithm Of Haze Degraded Images In Mongolia

In recent years,with the rapid development of economy,urban air pollution,especially haze weather,frequently occurs in Mongolia.It not only brings inconvenience to people's travel,but also challenges video monitoring in the field of computer and security.For Mongolia,it is facing the rapid development of economy,city and population,and the serious lag of transportation and road construction,which is also one of the key factors restricting the national economic development.As a developing country,the state of Mongolia's air ecosystem is declining rapidly.In recent years,with the population explosion and population urbanization migration,especially the increase of the yurt area around the capital Ulaanbaatar City,it has brought bad impact on the air environment.Secondly,the development of the city and industry also makes the air pollution in Ulaanbaatar city serious.The annual average value of PM10 has reached 279 ? g / m3,and the annual average value of PM2.5 has reached 63.0 ? g / m3,which is significantly higher than the upper limit value of 20 ? g / m3 and 10 ? g / m3 recommended by who.Therefore,the haze in Ulaanbaatar is more and more serious in winter.The appearance of haze weather causes the quality of the images collected by traffic monitoring equipment to decline seriously.The collected haze images have a negative impact on the application of computer vision in outdoor environment,such as object detection and video monitoring.Therefore,the research of image defogging algorithm has become a hot spot in the field of computer vision.In this paper,the image restoration algorithm is studied based on the haze image collected in Ulaanbaatar,the capital of Mongolia.Image de fogging improves the image quality by image de fogging algorithm to get the restored non fogging image.The main research contents include the analysis of the geographical environment,haze characteristics and haze image characteristics of Ulaanbaatar city;the analysis of non physical model and physical model two commonly used algorithms for haze removal,simulation experiments and results analysis,analysis and comparison of the advantages and disadvantages of various algorithms;the paper focuses on the analysis of what he Kaiming and others said to use the dark channel priori to remove fog from a single image The algorithm is proved to be effective for most of the fog images,but its disadvantage is that the algorithm runs for a long time,so it is difficult to meet the effectiveness requirements.