| Green waters and green hills are Jinshan and Yinshan,which have been deeply recognized by people nowadays.Because environmental problems are closely related to everyone,in most parts of China,especially in northern cities,the fog and haze problem is becoming more and more serious,which directly affects people's travel and even health.Nowadays,it is the daily life of many people to care about the air quality every day.However,it should be recognized that the weather quality obtained by the public is based on the data of meteorological observation stations,which are sparse and unevenly distributed,and generally do not locate in densely populated areas.At the same time,the air quality of urban areas must be very different from the surrounding areas,which leads to the inaccurate air quality results.In view of fog and haze phenomenon,the digital image field mainly studies how to remove it.Some scholars in the field of meteorological observation have studied the correlation between air quality and scattering coefficient by themselves.The uneven distribution of meteorological observation stations in China leads to inaccurate real-time air quality information.Combining with the research conclusions of scholars in two fields,this paper draws a conclusion that there is a strong correlation between the atmospheric transmittance obtained by optical imaging principle in fog removal and air quality,and uses BP neural network model to learn the implicit relationship between air quality and picture transmittance,so as to qualitatively determine its air quality by pictures.It is expected to provide a reliable air quality result for people going out in haze areas.The experimental results show that the overall recognition rate of the proposed scheme is 83.72% for the four categories of air quality: good,moderate,light polluted and heavily polluted,and over 90% for the good and moderate categories,which has reached the goal of qualitative determination of air quality categories. |
PDF Full Text Request