Font Size: a A A

Correlation Model Construction And System Development Of Related Indicators Of Image Quality And Air Quality

Posted on:2019-01-01Degree:MasterType:Thesis
Country:ChinaCandidate:B Q YangFull Text:PDF
GTID:2438330551960780Subject:Computer technology
Abstract/Summary:PDF Full Text Request
At present,the concentration of pollutants in the air is detected mainly by a variety of precision testing instruments.Because of the high cost of instruments,the state set up monitoring stations in every city to detect air quality,which is coarse-grained and cannot cover every corner of the city.The algorithm of air quality detection based on image aims to detect the air quality grade of local that user taken photo.Now the popularity of mobile devices makes it possible to fine-grained detect air quality through images.Therefore,it is a very meaningful topic.Based on the existing research,we propose a new method to detect PM2.5 concentration based on images taken by smartphones,which considers the influence of relative humidity.A strategy to overcome the relative humidity influence is presented.At the same time,we propose a general algorithm of air quality grade detection without site constraints.We also design and implement an air quality rating system based on images.The work of this paper mainly includes the following three parts:(1)A new algorithm of PM2.5 concentration detection based on image is proposed.The transmission matrix of the image is restored by the optical model and the dark channel prior.We analyze the influence of relative humidity on image transmission matrix and propose a relation model of image transmission matrix and PM2.5 concentration in air by eliminating the influence of relative humidity.The model parameters are obtained by using the machine learning method to fit the relationship model.The performance experiments on our data set show that our proposed algorithm is obviously superior to the existing methods.(2)A universal method of air quality grade detection based on image is proposed,which is not restricted by location.This method is an extension of the No-reference image quality assessment algorithm.We carry on the analysis of color channel information and gray channel information,and find features that can reflect air quality.We build the air quality grade detection model based on image features.This algorithm is compared with other existing algorithms on our data set from the estimation accuracy to the time consuming of the modeling,and the experimental results show that our algorithm is more suitable for air quality grade detection than other related algorithms.(3)An air quality grade detection system based on image is designed and implemented.The system can quickly and accurately estimate the air quality of the local that user taken photo.It can help people get the ambient air quality conveniently,plan the schedule reasonably,and avoid the harm of air pollution to the body.
Keywords/Search Tags:PM2.5, image transmission matrix, relative humidity, air quality grade, Noreference image quality assessment, local entropy
PDF Full Text Request
Related items