Particle Pollution Estimation from Images Using Convolutional Neural Network and Weather Features

Abstract

Airborne particulate matter with a diameter less than 2.5 micrometers $(\mathrm{PM}_{2.5})$ is one of the most harmful air pollutants, because $\mathrm{PM}_{2.5}$ can be inhaled into human body and cause serious health problems by transmitting hazardous chemicals deeply into lung and bloodstream. A reliable, easily accessible, and low-cost $\mathrm{PM}_{2.5}$ monitoring system can greatly help people raise public awareness of $\mathrm{PM}_{2.5}$ and reduce health hazards of air pollution. In this paper, we combine image and weather information to estimate $\mathrm{PM}_{2.5}$ indices of outdoor images using deep learning and support vector regression (SVR) techniques. The proposed method first uses a convolutional neural network (CNN) to predict the $\mathrm{PM}_{2.5}$ index based on image information, and then the $\mathrm{PM}_{2.5}$ predicted by CNN and two weather features, humidity and wind speed, are combined to yield final estimated $\mathrm{PM}_{2.5}$ index using a created SVR model. We assessed our method using two datasets collected from Shanghai City and Beijing City in China and experimental results demonstrated the effectiveness of the proposed method for $\mathrm{PM}_{2.5}$ estimation.