Deep Learning Techniques for Air Pollution

Abstract

One of the most hazardous pollutions all over the world is Air Pollution. Nearly 7 million people get killed each year by air pollution only. Normally 9 out of 10 people breathe highly polluted air. But in terms of India, Ambient air pollution ranked 7 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> in the overall health burden and 5 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> in mortality. As per WHO, 14 cities from the top 20 most polluted cities all over the world are from India which contains particulate matter of <2.5mm in diameter as a part of air pollutants. And, top 10 most polluted cities from the world are Kanpur, Faridabad, Varanasi, Gaya, Patna, Delhi, Lucknow, Agra, Muzaffarpur, and Srinagar. Which shows the severity of the study of air pollution in India. So with the help of different available deep learning techniques we can forecast the level of pollution in certain areas and also find the solution for limiting it. So from this study, researchers can easily find out the challenges and limitations of the research in deep learning field. And, they can also see what are the potential way to utilize CNN-LSTM for their research in the field of air pollution in India.