Abstract
As of today, increased air pollution has disrupted the air quality levels, deeming the air unsafe to breathe. Traditional systems are hefty, costly, sparsely distributed, and do not provide ubiquitous coverage. The interpolation used to supplement low spatial coverage induces uncertainty especially for pollutants whose concentrations vary significantly over small distances. This chapter proposes a solution that uses satellite images and machine/deep learning models to timely forecast air quality. For this study, Lahore is chosen as a study area. Sentinel 5-Precursor is used to gather data for Sulphur Dioxide (SO2), Nitrogen Dioxide (NO2), and Carbon Monoxide (CO) for years 2018-2021. The data is processed for several AI models, where convolutional neural networks (CNN) performed the best with mean squared error (MSE) 0.0003 for the pollutants. The air quality index (AQI) is calculated and is shown on web portal for data visualization. The trend of air quality during COVID-19 lockdowns is studied as well, which showed reduced levels of NO2 in regions where proper lockdown is observed.
TopLiterature Review
This section describes the conventional ways of monitoring air quality and discusses the research on the use of these methods. In addition, studies on using IoT, ML, and GIS data for air quality analysis are discussed. The section also discusses the importance of air quality as a parameter in COVID-19 pandemic related research.
Key Terms in this Chapter
Orbiting Carbon Observatory-2: It is an American environmental science satellite launched by NASA as a replacement for the Orbiting Carbon Observatory, which was lost in a launch failure in 2009.
Aerosal Optical Depth (AOD): It is the measure of aerosols (e.g., urban haze, smoke. particles, desert dust, sea salt) distributed within a column of air from the instrument.
Temporal Resolution: It refers the smallest possible time difference between two readings of a sensor.
Keras: Keras is an open-source software library that provides a Python interface for artificial neural networks.
Sentinel-5 Precursor: Sentinel-5 Precursor is an Earth observation satellite developed by ESA as part of the Copernicus Program to close the gap in the continuity of observations between Envisat and Sentinel-5.
Multivariate Time Series: Series that consist of more than one time-dependent variable and each variable depends on the past values of other variables.
Air Quality Index (AQI): An air quality index is used to communicate to the public on how polluted the air currently is or how polluted it’s forecast to become.
NetCDF: Network Common Data Form is a set of software libraries and machine-independent data formats to support creation, access, and sharing of array-oriented scientific data.
Spatio-Temporal Data: Data collected across space and time.
Cloud Virtual Machine: A virtual machine or operating system that runs over a cloud.
React.js: React is an open-source, front-end JavaScript library for building user interfaces or UI components.
Spatial Resolution: It refers to the size of the smallest feature that can be resolved by the satellite sensor.
ConvLSTM: It is a type of recurrent neural network for spatio-temporal prediction with convolutional structures in both the input-to-state and state-to-state transitions.
Tensorflow: Tensorflow is a symbolic math library based on dataflow and differentiable programming. It can be used across a range of tasks but has a particular focus on the training and inference of deep neural networks.
SMOG: A mixture of smoke and fog is called smog.
Express.js: It is a backend web application framework for Node.js. It is designed for building web applications and APIs.
JavaScript: A web-scripting language that creates animations and adds dynamic functionality for websites.
Node.js: Node.js is an open-source, cross-platform, backend JavaScript runtime environment that runs on the V8 engine and executes JavaScript code outside a web browser.
Satellite: It is an artificial body placed in orbit around the earth or moon or another planet to collect information or for communication.
TROPOMI: The TROPOspheric Monitoring Instrument is a satellite instrument on board the Copernicus Sentinel-5 Precursor satellite. It was launched in October 13, 2017.