A Smart Approach to Environmental Toxicology and Monitoring: Artificial Intelligence for Air Pollution Detection in the Coastal Regions of Bangladesh

Abstract

Globally, the contamination of the air by damaging agents presents a critical environmental and public health concern. These pollutants, arising from a spectrum of human and natural origins, impair air quality and create severe hazards for human health, ecological balance, and climatic conditions. Recognizing the intricate nature of air pollution, characterized by various gases and particulate matter, accurate and timely identification is indispensable for devising and enacting effective solutions to protect our planet and its inhabitants. Monitoring air pollution in dynamic places is much critical due to various emissions such as industrial, automotive, biogenic and seasonal fluctuations as well. Traditional air quality sensors often lack the real-time data needed for effective intervention and pollution control. To address this, this research introduces an AI-driven solution for real-time air pollution monitoring, particularly for the coastal region of Bangladesh. The approach utilizes a comprehensive dataset of PM2.5, PM10, and NOx from governmental and regional monitoring stations. Employing supervised learning techniques with machine learning and deep learning techniques, the proposed system can accurately (97%) predict pollution levels in real-time and identify pollution trends. The findings highlight the capability of Artificial Intelligence in enhancing air quality governance and supporting public health decisions in densely populated areas.

FENI UNIVERSITY JOURNAL, 2025, 4(1), PP. (25-52)