Monitoring Toxic Gases in Dhaka Brick Kiln Areas & Comparing CNN vs. ResNeT for Brick Kiln Identification Via Satellite Imagery in Bangladesh

In developing nations like Bangladesh, clay bricks are essential for construction, but the firing process generates exhaust and toxic gases that degrade air quality and harm local flora and fauna. Ensuring compliance with environmental regulations is crucial for clean environments and public health. In South Asia, small-scale, informal brick producers are significant polluters, making monitoring and regulation challenging. This study presents a cost-effective, scalable method for identifying brick kilns using high-resolution satellite imagery from Bangladesh, sourced from the Kaggle database. Utilizing convolutional neural networks (CNN) and Residual Neural Networks (ResNet) deep learning models on 6,147 satellite images, the approach achieves 100% accuracy and precision in identifying chimney kilns. Additionally, the research employs an Internet of Things (IoT)-based system with Arduino Uno, NodeMCU ESP8266, and MQ-7, MQ-8, and MQ-136 sensors to detect Carbon Monoxide (CO), Hydrogen (H2), and Sulfur Dioxide (SO₂). Experimental results show significantly higher sensor readings in polluted conditions: MQ-7 (clean air: 60.0650 ppm, polluted air: 500.0423 ppm), MQ-8 (clean air: 20.5425 ppm, polluted air: 150.6263 ppm), and MQ-136 (clean air: 70.1921 ppm, polluted air: 200.0489 ppm).