Enhancing Coal Quality Control with Computer Vision for a Sustainable Green Energy Transition

More than 860 million people worldwide live in energy poverty. While many developed nations are making a strong transition to renewable energy sources, developing countries continue to face significant challenges in adopting alternative forms of energy on a large scale. The International Energy Agency (IEA) projects that coal will remain a critical source for meeting the power needs of developing nations for the coming decades. Therefore, improving coal quality is vital for reducing environmental impact and facilitating a cleaner energy transition. Coal quality directly influences emissions, combustion efficiency, and the effectiveness of carbon capture technologies, making rigorous quality control essential for sustainability. Ultimately, coal quality plays a key role in determining the level of environmental impact. Coal quality control can be challenging due to the inherent limitations of manual assaying. This work is devoted to exploring the currently available solutions for coal quality assessment and demonstrating a simple methodology based visual range imaging for assessing coal quality. The methodology considers (i) image data collection from known samples, (ii) training a machine learning model, and (iii) evaluating the model's performance against existing methods. The model, trained using transfer learning, achieved over 95% accuracy on a test set, demonstrating its reliability and effectiveness. The integration of machine learning supports efforts to develop a sustainable value proposition for clean coal, in line with sustainable development goals.