A novel intelligent stereo vision approach for blast-induced fragmentation size distribution: Case study at Golgohar open-pit mine, Iran

Rock fragmentation assessment is a critical aspect of evaluating blasting operations in the mining industry, offering insights for optimising procedures and subsequent mining processes. Traditional methods, such as sieving, while accurate, are costly and time-consuming. Digital image analysis methods have emerged as alternatives, yet they face challenges in accurately handling fragments of varying sizes. Recent advancements in computer vision, particularly deep learning techniques, offer promising solutions to automate and refine fragmentation analysis. This paper introduces a novel intelligent stereo vision approach utilising YOLOv8, an advanced deep learning model, for blast-induced fragmentation size distribution assessment at the Golgohar open-pit mine in Iran. By training YOLOv8 with an automated self-labelled dataset and evaluating its performance against real mine site data, this research aims to enhance the accuracy and efficiency of blast fragmentation analysis. The proposed method leverages stereo images and machine learning to provide precise measurements, leading to improved optimisation of blasting parameters and potential economic benefits compared to traditional methods. Through rigorous validation processes and case studies across diverse scenarios, the proposed approach demonstrates robustness and reliability in assessing blast fragmentation, offering a practical and efficient solution for the mining industry's evolving needs.