A Big Coal Block Alarm Detection Method for Scraper Conveyor Based on YOLO-BS

Area of Science:

• Mining Engineering
• Computer Vision
• Artificial Intelligence

Background:

• Underground coal mining faces challenges with large coal blocks causing scraper conveyor congestion.
• Accurate detection of oversized coal blocks is crucial for preventing operational disruptions.

Purpose of the Study:

• To develop an advanced algorithm for detecting large coal blocks on underground mine scraper conveyors.
• To improve the accuracy and real-time performance of coal block detection systems.

Main Methods:

• Proposed the YOLO-BS (YOLO-Big Size) algorithm, modifying YOLOv4 by incorporating a transform module and optimizing the PAnet module.
• Integrated SimAM attention mechanism to mitigate feature loss and enhance model convergence.
• Utilized focal loss to address sample imbalance and developed a formula for calculating congestion rates.

Main Results:

• The YOLO-BS algorithm demonstrated a processing speed of 80 frames per second (fps).
• Achieved a correct alarm rate of 93% in detecting abnormal coal block phenomena.
• Validated performance using a dataset of 12,000 images from Daliuta Coal Mine.

Conclusions:

• The YOLO-BS algorithm meets the stringent real-time and accuracy requirements for underground mine conveyor monitoring.
• This AI-driven approach offers a viable solution for mitigating coal congestion issues.