Intelligent assembly and disassembly method for coal mine drill pipe based on object detection

The assembly and disassembly of drill pipes are crucial operations in coal mine drilling, and their automation is essential for both worker safety and operational efficiency. However, due to the complex connection mechanisms of drill pipes, the harsh mining environment, and the lack of relevant research, intelligent disassembly remains a significant challenge. This study proposes a novel quadrangular spiral drill pipe and an intelligent assembly and disassembly method. The drill pipe features a convex quadrangular prism male end with a locking pin and a matching concave female end with an unlocking pin, enabling bidirectional torque transmission and axial tension transfer. Chamfered interfaces (10×5 mm and 20×10 mm) allow self-aligning assembly under angular deviations up to 10°, effectively reducing coupling precision requirements. Moreover, a lightweight YOLOv5-based detection model enhanced with a MultiScale Efficient Channel Attention (MSECA) module is developed for robust detection of key components (pin, box, lock, and unlock) under low-contrast and noisy underground conditions. Furthermore, a geometric projection-based method is introduced to calculate the relative angles of the pin and box, providing a reliable decision basis for intelligent alignment and disassembly. Experimental results verify the effectiveness of the proposed method: the maximum angle error for the pin is 5°, with an average of 2.9°, and for the box, the maximum is 8.9°, with an average of 6.1°; the overall maximum pin–box deviation is 8.8°. During unlocking, the box shows a maximum error of 4.2°, with an average of 2.85°. Coupling and disassembly tests further confirm the robustness and reliability of the proposed system.