A novel coal-rock recognition method in coal mining face based on fusing laser point cloud and images

Abstract

Rapid and accurate recognition of coal and rock is an important prerequisite for safe and efficient coal mining. In this paper, a novel coal-rock recognition method is proposed based on fusing laser point cloud and images, named Multi-Modal Frustum PointNet (MMFP). Firstly, MobileNetV3 is used as the backbone network of Mask R-CNN to reduce the network parameters and compress the model volume. The dilated convolutional block attention mechanism (Dilated CBAM) and inception structure are combined with MobileNetV3 to further enhance the detection accuracy. Subsequently, the 2D target candidate box is calculated through the improved Mask R-CNN, and the frustum point cloud in the 2D target candidate box is extracted to reduce the calculation scale and spatial search range. Then, the self-attention PointNet is constructed to segment the fused point cloud within the frustum range, and the bounding box regression network is used to predict the bounding box parameters. Finally, an experimental platform of shearer coal wall cutting is established, and multiple comparative experiments are conducted. Experimental results indicate that the proposed coal-rock recognition method is superior to other advanced models.

1. Introduction

As one of the core equipment in coal mining face, shearer is a significant guarantee for realizing unmanned mining. In the process of coal wall cutting, complex working conditions such as large gangue, coal-rock fracture, and sudden changes in coal-rock distribution frequently cause the cutting power of shearer to be unstable and mismatched, which greatly reduces the coal mining efficiency and the equipment service life. To achieve the optimal coal-rock cutting performance, it is necessary to adaptively control the traction speed and drum height based on the distribution of coal-rock in coal seams [1,2]. However, the coal mining face is characterized by constant fluctuations and complex working conditions. How to effectively identify the position of rock layers and coal seams in the cutting coal wall is still a key technical bottleneck for intelligent coal mining.

In this regard, many scholars have carried out a series of research work and formed various coal-rock recognition methods. Zhou et al. [3] proposed a coal and gangue recognition method based on R value for dual-energy X-ray of Geant4 simulation. Si et al. [4] developed a sensing identification method for shearer cutting state based on the modified multi-scale fuzzy entropy and support vector machine. Zhang et al. [5] converted the vibration signals of shearer key components into two-dimensional time-frequency maps and performed cutting state recognition by deep convolutional network. Zhang et al. [6] proposed a coal and rock type recognition method based on mechanical vision. Liu et al. [7] provided an in-depth analysis of the electromagnetic signals of coal and rock during drilling through simulations and experiments for identifying coal and rock properties. Si et al. [8] developed a novel coal-rock recognition method for coal mining working face based on laser scanning technology. It is evident that remarkable achievements have been made in the field of coal-rock recognition. Nevertheless, the following issues still exist: (1) Radiation sources have certain radioactive hazards and weak penetration, limiting its underground application; (2) Image and vibration signals are strongly interfered due to the low illumination, high dust, and strong noise generated by the complex coal mining face; (3) Laser point cloud is sparse in underground coal mine, which leads to low accuracy of coal-rock identification. In short, above traditional sensing technologies cannot obtain high-quality data in complex environments.