213 / 2023-09-22 20:52:13

Research on Coal Mining strata movement Based on Comprehensive Remote Sensing dynamic Observation in Semi-desert Area

InSAR,Deep learning,Coal mining,Ground fissure,Strata movement

Underground coal mining causes stress changes in rock and soil, leading to overlying rock failures and deformation of soft or loose soil, which can cause damage to the ecological environment and building facilities. Mining-induced fractures can also lead to air flow, resulting in underground coal spontaneous combustion and even mine fire disasters, posing a serious threat to people's lives, property, and mining production. Traditional field investigation and ground measurement methods are time-consuming, laborious, and easily affected by human subjectivity. It is difficult to obtain real-time, efficient, systematic, and comprehensive information on surface rock displacement. InSAR monitoring can detect subtle displacements over vast regions, while UAV provide the benefits of enhanced spatial resolution, flexibility, lightweight, and terrain independence. Deep learning can automatically extract targets after establishing a recognition model through training samples. In this study, InSAR technology, UAV aerial photography, and deep learning models were used to analyze the changes in surface rock movement, mining-induced fractures, and rock displacement parameters of a large-scale fully mechanized coal mine in the Shenfu coalfield, located on the northern edge of the Loess Plateau in northern Shaanxi and the southeastern edge of the Mu Us Desert, where the thickness of the coal seams is about 3.30–3.82 m. The findings demonstrate that by combining remote sensing techniques with artificial intelligence methodologies, it is possible to accurately detect and monitor both ground subsidence and fractures within the mining area under this specific working condition.

Aiming to detect wide-area dynamic strata movement in the mining area, medium- to high-resolution satellite remote sensing data is employed to rapidly capture strata movement and associated surface deformation at the sub-meter scale during different stages of mining in the area. Time-series deformation of the ground surface within the mining area can be observed with PALSAR-2 L-band SAR data, which reveal the process of surface deformation in underground mining area. For the identification of coexisting ground fissures in areas with severe deformation, UAV  is utilized for its high precision, fast and flexible acquisition of centimeter-level images, and easy operation. With this technology, we obtain centimeter-level images. Deep Residual Shrinkage U-Net (DRs-UNet) was used to quasi-automatically extract the spatial distribution patterns and characteristics of mining-induced ground fissures. The strike and uphill boundary angles of working face 22308 were determined at 26.27° and 46.18°, respectively, employing surface deformation subsidence boundary by InSAR result and working conditions. The integration of UAV images and InSAR surface deformation matches well with underground coal mining progress and surface strata movement, acting as a valuable guide for surface environmental protection, underground ventilation, mining pressure, and drainage.