Abstract: Accurate identification and quantitative characterization of mining-induced fractures in overlying strata are fundamental to controlling surface subsidence, preventing mine water hazards, and enabling the co-extraction of coal and gas. However, such fractures are typically distributed within hundreds to thousands of meters of overlying strata , rendering direct observation and in situ quantification nearly impossible. This challenge has become a common bottleneck in related engineering practices. Physical modelling experiments provide an effective method of studying fracture evolution, yet current studies lack systematic approaches for quantitative characterization, limiting the precision of fracture description and constraining research on their spatial distribution and dynamic evolution. To overcome these limitations, this study integrates physical modelling experiments with digital image processing to develop a quantitative method for fracture extraction and analysis. The proposed approach automatically identifies and maps fracture geometry and spatial distribution, and establishes an evaluation framework based on key parameters, including fractal dimension, dip angle, spacing, length, density, and connectivity. Comparative analyses between measured fractures in physical models and image-based results confirm the accuracy of the extraction and quantification methods. By applying this method, the fracture development in different overlying strata regions and the division of the “three vertical zones” are further examined. The results reveal pronounced heterogeneity in the spatial distribution of fractures, and the use of fractal dimension and connectivity enables quantitative delineation of the “three vertical zones”. This methodological innovation transforms the characterization of mining-induced fractures from qualitative description to quantitative assessment. Overall, the proposed approach addresses the shortcomings of traditional qualitative methods, enables precise quantification of fracture geometry and distribution, and provides a reliable tool for studying overlying strata movement and fracture evolution in physical modelling experiments, as well as a methodological reference for related numerical simulations.