An Integrated Color Space–Thermochemical Framework for Predicting Coal Spontaneous Combustion Liability

Coal spontaneous combustion (CSC) remains a critical safety, environmental, and economic concern in coal mining, storage, and transportation, while conventional prediction methods based on thermochemical indices and proximate analysis often show limited reliability for heterogeneous coal datasets. This study develops an integrated framework to improve CSC prediction by combining traditional combustion indicators with quantitative surface color analysis. A total of 216 coal samples, representing diverse ranks and geological origins, were systematically evaluated. Standard liability parameters, including Crossing Point Temperature (CPT), FCC index, and Average Temperature Increase (ATI), were determined experimentally, while surface characteristics were quantified using digital image analysis based on RGB and HSV color spaces. Statistical evaluation revealed strong and consistent correlations for all indicators, with high predictive accuracy, particularly for CPT (R2 = 0.9783), and similarly robust relationships for FCC (R2 = 0.8805) and ATI (R2 = 0.9183). The enhanced performance is attributed to the sensitivity of color parameters to early-stage oxidation processes, which precede measurable thermal and gaseous changes. These findings demonstrate that integrating image-based features with conventional indices provides a reliable, nondestructive, and scalable approach for assessing CSC liability, offering significant potential for both laboratory studies and practical coal safety management applications.