Abstract: The accuracy and efficiency of coal-gangue identification are not only critical supports for clean coal utilization,but also key components of the core technology for intelligent coal-rock recognition in the construction of intelligent mines. Owing to the prominent advantages of infrared thermography,such as high recognition precision,low susceptibility to operational environmental interference,simple operation,and low cost,samples of bituminous coal and roof sandstone were collected from the 2067 fully mechanized working face in Dongpang Mine. Active thermal excitation was adopted to obtain the thermal infrared response information of coal and gangue,and the thermal excitation temperature change index was quantitatively calculated to achieve effective discrimination between the two materials. Accordingly,a coal-gangue identification method for belt conveyor transportation was proposed. The experimental results indicate that during both heating and cooling stages of thermal excitation,the thermal radiation temperatures of bituminous coal and roof sandstone exhibit significant variations,and the infrared response characteristics of coal and gangue show distinct differences,with the temperature variation amplitude of bituminous coal being notably larger than that of sandstone. Based on the evolution of radiation temperature with time under thermal excitation,the thermal excitation temperature change index was introduced,and its applicable conditions were clarified. Under the experimental conditions employing a specific heat source for active heating and cooling excitation,the threshold for distinguishing bituminous coal from roof sandstone was established and verified experimentally. Meanwhile,in view of the low illumination and high dust concentration in underground mines,a coal-gangue recognition scheme was proposed,which involves arranging a purification constant-temperature regulation zone,a heating zone,a coal-gangue recognition monitoring zone,and a sorting zone on the belt conveyor. This provides a new technical approach for intelligent coal-gangue recognition in mining operations.