A coal gangue near-infrared image identification technology belongs to the image processing technology field, is characterized by being an object identification technology under the near-infrared vision, and comprises the concrete steps of 1) extracting the gray value information within objects one by one, wherein the step comprises utilizing a Canny edge detection operator to extract the edges of the objects and extracting the gray value information within the edges separately; 2) gathering and analyzing the gray value information within the edges separately; 3) determining the objects. The algorithm analyzes the objects one by one based on extracting the edges of the objects, effectively avoids the interference of the background with the identification results, increases the difference between the characteristic intervals of the coal gangue and the coal under the near-infrared vision, still can guarantee the robustness of the algorithm even on a worse visual condition, and can identify the objects accurately.