Based on the research on the pore feature extraction of the sintering machine tail section, this paper proposed a multithreshold method based on OTSU three iterations of RGB color channels. In order to ensure the accuracy of stomatal feature extraction and reduce the loss of image details, this method not only adopted the OTSU method to calculate the threshold components of the RGB color channels of the tail section image for three iterations, but also determined the weights by the ratio of the RBG three color channels. Finally, a global threshold and two local thresholds were calculated to perform threethreshold image segmentation. The author used the obtained global threshold to segment the flaming red layer and the black ore layer of the crosssectional image of the sintering machine tail, and then used two local thresholds to respectively segment the pores embedded in the flaming red layer and the black ore layer. In the end, the author extracted the stomatal features of the tail section of the sintering machine. Experiment shows that compared with conventional methods, this method can extract more stomatal features and details of the tail section of the sintering machine.At the same time, the extracted stomatal feature contours are better and the accuracy of stomatal feature extraction is high.