In sintering production, the trolley side plate of the sintering machine is prone to crack due to the large pressure and temperature difference. The appearance of crack will affect the quality of the sinter, and in severe cases, the side plate may break and cause production accidents. In this article, a camera is adopted to obtain the video image data of the side plate of the sintering machine in operation, and then the labelme is used to label and process the data. Based on this network recognition of DeepLabV3 +, a method of detecting cracks on the trolley side plate of the sintering machine is proposed at last. The results show that the method has a good recognition effect on the crack detection of the trolley side plate, which provides a new solution for the crack detection of the trolley side plate during the sintering process.