As the carrier of solar cell, the quality of silicon wafer directly determines the conversion efficiency of the cell. At present, in the process of cutting silicon rods into silicon wafers by the front-end process of silicon wafers, the cutting process will cause surface dirt, thickness, size and hidden cracks and other defects of silicon wafers. Therefore, after cleaning silicon wafers, it is necessary to detect the separation quality of silicon wafers and separate them with different qualities. Using 3D vision to carry out online detection of thickness and other sizes, it can improve product qualification rate, improve production efficiency, reduce production cost, and avoid defective silicon wafers entering solar photovoltaic modules.