Abstract: A binocular vision detection and positioning method is proposed to solve the problem of post-processing of hole support residue in metal 3D printing parts. In this paper, the measuring system detects the hole position of the workpiece based on the arc ellipse recognition algorithm, calculates the pose of the hole position to be machinized in the optical center coordinate system of the main camera of the binocular module, and converts the pose to the base coordinate system of the robot arm on the machining site, providing coordinate information for off-line programming and automatic machining. Firstly, the binocular camera is calibrated and the measurement accuracy of the hardware system is checked. Then, the position ellipse features of the complex workpiece to be processed are extracted, and the left and right image ellipses are collected and binocular matching based on the pole-corrected images. Finally, the coordinates of matching point pairs are calculated according to the principle of multi-view, and the spatial ring with coordinate information is output. The experimental results show that the measurement error is less than 0.20 mm. The results show that the maximum error of the system is less than 0.84%, the error of the center space is less than 0.50 mm, and the maximum error of the attitude of the ring is less than 0.5°.