Research on automated segmentation for pig carcass of robot system

At present, the proportion of semi-mechanized production methods in China’s meat cutting industry is very high, and the boning operation can only be completed manually. In order to solve the problems of low efficiency, easy pollution and high cost in traditional meat cutting industry, a set of fine automatic cutting robot system was developed. In this system, X-ray machine is used to obtain the position outline information of pig carcass bones, and depth camera is used to obtain the information of pig carcass surface elements. The obtained pig carcass data is analyzed and processed based on deep learning, and the coordinates of the grasping position and the key points of the cutting path are obtained to guide the robot to complete the fine cutting. The end flange of the cutting robot is installed with a self-developed adaptive cutting tool, which can autonomously avoid the dead point position of the bone, compensate for the deviation between the ideal cutting path and the planned path, and correct the cutting path to a certain extent, making the cutting process more smooth and efficient. Based on workbench, the strength analysis of the key parts of the cutting tool was carried out, and the Von Mises stress was 79.921 Mpa, which met the strength requirement. After testing, the ability of a single set of cutting systems to handle pig carcass is 90 pieces/hour and the loss rate is less than 5%.