Research And Experiment Of Sheep Carcass Robotic Cutting System Based On 3D Point Cloud

The realization of automatic cutting of sheep carcasses is of great significance for improving the efficiency of slaughtering and process.Due to the lack of automatic cutting technology,at present the cutting of sheep carcasses in most enterprises in China are generally based on the production method of artificially assisted assembly lines,with low production efficiency and large number of workers being its major existing issues.In view of the above-mentioned status quo,this subject puts forward and designs a robotic sheep carcass cutting system based on 3D point cloud,which combines point cloud with robotics technology and cooperates with the profiling clamping device to realize the automatic cutting of sheep carcasses.The main research content and conclusions of this subject are as follows:(1)This subject studies the 3D point cloud segmentation of sheep carcasses.First,we use reverse modeling technology to obtain samples of sheep carcass point cloud;then,we preprocess the obtained point cloud samples with removing outlier noises and simplifying the number of points;finally,we introduce a sheep carcass point cloud segmentation method based on surface concavity.This method divides the sheep carcass 3D point cloud into hexads through following steps: supervoxelizing the point cloud,judging the convexity of common edges,setting evaluation functions of scores,and confirming the segmentation plane.The average accuracy of the segmentation,the average recall rate,and the average of the F1 value and the average overall accuracy rate were 0.923,0.913,0.918,0.921,respectively,and the average time for processing a single whole sheep carcass point cloud was 18.82 s.Test of multi-split combined point cloud segmentation,test of multi-body sheep carcass point cloud segmentation and comparison test of multi-segmentation algorithm show that the method of this subject has good segmentation accuracy and real-time performance,better comprehensive segmentation ability.(2)This subject designs and trial-manufactures profiling clamping device for sheep carcass robotic of cutting use.To meet the need of maintaining stability of sheep carcasses during the robotic cutting and of keeping the cutting parts clamped firmly,what is therefore designed profiling clamping device that can absorb and clamp an upside-down sheep carcass,which contains many parts,including adsorption module,clamping module,profiling guide rail.What’s worth mentioning is that the clamping module and the profiling guide rail are designed referring to the curve of sheep carcasses.The adsorption module is used to absorb backbone of the upside-down sheep carcass,and the clamping module is used to fix both sides of the sheep carcass with the module able to move on Z axis to the cutting position following robotic action.While the robot is cutting the carcass following cutting sequence,the whole sheep carcass can maintain stable,and the cutting parts well clamped with through the cooperation of the various parts in the device.(3)This subject designs and constructs sheep carcass segmentation control system.To meet the requirements of coordination between clamping device and robotic action in sheep carcass cutting process,control program of this system is adopted.In the aspect of the profiling clamping device,the hardware selection of each actuator is carried out,and the control program is designed based on the idea of sequential control.On this basis,the human-machine interaction interface is developed,and the automatic/manual operation of the clamping device can be selected through the touch screen.In terms of robot control,the robot hand-eye coordinate transformation is carried out,and the robot motion control is realized based on the robot offline control software.(4)This subject conducts sheep carcasses robotic cutting experiments.By integrating the method of point cloud segmentation,control of robotic motion and profiling clamping device,this subject forms sheep carcass robotic cutting experimental platform,and carries out robotic cutting experiments of whole sheep carcass.The experimental results show that this system achieve the desired effect with 1.53%segmentation loss rate,59.3 s cutting time and 10.2 mm average cutting offset distance.