Research On Fish-shape Bait Workpiece Grabbing Based On 3D Machine Vision

With the deepening of national industrial structure adjustment and the transformation and upgrading of modern manufacturing industry,the intelligent robot grasping system has been widely used in modern factories.Nowadays,the demand for fish-shaped bait workpieces is increasing,and the process is usually in a disorderly state,which is an extremely typical demand scene for disorderly grasping.In response to this huge industrial demand potential,robots with vision systems have emerged.In order to reduce the problems of manual selection and placement in the production process of fish-shaped bait workpieces,and improve production efficiency,this paper proposes a grasping system for fish-shaped bait workpieces based on 3D machine vision.Through a six-degree-of-freedom robot carrying an intelligent camera to extract target information of fish-shaped bait workpieces in the industrial environment,the point cloud information is processed and fed back to the robot arm,and the robot arm accepts instructions to complete the grasping operation.The specific research contents are as follows:(1)Image acquisition module.A camera image acquisition system is built to obtain point cloud information of scene target objects.An improved point cloud filtering algorithm that adaptively removes discrete noise is proposed to remove noise from the workpiece point cloud information obtained by the image acquisition module,and the denoising effect of the improved algorithm is experimentally verified.(2)Image processing module.The point cloud is improved and voxel filtering is used to reduce the number of points.Point cloud segmentation is completed,and multiple point clouds are segmented into single point clouds as a whole.Point cloud stitching is completed using feature points,and the template point cloud is generated.The single object point cloud after point cloud segmentation is registered with the template point cloud using the iterative closest point(ICP)method.(3)This paper presents the design and implementation of a 3D vision-based robotic arm grasping system for fish-shaped baits.The system utilizes the feedback of image information acquired by a 3D camera to guide the robotic arm grasping system.The hardware modules of the platform were built,and the communication between the modules was established.The hand-eye calibration between the robotic arm and the 3D camera was performed.The grasping function was integrated,and a software control operation was designed using an upper computer.The grasping trajectory design and grasping experiments for fish-shaped baits by the robotic arm were completed,and it was found that the fishshaped bait grasping system with the 3D robot met the requirements for grasping fishshaped baits.