- Industry > Metal·Plastics
- Application > Inspection
- Application > Loading·Unloading
- Sector > Manufacturing
- #3DVision
- #AIVisionInspection
- #BinPicking
- #IndustrialCamera
- #LiDARSensor
- #LogisticsInspectionAutomation
- #MachineVision
- #MachineVisionCamera
- #MechMind
- #MetalParts
- #SortingAutomation
- #VisionCamera
- #VisionInspectionSystem
- #VisionSensor
Process Overview
This solution is an automated feeding system for automotive parts. It combines industrial 3D cameras and AI vision algorithms from Mech-Mind, a global leader in 3D vision. Its core hardware, the Mech-Eye PRO M-GL 3D camera, works with deep learning algorithms to enable 3D bin picking: the robot precisely recognizes highly reflective metal CV joints (constant velocity joints) piled randomly inside deep, large bins and transfers them to the next process.
The system was designed to overcome the complex geometries and surface reflections of metal parts, which conventional 2D vision or 3D sensors struggle to handle. This makes efficient robotic automation possible even in harsh manufacturing environments. Its key strength is how three things work together so parts can be fed unmanned and without interruption: high-speed scanning with a wide field of view (FOV), AI object recognition that filters out surrounding interference such as plastic film, and intelligent collision-avoidance motion control.
Project Background and Purpose
The automotive parts industry faces several challenges. Heavy metal parts are handled by high-intensity manual work, which causes worker fatigue, risk of musculoskeletal disorders and bottlenecks in part supply. Manufacturers previously installed expensive mechanical alignment devices such as part feeders to address this. However, these devices lacked flexibility and took up a lot of floor space. This project was carried out with the following objectives:
Fully automate heavy, high-intensity, repetitive feeding tasks
Precisely recognize highly reflective metal materials and parts with complex curved surfaces
Pick reliably in environments with heavy optical noise, such as tangled parts and plastic film
Build a flexible supply process that runs in tight spaces without mechanical alignment devices
Achieve unmanned operation through a high bin-emptying rate
The system goes beyond simply adding a vision sensor. It is an intelligent robotics solution that links visual data acquisition, AI analysis and robot motion control into one seamless workflow.
Components
- Robot
Mech-Eye PRO M-GL Industrial 3D Camera
Extended field of view (FOV) that covers an entire large bin
Anti-reflection optics that suppress strong light reflections from machined metal parts
Precise, fast 3D scanning that is robust against ambient light
- Peripherals
AI Vision Software (Mech-Vision & Mech-DLK)
Analyzes captured 3D point cloud data in real time
Uses deep learning to accurately separate overlapping parts from the plastic film at the bottom of the bin (segmentation)
Precisely estimates the 6-DoF position and orientation of complex-shaped CV joints
Intelligent Robot Motion Control Software (Mech-Viz)
Generates optimal robot trajectories based on coordinates from the vision system
Automatically calculates paths that predict and avoid collisions between the robot arm, the gripper and the edges of large bins
Lets users program robot motions through an intuitive graphical UI, without writing code
Workflow
- STEP 1.
- 3D scanning and data acquisition
- STEP 2.
- Point cloud generation and AI recognition
- STEP 3.
- Optimal picking strategy and trajectory planning
- STEP 4.
- Robot gripping and transfer
- STEP 5.
- Repeated cycles and bin emptying
※ The content provided by Bigwave Robotics is protected under U.S. and Canadian copyright and intellectual property laws. Unauthorized reproduction, distribution, or use of this content is strictly prohibited.
