OdinVision delivers a fully automated 2D backlit machine vision QC system for automotive stamped body panels, replacing slow, error-prone manual inspection for missing holes across multi-surface sheet metal components. Powered by robust blob analysis algorithms, our system reliably detects missing punched holes on automotive metal panels and outputs real-time Pass/Fail signals to PLCs and robotic handling systems. Optimized for high-volume stamping lines serving automotive OEMs and Tier 1 suppliers across North America, this large field-of-view vision solution reduces false reject rates and meets stringent automotive manufacturing quality standards.
1. Customer Inspection Requirements
- Target Part: Automotive stamped sheet metal, sample dimension 321mm × 196mm, maximum compatible workpiece up to 4000mm × 400mm
- Core Task: Detect missing stamped holes across three vertical surfaces of metal panels
- Signal Output: Trigger a Fail signal for parts with missing holes, and a Pass signal for fully qualified parts
- FOV Demand: Large single field of view to minimize multi-shot stations and boost line throughput
2. Complete Vision Hardware Bill of Materials
- 5MP CMOS industrial camera
- 12mm fixed focal length industrial lens
- White backlight illumination (mounted beneath parts to eliminate surface reflection noise)
- OdinVision dedicated vision processor
- Industrial cables and I/O communication wiring kits
- Working distance: 150mm, effective field of view: 370mm × 300mm
3. Inspection Workflow & Core Algorithm Logic
- Image Capture: Robots position workpieces perpendicularly to the backlight module for consistent image acquisition.
- Blob Analysis: The algorithm extracts all circular hole feature regions and records standard hole count and coordinate positions from golden samples.
- Threshold Judgment: The system compares detected hole quantity against the golden template and highlights defect zones if holes are missing.
- Visual Result Display: Green cross markers show the center position of every hole present on Pass parts; red circular markers flag missing-hole areas on Fail parts.
- Industrial Communication: Pass/Fail results are transmitted to PLCs and robot controllers via Ethernet TCP/IP, RS232 serial ports, or digital I/O modules.
4. On-Site Installation & Calibration Guidelines
Optical Assembly Rules Align the camera, lens, and backlight on a coaxial vertical axis relative to the sheet metal surface to avoid reflection-induced false detections. The gap between the backlight and workpiece is adjustable to fit varying metal thicknesses.
Robotic Positioning Standards Maintain fixed shooting coordinates for identical part models. Design robot motion paths that avoid blocking the camera’s field of view. For oversized stamped components, split the inspection into segmented shot zones.
Multi-Protocol Compatibility Native support for TCP/IP, RS232, and digital I/O enables seamless integration with mainstream North American PLC and robotic brands.
5. Expandable Upgrade Options
Add 3D vision modules to measure hole diameter tolerance, hole position offset, and sheet metal warpage. Enable parallel multi-task inspection to detect scratches, indentations, and burr defects simultaneously.