Visual Inspection and Rejection System is an intelligent quality control device based on machine vision technology. It captures product images using industrial cameras, analyzes defects with image processing algorithms, and automatically rejects non-conforming products via linked actuators.
Visual Inspection and Rejection System is an intelligent quality control device based on machine vision technology. It captures product images using industrial cameras, analyzes defects with image processing algorithms, and automatically rejects non-conforming products via linked actuators. System Core Components The Visual Inspection and Rejection System typically consists of a Vision Inspection Unit and a Rejection Execution Unit. Both are controlled by an industrial computer, forming a closed-loop "detection-judgment-rejection" process.
Specification Parameter
Module
Components and Functions
Vision Inspection Unit
Industrial Camera: High-resolution (5MP~20MP), high-speed (≥300fps) area array/line scan cameras, adaptable to different product sizes and line speeds.
Lighting: Customized lighting (e.g., ring light, coaxial light, backlight, UV light) to highlight foreign object/label features.
Lens: Telecentric lens (to eliminate distortion) or wide-angle lens (for large field of view coverage).
Image Processing System: Industrial computer + vision algorithms (OpenCV, deep learning models) for defect identification and coordinate localization.
Trigger Module: Encoder/photoelectric sensor (synchronizes with line speed, triggers camera capture).
Rejection Execution Unit
Pneumatic Rejection Device: High-speed air valve (response time ≤10ms), suitable for lightweight/small products (e.g., bottled beverages, PCB boards).
Collection Device: Rejected product chute/bin, separated from the acceptable product channel.
Product Application
The equipment can achieve detection of various defects such as tablet defects, chipped corners, cracks, black spots, scratches, dents, foreign matter, mixed tablets, cracks, and pockmarks, meeting GMP requirements. Foreign Object (Oil Stains, Metal Shavings, Hair) Detection and Label (Sleeve/Adhesive Label Misalignment, Missing Labels) Detection
Product Feature
1
Camera Resolution & Field of View Ratio: E.g., To inspect a 500mm×300mm product surface, camera resolution ≥2592×1944 pixels is required (single pixel coverage ≤0.2mm).
2
Compensating for Production Line Vibration: Real-time synchronization via encoder for camera triggering to avoid motion blur (shutter speed ≤1/5000 sec).
3
Label-Product Conformance Modeling: Pre-capture images of qualified labels as benchmarks, recording features like edge coordinates, area, aspect ratio.
4
Dynamic Compensation: Adjust camera exposure time based on line speed (e.g., shorten exposure for high-speed scenarios to prevent motion blur).