Industrial-AI-QC-Labeling-Traceability-System

AI enabled Industrial QC-LABELING-TRACEABILITY Automation System

Team Teenage Engineering

TEAM MEMBERS: Pavan K. & Adarsh Singh MENTOR: PROFS. BHARAT TANK

FIG 1 :

Industrial AI QC Labeling and Traceability System

A fully automated end-to-end solution for PCB labeling, visual inspection, and product traceability using AI, computer vision, and real-time logging.

VIDEO: 1 https://drive.google.com/file/d/1ATr-kjntGiZoPwIW6M74zLSkXyEIK4Zf/view?usp=drivesdk

Project Overview

This project is a practical industrial automation system built to streamline the PCB (Printed Circuit Board) manufacturing process. It automatically labels each PCB with a QR code, checks its quality using AI, and logs all data for traceability. The result is a faster, smarter, and more reliable quality control system.

CLICK HERE TO VIEW SCHEMATICS

FIG 2 :

Key Features

Automated Labeling System

• Custom-built labeler for applying 1x2 inch QR code stickers
• QR codes are generated with unique serial numbers
• Accurate and repeatable sticker placement using mechanical guides

AI-Powered Quality Control

• Computer vision-based real-time inspection
• OCR (Optical Character Recognition) to read component part numbers
• Pass/Fail status automatically assigned using a machine learning model

Product Traceability

• Complete lifecycle logging for every PCB
• Public verification available at verify.stromlabs.tech/qrid
• Firebase database integration for real-time product records
• 24x7 online product verification for customers and admins

Homepage :

Product Verification :

Enter ProductId as it is :

Product details visible :

Employee Login :

Enter Employee Details properly :

Webiste opens that provides various features which allows employee to perform various tasks or operations on data logs :

Webpage that opens on scanning the QR code given on the product label :

#

Mechanical Automation

FIG 3 :

• Conveyor belt with fine-tuned tension control
• IR sensors detect PCB presence for timing and control
• Raspberry Pi is used for controlling the system and also for loading models that help in image detection for QR and product verification using an camera connected directly to RPI.

System Architecture

Hardware Components

Component	Specification	Purpose
Frame	MDF Sunboard + 3D Printed Parts	Structural body of the system
Conveyor Motor	IG32 Motor + Cytron Driver	Drives the conveyor belt
Conveyor Belt	Cotton Fabric with ±15mm adjustment	Moves PCBs across the system
Structural Support	8mm Steel Rods + Flange Bearings	Provides mechanical stability
Processing Units	Raspberry Pi 5 + ESP32	Handles AI, control logic, and sensors
Vision System	RPi2Cam Module (NoIR)	Captures images for inspection and scanning
Rejection System	Servo motor-based actuator	Redirects faulty PCBs
Detection	IR Sensor	Triggers events when PCB is detected

FLOWCHART:

Software Stack

• Backend: Python for model creation , OpenCV for image detection and Firebase for Database management
• AI/ML: Custom trained model for PCB detection and TensorFlow or PyTorch for defect detection
• OCR: EasyOCR for reading component labels
• Database: Firebase Realtime Database
• Web Platform: Hosted on GitHub Pages for now on our custom domain stromlabs.tech
• Communication: ESP32 to Raspberry Pi 5 over serial/I2C

System Workflow

1. PCB Entry → Detected by IR sensor
2. Conveyor Stops → Label is applied
3. QR Code is Scanned → Serial number recorded
4. AI Quality Inspection:
   • OCR text verification
   • Visual defect detection
   • Component identification
5. Database Logging → All data stored in Firebase
6. Pass/Fail Classification
7. Product Routing:
   • PASS → Sent forward on main conveyor
   • FAIL → Removed from the converyor belt manually(for now).

Quality Control Parameters

Each PCB is inspected for the following:

• Device ID: Unique product ID assigned
• Batch ID: Tracks production batch for collective product data access
• Manufacturing Date: Timestamp for traceability
• RoHS Compliance: Environmental compliance status
• Serial Number: Encoded in the QR code for direct verification instead of manually entering data
• Visual Defects: Detected via custom trained AI model
• Component Verification: Checked using OCR

Traceability System

Web Verification Platform

• URL: verify.stromlabs.tech/qrid
• Public Access: Customers can verify their product directly via scanning the QR or manually entering the serial serial number on the verification website.
• Admin Panel: See full history and statistics of product data logs and also possible to download data to store on local computer
• Live Sync: Firebase keeps all data up to date

QR Code Integration

• Simple scan-to-verify system
• No manual entry of serial numbers
• Full manufacturing history available instantly
• Clean UI for both customer and admin views

Technical Specifications

Mechanical Design

• Material: MDF Sunboard and 3D printed joints
• Belt System: Adjustable conveyor with ±15mm tension control
• Labeling: High-precision mechanical alignment for precise labeling
• Rejection System: Manual for now but can be updated to automatic at later stages.

Control System

• Main Controller: Raspberry Pi 5
• Sensor Controller: ESP32
• Protocol: Serial/I2C/SPI communication
• Software: Multi-threaded Python application for real-time control

AI/ML Capabilities

• Image Processing: Handled by OpenCV (stands for Open Computer Vision) for direct image capturing
• OCR Engine: EasyOCR for small text detection
• Defect Detection: Custom-trained CNN (Convolutional Neural Network)
• Accuracy: >95% on known PCB defects

Performance Metrics

• Speed: 30–45 seconds per PCB (label + inspection)
• Label Accuracy: 90.5% placement precision (Achived Mechnically)
• Defect Detection: 95%+ accuracy
• System Uptime: 24/7 operation tested
• Verification Time: <2 seconds via web

Applications

• Electronics Manufacturing: Real-time QC in assembly lines
• Automotive Industry: Component-level traceability
• Medical Devices: Regulatory compliance checks
• Consumer Electronics: Product authentication
• Industrial Automation: Smart factory integration

Future Enhancements

• Edge Computing: Local AI inference to reduce latency
• Cloud Integration: Deployment on AWS or Azure
• Mobile App: Easy customer-side product scanning
• Predictive Analytics: AI to detect failure trends
• Multi-Product Support: Configurable for different PCB types
• Rejection System: using servo motors for auto removal of defected products.

Team & Contributions

This project showcases full-stack development in hardware + software:

• Mechanical design and 3D fabrication
• Embedded electronics and automation
• AI model training and image processing
• Database and cloud integration
• Web-based verification platform

By Pavan Kalsariya & Adarsh Singh