Demo Catalog
Maintenancedemo-0015

Computer Vision Quality Inspection

Camera inspects products for defects in real time. YOLO detection plus a vision language model provides pass/fail verdicts with natural language defect descriptions.

gpucomputenvidiahpaihpzgxn001in-progress
Request This Demo
Run Count0
Contactspainter
Student MaterialsPublic Repo
Admin / CI-CDPrivate Repo

Overview

A two-stage inspection pipeline: YOLO runs at frame rate for fast object detection, while a vision language model (LLaVA) provides detailed defect analysis every 2 seconds with natural language descriptions of what's wrong.

What You'll Learn

  • Computer vision for manufacturing quality control
  • YOLO object detection + vision language model pipeline
  • Reference image comparison for defect detection
  • Real-time inspection statistics and pass rate tracking

Key Capabilities

  • Real-time PASS/FAIL/REVIEW verdicts with confidence scores
  • Natural language defect descriptions (not just labels)
  • Reference image upload for comparison-based inspection
  • Inspection statistics dashboard with pass rate tracking
  • Image upload mode for offline inspection