Course Objective & Scope
On production lines, defective samples are scarce, new defect types appear unexpectedly and cycle times are measured in milliseconds. Starting with optical setup (camera, lens, lighting, triggering), this programme covers both supervised defect detection (YOLO, segmentation) and unsupervised anomaly detection trained solely on good parts (anomalib: PatchCore, PaDiM, EfficientAd).
Pixel-level anomaly maps are built on metal, textile, automotive, food and electronics (PCB) examples, OK/NOK thresholds are calibrated and PLC/SCADA integration is examined. Participants leave with an end-to-end technical design and prototype for an inspection station on their own line.
Who Should Attend?
- Computer vision engineers building inspection systems for factories and production lines
- Industrial, quality and mechatronics engineers working in automation and R&D
- Technical teams leading smart-factory (Industry 4.0) transformation projects
- Prerequisite: Python knowledge and basic familiarity with image processing or machine learning
Key Takeaways
- Specify the right lighting (dark-field, back-light, coaxial) and camera setup for a defect type
- Train anomaly models from defect-free samples alone using PatchCore and EfficientAd when faults are rare
- Catch surface scratches, cracks, stains and missing assembly parts with segmentation and detection models
- Calibrate decision thresholds against overkill (false reject) and escape (false accept) rates
- Run inference within line cycle times and signal PLCs over OPC-UA or Modbus
Curriculum
Industrial Imaging, Optics and Lighting
- Area-scan and line-scan cameras, global shutter, lens and working-distance calculations
- Lighting techniques for reflective metal, glass and textured surfaces, plus hardware triggering
Classical Metrology and Template Matching
- Sub-pixel edge detection, dimensional/angle measurement and calibration
- Fast presence/absence checks and Region of Interest (ROI) alignment
Supervised Defect Detection and Surface Segmentation
- High-resolution segmentation for scratches, burrs, holes and stains
- Synthetic defect generation and tiling strategies for tiny flaws
Unsupervised Anomaly Detection (anomalib)
- Training on good (OK) parts only: PatchCore, PaDiM and EfficientAd
- Pixel-level anomaly heatmaps and AUROC / F1 threshold optimisation
Line Integration, PLC Communication and Continuous Improvement
- Overkill versus escape trade-offs, cycle time and deterministic inference
- GigE Vision, PLC integration via OPC-UA/Modbus and active field data collection
Other Courses in This Category
-
Inpatient fall prevention, ICU monitoring, hand-hygiene and PPE compliance, operating-room instrument counting and privacy-preserving edge anonymisation, for hospital digital transformation and healthtech teams.
- Live Course
- 5 modules
-
DICOM/NIfTI processing, 2D/3D lesion and organ segmentation on CT/MRI/X-ray with MONAI and nnU-Net, whole-slide pathology (WSI) analysis and clinical validation, for medical imaging and biomedical AI teams.
- Live Course
- 5 modules
-
Reception and lobby queue/crowd analytics, buffet tray replenishment and food-waste measurement, car-park ANPR, pool/facility safety and privacy-compliant anonymous analytics, for hospitality technology teams.
- Live Course
- 5 modules
-
Model conversion with ONNX Runtime, TensorRT, OpenVINO and TFLite, INT8 quantization, camera pipelines on NVIDIA Jetson and Raspberry Pi, GStreamer and OTA updates, for teams running vision models at low latency in the field.
- Live Course
- 5 modules
-
Knowledge distillation, structured pruning, Quantization-Aware Training (QAT), NAS and Torch-TensorRT profiling, for engineers shrinking and accelerating large computer vision models with minimal accuracy loss.
- Live Course
- 5 modules
-
Computer Vision
Comprehensive Computer Vision training ranging from classical image processing techniques to PyTorch-based CNN, Vision Transformer (ViT), Swin Transformer, and ConvNeXt architectures. Hands-on coverage includes object detection with YOLO and DETR, segmentation with SAM and U-Net, video analysis, 3D vision (NeRF, Gaussian Splatting), Diffusion models, and model deployment with ONNX/TensorRT.
- Live Course
- 13 modules
Get in touch
