Skip to content Skip to content

AI 301: Machine Vision at the Edge

Access the full course

Full course content (video lessons, exercises, downloadable files, and QR references) is part of Pro.

The preview above gives you the full course structure, learning outcomes, prerequisites, and what you will be able to do after completion. Free for everyone.

  • ✓Video lessons with real 3D printing setups
  • ✓Calibration parameters
  • ✓Downloadable STL examples + gcode profiles
  • ✓AI tutor for questions
  • ✓Completion certificate
See Pro →

Defects in 3D printing cost time, material and reputation. AI edge vision enables real-time detection without cloud latency. This course builds a working pipeline – from camera to automatic print stop.

What you will learn

  • What edge inference is and why a cloud alternative does not fit
  • Hardware – Raspberry Pi, NVIDIA Jetson, Coral TPU, ESP32-CAM
  • Computer vision foundations – convolutional networks for defects
  • Defect classes – spaghetti, layer shift, under-extrusion, warp
  • Data collection and labeling strategy for a small dataset
  • Transfer learning from pretrained models (MobileNet, YOLO)
  • Integration with printer firmware for an automatic stop

Course outline

  • Lesson 1: AI in 3D printing – realistic expectations
  • Lesson 2: Edge hardware overview and benchmark
  • Lesson 3: Camera, lighting and mounting
  • Lesson 4: Dataset – capturing and labeling defects
  • Lesson 5: Transfer learning on MobileNet/EfficientNet
  • Lesson 6: YOLO for real-time detection
  • Lesson 7: Edge-device deployment and optimization
  • Lesson 8: Integration with OctoPrint, Klipper or native firmware

Prerequisites

A completed course-iot201-arduino-esp32 or equivalent experience with microcontrollers. Basic Python. No mandatory prior ML experience – the course introduces the concepts gradually.

Outcome

You deploy a working defect-detection model on an edge device that monitors the print, reacts to common failure modes and logs history for quality analysis.

Who it is for

Technology engineers in small and mid-sized print farms, printer-software developers, students of applied AI, and service-bureau owners seeking a competitive edge.