Skip to content
Hubs

Computer vision hubs, stage by stage.

Each hub answers one big question, then goes deep on every step, with links to the blog and the docs.

Full hub

Image Annotation for Computer Vision

Image annotation is adding labels to images so a model can learn from them: a class for the whole image, a box around each object, or an outline of its exact shape. Good annotation starts before the first label, with clear classes and written guidelines, and ends with review, versioning and an export in the format your model expects.

1 · Define the task and the classes2 · Collect and curate images3 · Write labeling guidelines4 · Choose the label type+6 more
Read the hub
Outline

Automated Visual Inspection

Automated visual inspection uses a trained model to find defects in images of parts or products. It is computer vision in manufacturing at its most practical. Start by framing the question, capture images that match the line, use anomaly detection when defects are rare, and prove the model on held-out images before it decides anything.

1 · Frame the inspection question2 · Capture: camera, lighting, angle3 · Good parts first: anomaly detection4 · Label defects+6 more
See the outline
Outline

Drones for Solar Panel Inspection

Drones for solar panel inspection fly a radiometric thermal camera and an RGB camera over a solar site, finds modules running hotter than their neighbours, locates them to the panel, and ranks them by severity in a report technicians can act on.

1 · Plan the inspection2 · Check the conditions3 · Fly the pattern4 · Check the photos+6 more
See the outline
Outline

PCB Inspection Methods

PCB inspection checks assembled boards for missing, wrong, misaligned or badly soldered parts. Automated optical inspection (AOI) machines do it inline after programming; golden-board methods compare each board with known-good references; AI helps cut false calls and programming time.

1 · Where inspection sits on the SMT line2 · Defect types3 · AOI, AXI and manual inspection4 · Programming: CAD, library, golden board+6 more
See the outline
Outline

Train and Evaluate a Vision Model

Training a vision model starts with a clean train and test split and ends with honest evaluation: precision and recall, the F1 score, a confusion matrix and mean average precision (mAP) for detection. Compare versions on the same held-out images before you switch.

1 · Split train and test2 · Augment the data3 · Object detection vs image classification4 · Train+6 more
See the outline