A 13-chapter beginner tutorial that takes you from install to a running production inspection, plus a reference library of 59 component guides across 14 categories — all available in-app under Help ▸ Guide.
Thirteen chapters, in order — from first launch to a production line.
Install, log in, and take your first look. Core concepts and a map of what VisionBase Studio does.
The thin ribbon, floating Vision Windows, and the Tool, Sequence, and Control panels — plus the viewport and key shortcuts.
Hands-on: Threshold → Blob Analysis → Inspector on the simulated camera, with ROIs and a PASS/FAIL verdict.
Discover, connect, snap, and go live. Exposure, gain, and manual / timer / external trigger modes.
Piped vs source input, ROI shapes, fixtures for moving parts, breakpoints with step/retry, teach, conditional gates, and undo.
Which component to reach for — a decision guide across preprocessing, detection, measurement, and AI.
Turn measurements into verdicts with the Inspector rule engine; read yield, statistics, SPC, history, and exports.
The three-layer model — .vbproj, .vbseq, .vbrcp — portability, and recipe compatibility verdicts.
From pixels to millimetres: line, checkerboard, and multi-point / homography calibration and .vbcal for real-world units.
Continuous run, Operator mode, image logging and replay, scheduled reports, traceability, audit, and the live dashboard.
TCP, MQTT, OPC UA, Modbus, databases, MES, and MCP; digital I/O and the trigger bridge for PLC reject.
Collect, label, train, and infer — classifier, detector, and anomaly models — and let the assistant build a sequence.
Common problems, where the logs live, a shortcut cheat-sheet, a glossary, and how to get help.
No camera yet? The simulated camera plays back an image folder, so you can complete the whole tutorial before hardware arrives.
Every component ships with its own in-app guide — 59 guides across 14 categories, reachable from the Tool Panel context menu.
Gray conversion, threshold, edges, morphology, filters, masking, pixel mapping.
Blobs, template matching, barcodes, OCR, lines, circles, contours, defects.
Dimensions, the Inspector, calipers, colour comparison, brightness, focus.
Fixtures, calibration, and calibration from fiducials.
Classifiers, detectors, segmentation, anomaly, labelers, and trainers.
Heatmaps, point clouds, plane fitting, stereo, volumetric measurement.
Expression evaluator, conditional gate, accumulate ROI.
TCP, MQTT, OPC UA, Modbus RTU, SQLite, and file logging.
The camera system and Modbus TCP digital I/O provider.
Stuck? The built-in assistant answers from these same guides.
Press Ctrl + I (Engineer role and above) to open the assistant. It retrieves answers from the component guides on-device, can design a sequence from a plain-language description, and runs disciplined tuning experiments on the live station. Bring your own model — Claude API, an OpenAI-compatible endpoint, or an offline local model via Ollama.
The assistant cites this guide directly — it's part of the on-device knowledge base — so answers stay grounded in how VisionBase actually works.
Go deeper on the parts that matter to you.