Software · 2026

Precision dimensioning from multi-angle images

A Python measurement pipeline that turns five camera views of a part into dimensions with stated uncertainty — and a DIN-style drawing.

48.02 ±0.0320.00 ±0.0232.01 ±0.03⌀6.01 ±0.02V1V2V3V4V5U = 0.012 mm (k = 2) · GUM5 VIEWS FUSED · SUB-PIXEL EDGESSCALE 2:1 · DIN 406AUTO-GENERATED DRAWING
Illustration. Replace with project photos via assets/config.js.
Client
Industrial quality control, Germany (confidential)
Role
Software & measurement engineering
Tools & stack
Python · OpenCV · NumPy · point clouds · VS Code
Status
Delivered

The problem

Dimensional checks were slow and operator-dependent. The client needed measurements from images that could stand next to a gauge reading — with an uncertainty, not just a number.

What we built

  • Image and point-cloud loading, preprocessing, and PCA-based alignment of the part
  • Sub-pixel edge measurement, calibrated against a pixels-per-millimetre reference
  • Fusion of five views into one result with a GUM-compliant uncertainty budget
  • Automatic output as a DIN-style technical drawing with dimensions and tolerances
  • Packaged as a VS Code workspace with debug configurations and setup scripts, so the client's engineers can run and extend it

Result

Repeatable measurements in seconds, every value carrying its uncertainty — traceable enough for a QC record, fast enough for the line.

For businesses that need it to work in the field, not in a demo.

Plantations, factories, workshops and infrastructure operators across Indonesia — and product companies in Europe. Tell us the problem; we will tell you honestly whether we are the right team.