Overview
What this challenge is about.
You will design and prototype a calibration workflow using a printed ChArUco board (a chessboard with embedded ArUco markers). You receive a sample dataset of 200 raw frames per camera plus the manufacturer's nominal intrinsics. Calibrate each camera individually (intrinsics + lens distortion), then estimate the relative pose between the four cameras (extrinsics). Success means a reprojection error below 0.5 pixels per camera and a documented per-camera-pair rotation/translation error below 1 degree and 5 mm respectively. Wrap the pipeline in a single Python script with a printable PDF runbook for the field technician.
The Brief
What you'll do, and what you'll demonstrate.
Build a 15-minute, technician-runnable calibration pipeline that brings a four-camera rig back to sub-pixel reprojection accuracy and sub-degree inter-camera pose error.
Earning criteria — what you'll demonstrate
- Apply the pinhole camera model and lens-distortion coefficients to real imagery
- Estimate camera extrinsics from shared fiducial observations across views
- Validate calibration quality via reprojection error and pose-consistency checks
- Translate a research-grade pipeline into a field-deployable tool
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Machine Learning Engineer
Packaging a perception preprocessing step as a reliable, reproducible pipeline mirrors how MLEs ship feature pipelines: deterministic outputs, error budgets, and runbooks for ops.
This challenge sharpens
- python
- tooling-design
- image-processing