Vision lighting and camera placement: internal test note
An internal test of lighting geometry before model training for dark, glossy parts.
- Owner
- content-team
- Reviewed
- 2026-09-12 · Next review 2026-12-12
- Evidence level
- Internal test
- Citation scope
- Cite only as an internal-test method note; do not use it as a production accuracy or ROI claim.
Hypothesis
Controlling reflection with a repeatable LED and camera geometry will make shallow defects observable before model training.
Setup and hardware
Dark glossy injection-moulded parts on a fixed inspection bench; one part at a time and a human-reviewed label set.
Industrial camera · LED bar · Jetson Orin Nano Super · Custom YOLO · TensorRT
Method
Capture the same part under candidate angles, record lighting and working distance, then compare the reviewer's defect visibility before training.
Result
The geometry exercise exposed lighting as the first decision. No publishable precision, recall or throughput number has been approved.
Failed attempts
- A flat front light produced glare that hid shallow sink marks.
Limitations
- Lighting and camera geometry must be re-tested for each installation.
- The internal set is not a production acceptance set.
Next steps
- Create a site-specific capture protocol and label a balanced acceptance set.
- Repeat the test at the target line speed before comparing models.
Citation scope
Cite only as an internal-test method note; do not use it as a production accuracy or ROI claim.