Designing imaging cells that catch the defect you care about — without tripping on the ones you don't.
Sample part study
01Real parts imaged in our lab under candidate lighting, optics and algorithms to prove the defect can be separated from noise.
Lighting design
02Bright field, dark field, dome, coaxial, backlight — selected and modelled around surface, finish and contamination type.
Optics & sensor spec
03Lens, working distance, depth of field and sensor resolution chosen for the smallest defect at full line speed.
Algorithm architecture
04Rule-based, blob, edge or deep-learning — selected for repeatability, false reject rate and audit traceability.