UnitX has published a case study detailing an OptiX Large FOV vision system, running on the company's CorteX AI platform, that automates paint-surface inspection on vehicle rooftop panels for an unnamed automotive manufacturer. According to UnitX, switching from manual inspection to AI analysis cut the client's scrap and rework rate from 18 percent to 1.6 percent.
The rooftop is one of the industry's toughest "Class A" surfaces to inspect: it's large, glossy, and prone to defects like orange peel, solvent pop, pitting, and fine scratches that manual inspectors catch inconsistently. UnitX's setup uses four OptiX Large FOV units per inspection cell, each built around a 20-megapixel camera paired with rectangular bar lighting, to image the full panel. CorteX processes that imagery and automatically adjusts to black, silver, and white finishes without manual recalibration between production runs. Engineers running mixed-finish paint lines will care about that detail specifically, since color changes are typically where automated inspection setups need retuning or fail outright.
The system runs across three sequential quality gates, called Panel Stations 1 through 3, giving the line 100 percent inline verification before panels leave the paint shop. UnitX says inspections complete in 30 seconds, well under the client's 75-second cycle time target. Across all three stations, the company reports a zero percent false acceptance rate, meaning no defective panel was cleared as good. Station 2 posted a "perfect gate," with zero percent false acceptance and zero percent false rejection. Stations 1 and 3 held false rejection rates at or below 2.54 percent and 3.1 percent, respectively.
The figures come from a UnitX-authored case study rather than an independent audit, and the release does not name the manufacturer, plant, or vehicle program involved. That's typical for machine-vision vendors marketing into automotive, where OEM contracts often carry confidentiality terms that keep customer names out of public case studies.
