Development of an innovative AI solution for the detection of car damage

TÜV Rheinland and Adomea have developed a system for the automatic assessment of vehicle damage that produces a detailed image of the vehicle surface in just one minute. The high-precision measurement data is not only based on RGB images, but also uses reflection and curvature information to detect the smallest changes on the car body surface at the micrometer level. Adomea's measurement setup generated a database that offered high potential for the use of computer vision to further automate the damage assessment workflow.

Merantix Momentum was involved to detect defect classes such as scratches, abrasions, dents, etc. on the car body surface and set new standards in vehicle damage detection. The solution is later also suitable for various use cases - e.g. end-of-line inspection, vehicle logistics, lease return or other applications requiring precise assessment.