An Image Measurement System for Detecting Dents and Scratches on the Surface of Used Car Body Parts

As abandoned cars have become a huge social problem, procedures to reuse them are urgently required. Before reusing used car body parts, it is necessary to inspect them for defects such as dents and scratches. Currently, most companies around the world perform manual inspections. Manually detection of these used cars is time consuming and over-relies on individual judgement. In this paper, an automatic used car body parts detection system is proposed based on image measurement. A two-times shooting technique is taken by this system to distinguish the dent and scratch on the car body surface, and at the same time, a neural network-based target detection method is proposed to locate the position of dent. Furthermore, a scratch extraction method is also involved. The experiment result shows that the proposed method can be used to detect car body parts with reflective surfaces. And the two-times shooting technique reduces the time used to take images.