Automated Separation of Reflections for Vehicle Damage Assessment

Di Yang -&&- Srimal Jayawardena

COMPUTER VISION AND ROBOTICS SERIES PhD Monitoring

DATE: 2013-05-23
TIME: 16:00:00 - 17:00:00
LOCATION: RSISE Seminar Room, ground floor, building 115, cnr. North and Daley Roads, ANU
CONTACT: JavaScript must be enabled to display this email address.

ABSTRACT:
This project aims to establish an automated system to assess vehicle damage, including the boundary damage and the panel damage. Until now, boundary damage has been successfully assessed by implementing an improved watershed algorithm. In comparison to boundary damage, panel damage is far more difficult to be detected because of influences of reflections, such as spectrum spots, mirror images, and illumination changes. In order to detect such damage, reflections must be separated. Unfortunately, separation of reflections is a highly ill-posed problem. It is impossible to properly address such problem without incorporating any prior information that requires the user to partially mark the reflective features in the image. To avoid user-assisted process, we managed to build up an automated reflective-feature detector by combining other weak reflective-feature detectors utilising AdaBoost approach. Subsequently, the prior information for the reflections separation process now can be automatically extracted by our reflective-feature detector.
BIO:
http://people.cecs.anu.edu.au/user/3497

http://people.cecs.anu.edu.au/user/3788