Automated assessment of steel bridge paint condition using combined infrared and computer visions

Abstract

The assessment of steel bridge paint condition is often subjective because of its reliance on visual inspection methods. This study presents the development of an automated paint condition assessment (APCA) system designed for objective and quantitative evaluation of steel bridge paint condition. The APCA system integrates sensing, onboard processor, and unmanned mobile robot modules. The sensing module utilizes an invisible line laser beam to heat the inspection surface, while infrared and vision cameras capture thermal and vision images. The data processing module subsequently visualizes and quantifies the paint thickness and defects, providing automated paint condition ratings in accordance with current paint rating guidelines. The sensing module is mounted onto an unmanned mobile robot module equipped with magnetic wheels. The APCA system is deployed onto a steel bridge, where it is maneuvered across inspection areas using a remote controller. This study offers the following uniqueness: (1) automated detection, classification, and quantification of paint defects; (2) quantification and visualization of paint thickness; (3) capability to inspect areas that are difficult or hazardous for manual assessment; and (4) automatic paint condition rating aligned with existing paint rating guidelines. The performance of the proposed APCA system was validated through field bridge and specimen testing.