Structured Light Device and Algorithmic Enhancements for Automated Bridge Steel Bar Weld Inspection

Abstract

The structural integrity of steel bar welds is paramount for the operational efficiency of highway infrastructure. Traditional assessment techniques, primarily manual visual inspections and conventional measuring tools, are characterized by notable deficiencies in precision, reliability, and efficiency. Given the increasing demands for safety inspections within China's vast highway network, these traditional methods are insufficient to meet the contemporary requirements for rapid testing. To bridge this gap, this research introduces an innovative automated visual inspection system and associated technology specifically designed for steel bar welds. By leveraging the finite element method, the study simulates weld defects and develops sophisticated image preprocessing and enhancement algorithms, along with an improved version of the traditional Normalized Mean Square Error (NMSSE) algorithm. Experimental outcomes from both controlled laboratory environments and field validations demonstrate the effectiveness of the proposed methodologies in replacing traditional inspection techniques, providing rapid and accurate evaluations of weld quality. These advancements not only establish new technical benchmarks for the morphological assessment of infrastructure surfaces but also significantly enhance the maintenance and safety evaluations of aging bridge infrastructures, thereby contributing to the overall resilience and safety of highway systems.