ABSTRACT
The current management and maintenance of the in-service concrete bridges mainly depends on manual visual inspection, which cannot meet the increasing demand for inspection. Meanwhile, traditional appearance evaluation method for concrete bridge is subjective, and cannot reflect the technical condition of the bridge clearly. In this paper, we propose a computer vision-based method for the inspection and evaluation of concrete apparent damage in bridge pier. To be specific, an airborne imaging platform is used for image acquisition. Then several deep learning-based damage recognition models are designed to classify and segment the damage areas. Finally, by using a damage score calculation formula, a novel method for the evaluation of the concrete component apparent conditions is established. Besides, a field experiment was carried out on the concrete pier of a simply supported bridge, and a score heatmap describing the distribution of the damage on the entire pier was formed. This work has practical value to the management and maintenance of the concrete bridges in engineering, especially for the quantitative assessing of the bridge component condition.
