Automated inspection of concrete bridge components is crucial for timely maintenance and safety. This paper presents
an algorithm for automated crack severity assessment in simply supported bridge beams using UAV-based video analysis. The
proposed method integrates deep learning for crack detection with image processing techniques to measure crack dimensions and
introduces a novel severity metric that combines the crack’s mean width with its location along the beam. This addresses a gap in
existing approaches which typically evaluate crack severity only by width or length, without considering structural context.