Author(s): Shijun Pan; Keisuke Yoshida; Satoshi Nishiyama

Linked Author(s): Shijun Pan, Keisuke Yoshida

Keywords: Instance Segmentation; Riparian Crack Detection; River Monitoring; Smartphone

Abstract: In this paper, the authors present a simple free methodology that harnesses the potential of smartphone-derived photogrammetry 3D models for the AI-driven segmentation and quantification of riparian cracks. The approach the authors propose leverages the YOLOv7-seg model to precisely delineate cracks, employing photogrammetry techniques to construct an orthophoto, subsequently enabling the extraction and accurate measurement of crack dimensions. This research underscores the fact that the proposed methodology yields outcomes in both the detection and measurement aspects of riparian cracks. This investigation carries profound implications for the realm of environmental oversight and administration, as it heralds an innovative approach toward the identification and amelioration of issues pertaining to cracks in riparian asphalt surfaces. Through the rigorous analysis of this approach, not only crack detection and measurement but also holds promise for wider applications in the field of environmental assessment. It extends the boundaries of AI-driven methodologies in the context of environmental preservation.

DOI: https://doi.org/10.3850/iahr-hic2483430201-520

Year: 2024