Automated fracture aperture quantification in rock outcrops using image processing technique: The OpenFraQ algorithm

This is the authors' abstract. We don't add key points for this paper.

Understanding and quantifying rock fractures is one of the main goals in mining, geotechnical engineering, and engineering geological assessment. This study presents OpenFraQ, an automated image processing algorithm designed to extract and analyse open fracture aperture and relative geometries from 2D images. Unlike previous automatic or semi-automatic image processing approaches that only represent fractures as trace lines or segmented points, OpenFraQ introduces a novel geometric method that characterises fractures as both contour areas and overall traces, allowing discrimination between open fractures and non-open cracks. Using contour-based segmentation, edge detection, and width estimation methods, OpenFraQ effectively detects complex and irregular fracture traces, providing reliable estimates of the fracture aperture and tracing geometry. Validation against manual tracing and Hough transform confirms its accuracy and efficiency, particularly in processing large datasets and challenging discontinuity networks. Across test samples, OpenFraQ achieved an average precision of 63.34%, a recall of 63.05%, and an F1-score of 62.92%. The algorithm offers a practical tool for rapid geometrical fracture characterisation, enhancing our understanding of fracture networks, which serves as the starting point for further engineering geological analyses, such as stability assessment, stress and strain distribution, and rock mass permeability evaluation.