Automatic and Manual Fracture-Lineament Identification on Digital Surface Models as Methods for Collecting Fracture Data on Outcrops: Case Study on Fractured Granite Outcrops, Bangka

Abstract

Three-dimensional outcrop models, or Digital Surface Models (DSMs), have proved their capacity in many geoscience studies. Along with the advantage in the rapid acquisition, DSMs are capable of creating virtual models of fractured outcrops to be interpreted for further analysis. This paper reports the DSM robustness by comparing the result of fracture-lineament measurement using DSMs and discusses the possible causes of error that might occur. The first method applied in this study is the scanline method to collect fracture data directly from outcrops, measuring more than 1,400 fracture data. The second method is applying fully automatic and manual fracture identification by optimizing hill-shaded DSMs. Two well-exposed granite outcrops in Bangka, Indonesia, are designed for the pilot area. Structure-from-Motion (SfM) photogrammetry is utilized to generate the DSMs, where a series of aerial images are captured using Unmanned Aerial Vehicle (UAV). The images are then processed into hill-shaded DSMs to be automatically analyzed following the algorithm in PCI Geomatics software and manually assessed. The textures of DSMs are also used in fracture identification through RGB filtering as the third method. The results show that the semiautomatic measurement using RGB-filtering texture has the closest pattern to the scanline data compared to the hill-shaded DSM method. The differences rely on several conditions, such as the geometry and texture of the outcrops. Eventually, methods of fracture identification using DSM are expected to be capable as options in preliminary fracture data collecting on outcrops, especially when the scanline is unable to be performed.