Research paperAutomatic fracture detection based on Terrestrial Laser Scanning data: A new method and case study
Introduction
Fractures widely develop in rocks of various lithologies and are important for understanding the deformation mechanism (e.g., stress-strain property) in a specific area (Olson et al., 2009, Strijker et al., 2013, Zeng et al., 2012a). They reduce the rock strength by breaking the internal coherency, inducing geological hazards (e.g., landslides and debris flow) and increasing risk of collapse of buildings on the rock. On the other hand, however, they can form good reservoirs in tight sandstones, such as the Paleogene Ganchaigou Formation in the Qaidam Basin of western China (Feng et al., 2013, Li and Wang, 2001), and therefore play a significant role in oil & gas exploration (Gale et al., 2007, Olson et al., 2009). Recent studies of fractures are generally based on direct measurements in the field (Awdal et al., 2016, Watkins et al., 2015, Su et al., 2014), analysis of borehole loggings (Lacazette, 2009, Prioul and Jocker, 2009) and signal processing of high-resolution seismic data (Hart, 2006, Lohr et al., 2008, Masaferro et al., 2003). Among them, field measurement is the most prevalent way of high precision and accuracy, but usually time-consuming, dangerous and limited by reach of measurement. A new way of extracting fracture-related information (e.g., fracture geometries, surface density) efficiently, safely and precisely is quite necessary. Currently, non-contact measuring techniques, such as photogrammetry and LiDAR (Light Detection and Ranging), provide alternative approaches to in-situ measurement of fractures from high resolution images and 3-D point clouds of rock mass exposures.
The two-dimensional (2-D) image detection method (photogrammetry) extracts fractures according to changes of pixel intensities, and can be implemented both manually and automatically. Image auto-detection based on edge detection algorithms (Bao et al., 2015, Canny, 1986, Lopez-Molina et al., 2013, Sun et al., 2016) is faster and more objective than the manual one. Whilst, it is influenced largely by quality of the images (e.g., resolution, exposure, light condition), and the results usually contain much meaningless information (Ferrero et al., 2009). Image-based fracture identification, both manually and automatically, faces a nearly insurmountable problem of how to extract convincing information of fractures, which are 3-D in nature, from 2-D photos.
The Terrestrial Laser Scanning (TLS), an emerging LiDAR technology that provides high-resolution 3-D topography (millimetre accuracy) and color images of field outcrops, has been gradually changing the situation (Buckley et al., 2006, Buckley et al., 2008) in recent years. Terrestrial Laser Scanners are now small enough to be taken to the field even in tough geographic conditions, and able to capture outcrops (including topography and surface color) with high accuracy. Because of these advantages, a growing number of applications has been put forward in various fields, including engineering construction (Wang et al., 2014), agriculture and forestry managements (Liang et al., 2016, Ouédraogo et al., 2014), monitoring of topography and channels (Goodwin et al., 2016, Kuo et al., 2015) and particularly rock surface information extraction (Ahlgren and Holmlund, 2003, Hodgetts, 2013, Slob et al., 2002, Sturzenegger and Stead, 2009). Two kinds of approaches were proposed to detect fractures using 3-D TLS surface data, based on intersection lines between the fitting planes of rock mass surfaces (Slob et al., 2007, Gigli and Casagli, 2011) and the principle curvature of the vertices on the digital surface model of the rock mass (Umili et al., 2013) respectively. However, these two approaches are both indirect ways to identify fractures from TLS data, and the results rely generally on the data pre-processing. For instance, result of the former approach highly depends on the goodness of fitting planes and segmentation accuracy of the rock mass; whereas that of the latter one is largely influenced by the mesh quality and smooth degree.
In this paper, we proposed a new and simple approach to identify fractures directly and thus acquire surface density from 3-D surface model of natural outcrops generated from TLS data. We take one outcrop from the Shizigou anticline in the Qaidam Basin (NW China) as the case to validate the method and obtain optimal parameters. The results show that the proposed method can detect the fractures well and eliminate most redundant information if suitable parameters are chosen. Though requiring further improvement, the method provides for the first time an unbiased way to identify fractures, so that results from different data sources and interpreters can be well correlated. We believe that it will help a lot in analyzing the development of fractures and other related fields.
Section snippets
Data collection and pre-processing
Our study is based on TLS data collected from a Riegl VZ-1000 Terrestrial Laser Scanner with highest resolution of 3 mm. It's one of the mostly commonly-used scanner machine, and designed to be easily manipulated in the field. The scanner is equipped with a calibrated Nikon D300 digital camera mounted rigidly on top of the instrument body. Captured digital images are therefore registered to the scanner coordinate, allowing the RGB color to be later assigned onto the 3D geometry (McCaffrey et
Triangulation
Triangulation is to link the point cloud via a series of triangles with the aim of producing a solid surface (mesh, triangulated irregular network) that is essential to identify fractures (Buckley et al., 2008, García-Sellés et al., 2011). It is possible and appropriate to form a solid surface from a point cloud data, as a 2-D Delaunay triangulation (Delaunay, 1934) finds the best criteria for triangle creation automatically from projected points (McCullagh, 1998). In contrast, the 3-D Delaunay
Case studies and calibration of parameters
We now use the TLS data collected in the Shizigou anticline of the western Qaidam Basin (NW China) to validate the proposed method. We also measured the fractures in the field and on the photographs at the same outcrop to calibrate the result and obtain optimal values of the two parameters: the threshold angle (α), and the spatial resolution (R) of exported data.
Discussions
The method proposed in this paper can identify structural fractures based on TLS data to evaluate fracture development with many superiorities. Compared to the principle curvature (Umili et al., 2013) method, the detected lines contain more fragmented information which may be small scale fractures. Some structural fractures neglected easily by human operation can be identified within seconds accurately compared to field measurement and image detection. In addition, we can get different types
Conclusions
A wide range of fields are now making use of TLS data because of its high precision and accuracy. This paper proposed a simple and unbiased approach to identify fractures directly from 3-D surface model of natural outcrops generated from TLS data. One outcrop of Shizigou Anticline, Qaidam Basin is chosen to validate this method and obtain the optimal parameters, with surface density acquired from the field and image as references. In the case, the proposed method can accurately identify most of
Acknowledgements
The paper is supported by the National Key Scientific and Technological Projects (Grant nos. 2011ZX05009-001 and 2016ZX05003001-003), and the National Key Research and Development Plan Program (Grant no. 2016YFC060100X). The authors would like to thank the open source graphics library visualization tool kit (VTK) company for the use of their algorithm library, Wenyong Weng for helpful advice and Rui Wang for the implementation of some algorithms. We sincerely appreciate the four anonymous
References (47)
- et al.
Fracture patterns and petrophysical properties of carbonates undergoing regional folding: a case study from Kurdistan, N Iraq
Mar. Pet. Geol.
(2016) - et al.
Step edge detection method for 3D point clouds based on 2D range images
Opt. Int. J. Light Electron Opt.
(2015) - et al.
Dissolution and its impacts on reservoir formation in moderately to deeply buried strata of mixed siliciclastic–carbonate sediments, northwestern Qaidam Basin, northwest China
Mar. Pet. Geol.
(2013) - et al.
Supervised identification and reconstruction of near-planar geological surfaces from terrestrial laser scanning
Comput. Geosci.
(2011) - et al.
Semi-automatic extraction of rock mass structural data from high resolution LIDAR point clouds
Int. J. Rock. Mech. Min. Sci.
(2011) - et al.
Assessing the repeatability of terrestrial laser scanning for monitoring gully topography: a case study from Aratula, Queensland, Australia
Geomorphology
(2016) Laser scanning and digital outcrop geology in the petroleum industry: a review
Mar. Pet. Geol.
(2013)- et al.
Influence of range measurement noise on roughness characterization of rock surfaces using terrestrial laser scanning
Int. J. Rock. Mech. Min. Sci.
(2011) - et al.
Monitoring channel responses to flood events of low to moderate magnitudes in a bedrock-dominated river using morphological budgeting by terrestrial laser scanning
Geomorphology
(2015) - et al.
Terrestrial laser scanning in forest inventories
ISPRS J. Photogramm. Remote Sens.
(2016)
Multiscale edge detection based on Gaussian smoothing and edge tracking
Knowl. Based Syst.
The evaluation of unmanned aerial system-based photogrammetry and terrestrial laser scanning to generate DEMs of agricultural watersheds
Geomorphology
FEM analysis of deformation localization mechanisms in a 3-D fractured medium under rotating compressive stress orientations
Tectonophysics
Close-range terrestrial digital photogrammetry and terrestrial laser scanning for discontinuity characterization on rock cuts
Eng. Geol.
Fracture patterns in successive folding in the western Sichuan basin, China
J. Asian Earth Sci.
A subpixel edge detection method based on an arctangent edge model
Opt. Int. J. Light Electron Opt.
A new method for automatic discontinuity traces sampling on rock mass 3D model
Comput. Geosci.
Applications of terrestrial laser scanning for tunnels: a review
J. Traffic Transp. Eng.
Influence of structural position on fracture networks in the Torridon Group, Achnashellach fold and thrust belt, NW Scotland
J. Struct. Geol.
Cenozoic deformation history of the Qaidam Basin, NW China: results from cross-section restoration and implications for Qinghai-Tibet Plateau tectonics
Earth Planet. Sci. Lett.
Using 3-D outcrop laserscan for fracture analysis
Am. Assoc. Pet. Geol. Search Discov. Artic.
Terrestrial laser scanning in geology: data acquisition, processing and accuracy considerations
J. Geol. Soc.
Integration of terrestrial laser scanning, digital photogrammetry and geostatistical methods for high-resolution modelling of geological outcrops
Remote Sens. Spat. Inf. Sci.
Cited by (37)
Intelligent image-based identification and 3-D reconstruction of rock fractures: Implementation and application
2024, Tunnelling and Underground Space TechnologyScanline intersection similarity: A similarity metric for joint trace maps
2023, Computers and GeosciencesMachine learning-based classification of rock discontinuity trace: SMOTE oversampling integrated with GBT ensemble learning
2022, International Journal of Mining Science and TechnologyA state-of-the-art review of automated extraction of rock mass discontinuity characteristics using three-dimensional surface models
2021, Journal of Rock Mechanics and Geotechnical EngineeringSemi-automatic mapping of dyke and dyke-related fractures using UAV-based photogrammetric data: A case study from Sijiao Island, coastal Southeastern China
2020, Journal of Structural GeologyCitation Excerpt :IGCOC algorithm can provide reliable and intuitive cutouts for geologists. Structural fractures of natural outcrops typically appear as sharp edges, where the image intensity changes sharply (Cao et al., 2017; Zou et al., 2012). There are many existing algorithms to detect edges such as Canny, Sobel, Hessian, phase congruency and phase symmetry (Canny, 1986; Duda and Hart, 2003; Frangi et al., 1998; Miao et al., 2010; Morrone and Owens, 1987).
A deterministic-stochastic identification and modelling method of discrete fracture networks using laser scanning: Development and case study
2019, Engineering GeologyCitation Excerpt :Therefore, identification is the first problem we address. With efficient and precise identification of discontinuities, information extraction can be achieved immediately (Cao et al., 2017). In this study, a laser scanner with Real-time kinematics (RTK) was used for the field data acquisition (Bellian et al. 2005).