1.10.2018

Science Blog: Photogrammetry in fracture mapping with the aid of an unmanned aerial vehicle (UAV)

Teksti:
Nicklas Nordbäck, Senior Specialist
Jussi Mattila, Senior Specialist
Jon Engström, Senior Specialist

The essential principle in photogrammetry is triangulation, a science that dates back to 1480 and to the invention of perspective and projective geometry by Leonardo da Vinci. By capturing photographs from at least two different locations, so-called “lines of sight” can be established from cameras to points on a photographed object. By mathematically intersecting these lines, 3-dimensional coordinates of points can be created (Figure 1). When the locations of some points in the images or alternatively the camera locations are known, the resulting photogrammetric point cloud can also be accurately georeferenced and a 3D object generated from the points. From the photographs, RGB colour can also be assigned to each point, resulting in a photorealistic point cloud and 3D model.

Figure 1. Principle of photogrammetry.

Photogrammetry was first applied by the scientist Arago in 1840, using daguerreotype images. The first aerial photogrammetry was performed by Laussedat in the mid-19th century, using aerial kites and balloons. Since then, the most traditional way for geoscientists to acquire 3D information has been the use of aerial stereo photographs. Up until recent times, the cost and requirements of aviation have been limiting factors in the utilisation of aerial photos in photogrammetry. Starting from the times of Arago and Laussedat, photogrammetric processing techniques have developed in four separate phases, with each phase extending for about fifty years. Plane photogrammetry was used from about 1850 to 1900, analogue photogrammetry from 1900 to 1960, analytical photogrammetry from 1960 to 2010, and photogrammetry has recently experienced complete digitalization.

Since the beginning of the 21st century, the possibilities of the photogrammetric technique have greatly progressed thanks to the development of large-format digital cameras, affordable UAVs, capable computers and photogrammetric algorithms. Modern photogrammetric techniques enable a wide range of opportunities for geoscientific research. With the help of light and affordable UAVs equipped with high-resolution cameras and autopilot software, it has now become possible to survey large areas at close range and high resolution within minutes. This practice is especially useful for documenting dangerous, inaccessible or temporarily accessible sites, such as construction sites and rock quarries. Using modern photogrammetric software (such as Agisoft Photoscan or Pix4D), high-resolution and high-accuracy point clouds, 3D models, digital elevation models and orthomosaics can be simply produced using digital photographs as input data (Figures 2&3). In order to produce georeferenced outputs, ground control points (GCP) have to be used, but with modern surveying equipment (e.g. differential GPS), such measurements can easily be carried out. For more information on photogrammetric processing with Agisoft Photoscan the instructions by James (2017) are very useful.

Figure 2. The upper picture displays a photogrammetric 3D model of a road cut in Inkoo, while the lower picture displays the photogrammetric 3D point cloud for the same road cut with point normals coloured according to the dip direction, highlighting, for example, fracture planes with different directions.

 

Figure 3. Examples of photogrammetric 2D outputs from an outcrop at Kopparnäs, Inkoo. Upper left, a high-resolution 2D digital elevation model (DEM). Upper right, a high-resolution 2D orthomosaic with digitised fracture traces coloured according to the orientation of the fracture trend. Below, a rose plot of fracture trend orientations from Kopparnäs.

 

During the latest fieldwork campaigns, and especially during this spring, GTK’s Bedrock Construction and Site Assessment Unit (KAS) has developed its in-house capacity and competence in the utilisation of UAV photogrammetry for geological mapping purposes, mainly focusing on structural geology and fracture mapping (Figures 4-6). So far, KAS has utilised UAV photogrammetry for documenting trenches and landforms in relation to postglacial faults, vegetation in relation to peatland studies, outcrops, quarries, road cuts and underground tunnels in relation to bedrock construction and nuclear waste studies. The current UAV set-up and photogrammetric equipment, consisting of practically consumer-level equipment, has proved to be a rapid, powerful and accurate tool for detailed documentation of large areas in 3D. For fracture mapping, high-resolution orthomosaics and point clouds are produced and analysed with software such as ArcMap and CloudCompare. UAV photogrammetry can be used to determine fracture traces and orientations, which can be further employed in the assessment of fracture densities and their regional variability, fracture trace length distributions and fracture orientations. UAV photogrammetry is an efficient and practical method, particularly for surveying large outcrops, road cuts and areas where access is otherwise limited. However, it must be noted that the photogrammetric method for fracture mapping is no substitute for traditional structural geological mapping to be carried out in the field. Small details, such as fracture fillings and fracture kinematics, cannot usually be determined from photogrammetric material, and complementary field mapping is needed to obtain further information on these aspects. For fractures with trace lengths shorter than 0.5 m, it is also challenging to obtain an accurate estimate of the fracture orientation using photogrammetric data. However, UAV photogrammetry and traditional structural geological mapping can complement each other: UAV photogrammetry provides a tool for efficient surveying of large areas, providing a method for acquiring information on the regional variation of geological structures and accessing large data sets in a very minimal time. Traditional structural geological mapping is then used to complement the data set by focusing on the details of the geological structures and acquiring data that are not available from the photogrammetric models. By first carrying out a UAV photogrammetric survey and analysing the data from a larger area, it is possible in the office to accurately define the smaller areas where complimentary structural geological mapping should be carried out. By combining these two approaches, structural geologists now have an efficient methodology to produce high-quality and accurate data from relatively large areas. However, well-exposed outcrops, which are not always available, are a necessity for using the methodology. For more information on 2D and 3D fracture mapping the reader is referred to Thiele et al. (2017).

References:

James, Mike. 2017. SfM-MVS PhotoScan image processing exercise. https://www.researchgate.net/publication/320407992_SfM-MVS_PhotoScan_image_processing_exercise

Thiele et al. 2017. Rapid, semi-automatic fracture and contact mapping for point clouds, images and geophysical data. https://www.solid-earth.net/8/1241/2017/

Figure 4. Drone surveying and structural geological mapping of outcrops at Kopparnäs, spring 2018 (Photo: Nicklas Nordbäck).
Figure 5. Photogrammetric point cloud of a road cut in Inkoo. The locations, trace lengths and orientations of 133 fractures were mapped using the Compass plugin of CloudCompare software. Orientations are displayed in a pole plot with contour overlay (equal area, lower hemisphere projection).

 

Figure 6. Drone surveying at Kopparnäs, spring 2018 (Photos: Jon Engström).
Nicklas Nordbäck

Teksti: Nicklas Nordbäck

Nicklas Nordbäck (M.Sc.) is a Senior Specialist in the Bedrock Construction and Site Assessment Unit at GTK. He is a structural geologist with over 14 years of experience in site investigation and characterization activities in relation to the construction of an underground final repository for high-level nuclear waste in Finland and is specialized in brittle structures and UAV photogrammetry.

Jussi Mattila

Teksti: Jussi Mattila

Dr Jussi Mattila is a Senior Specialist in the Bedrock Construction and Site Assessment Unit at GTK. During his career, Dr Mattila has been involved in geological site investigations, 3D modelling and monitoring activities in relation to the siting of a nuclear waste repository and is specialized in structural geological analysis and modelling of bedrock structures.

Jon Engström

Teksti: Jon Engström

Jon Engström (M.Sc.) is a Senior Specialist in the Bedrock Construction and Site Assessment Unit at GTK. He is a structural geologist with over 14 years of experience in site investigation and characterization activities in relation to the construction of an underground final repository for high-level nuclear waste in Finland and is specialized in ductile structures.