Dr. Dirk Hoffmeister


Dissertation Title

Feasibility studies of terrestrial laser scanning in Coastal Geomorphology, Agronomy, and Geoarchaeology

Supervisor: Prof. Dr. Georg Bareth
Advisor: Prof. Dr. Helmut Brückner


Terrestrial laser scanning (TLS) is a newer, active method of remote sensing for the automatic detection of 3D coordinate points. This method has been developed particularly during the last 20 years, in addition to airborne and mobile laser scanning methods. All these methods use laser light and additional angle measurements for the detection of distances and directions. Thus, several thousands to hundreds of thousands of polar coordinates per second can be measured directly by an automatic deflection of laser beams. For TLS measurements, the coordinates and orientation of the origin of the laser beam can be determined to register different scan positions in a common coordinate system.

Read more

These measurements are usually conducted by Global Navigation Satellite Systems or total station surveying, but also identical points can be used and data driven methods are possible. Typically, accuracies and point densities of a few centimetres to a few millimetres are achieved depending on the method. The derived 3D point clouds contain millions of points, which can be evaluated in post-processing stages by symbolic or data-driven methods. Besides the creation of digital surface and terrain models, laser scanning is used in many areas for the determination of 3D objects, distances, dimensions, and volumes. In addition, changes can be determined by multi-temporal surveys.

The terrestrial laser scanner Riegl LMS Z-420i was used in this work in combination with the Differential Global Positioning System system Topcon Hiper Pro, based on Real Time Kinematic (RTK-DGPS). In addition to the direct position determination of the laser scanner, the position of a self-developed reflector on a ranging pole was measured by the RTK-DGPS system to accurately derive the orientation of each measured point cloud. Moreover, the scanner is equipped with an additional, mounted camera Nikon D200 to capture oriented pictures. These pictures allow colouring the point cloud in true colours and thus allow a better orientation. Furthermore, the pictures can be used for the extraction of detailed 3D information and for texturing the 3D objects. In one of the post-processing steps, the direct georeferencing by RTK-DGPS data was refined using the Multi Station Adjustment, which employs the Iterative Closest Point algorithm. According to the specific objectives, the point clouds were then filtered, clipped, and processed to establish 3D objects for further usage.

In this dissertation, the feasibility of the method has been analysed by investigating the applicability of the system, the accuracy, and the post-processing methods by means of case studies from the research areas of coastal geomorphology, agronomy, and geoarchaeology. In general, the measurement system has been proven to be robust and suitable for field surveys in all cases. The surveys themselves, including the selected georeferencing approach, were conducted quickly and reliably. With the refinement of the Multi Station Adjustment a relative accuracy of about 1 cm has been achieved. The absolute accuracy is about 1.5 m, limited by the RTK-DGPS system, which can be enhanced through advanced techniques. Specific post-processing steps have been conducted to solve the specific goals of each research area.

The method was applied for coastal geomorphological research in western Greece. This part of the study deals with 3D reconstructed volumes and corresponding masses of boulders, which have been dislocated by high energy events. The boulder masses and other parameters, such as the height and distance to the current sea level, have been used in wave transport equations for the calculation of minimum wave heights and velocities of storm and tsunami scenarios and were compared to each other. A significant increase in accuracy of 30% on average compared with the conventional method of simply measuring the axes was detected.

For comparison, annual measurements at seven locations in western Greece were performed over three years (2009-2011) and changes in the sediment budget were successfully detected. The base points of the RTK-DGPS system were marked and used every year. Difficulties arose in areas with high surface roughness and slight changes in the annual position of the laser scanner led to an uneven point density and generated non-existing changes. For this reason, all results were additionally checked by pictures of the mounted camera and a direct point cloud comparison.

Similarly, agricultural plants were surveyed by a multi-temporal approach on a field over two years using the stated method. Plant heights and their variability within a field were successfully determined using Crop Surface Models, which represent the top canopy. The spatial variability of plant development was compared with topographic parameters as well as soil properties and significant correlations were found. Furthermore, the method was carried out with four different types of sugar-beet at a higher resolution, which was achieved by increasing the height of the measurement position. The differences between the crop varieties and their growth behaviour under drought stress were represented by the derived plant heights and a relation to biomass and the Leaf Area Index was successfully established.

With regard to geoarchaeological investigations in Jordan, Spain, and Egypt, the method was used in order to document respective sites and specific issues, such as proportions and volumes derived from the generated 3D models were solved. However, a full coverage of complexly structured sites, like caves or early settlements is partially prevented by the oversized scanner, slow measurement rates, and the necessary minimum measurement distance. The 3D data can be combined with other data for further research by the common georeference.

The selected method has been found suitable to create accurate 3D point clouds and corresponding 3D models that can be used in accordance with the respective research problem. The feasibility of the TLS method for various issues of the case studies was proven, but limitations of the used system have also been detected and are described in the respective chapters. Further methods or other, newer TLS systems may be better suited for specific cases.



Back to Alumni overview