You are here: Home Projects A1) Remote Sensing

A1) Remote Sensing

Remote sensing based methods for the assessment of forest structures

Barbara Koch & Holger Weinacker
Doctoral researchers: Xiang Liu (associated; since 2018) & Martin Denter  (from 2019)

University of Freiburg, Faculty of Environment & Natural Resources, Institute of Earth & Environmental Sciences,
Chair of Remote Sensing and Landscape Information Systems


Forest structures at the landscape, stand and tree levels are linked to various functions of our forests. These functions range from the provision of resources (e.g. drinking water, wood, etc.) to climatic stabilization effects, recreational use, and critical support for biodiversity.

Forest management influences the structural diversity and richness of our forests, and thus, their functionality beyond just timber production. Hence, describing structure is not trivial and there are numerous concepts for its quantification at different scales (e.g. tree, stand, landscape). Remote sensing has a long tradition of efficient assessment of forest structures, especially on the landscape and stand level.

Nowadays, laser-based distance measuring systems (i.e. LiDAR, Light Detection and Ranging) and photogrammetric methods (i.e. SfM, Structure from Motion) form the basis for the three-dimensional capture of forest geometries. Improved sensor technologies and the ability to move the sensors closer to the object of interest (close range) now enables us to record individual trees and stands at a high level of detail.

In addition, also new instruments allow more efficient terrestrial remote sensing based assessment of structures. Efficient and robust evaluation methods of close-range data from UAV and terrestrial platforms are a topic of current research.


Research questions and hypotheses

Forest biodiversity is influenced at all spatial scales. Landscape diversity and fragmentation must be recorded using methods that can record the entire landscape in a correspondingly detailed and efficient manner, even over time. Airborne aerial photographs and satellite images are particularly suitable for this purpose.

For the analysis of stands and individual trees, lightweight and mobile sensors and platforms are advantageous, as they can capture the object of investigation from close up and from a variety of angles. Terrestrial laser scanning systems (e.g. MLS, TLS) and drone-based images (e.g. UAV-SfM) are particularly useful for this purpose.

In project A1, we are dealing with the questions of how such recordings can be performed in the most efficient way and how to generate meaningful information from the acquired data. Merging information from these different levels and sensor systems is a further challenge. This leads to our research questions:

  • Which sensor systems, in combination with which platforms, are particularly suitable for measuring forest structures on different scales? Specifically, the use of small UAVs with lightweight cameras, which allows for a significant reduction of costs, for the acquisition of single tree structures and the use of mobile terrestrial laser scanning systems for forest structure monitoring shall be evaluated.
  • Which algorithms and new processing approaches are required to obtain high-quality information for the modelling of structural elements from the different remote sensing constellations?
  • Which conclusions can be drawn for biodiversity management based on the remote sensing-based structural descriptions?


Approach, methods and linkages

On the landscape level, A1 assesses the crown structures and standing deadwood for the entire ConFoBi study plots based on aerial photographs. A1 generated orthomosaics and digital elevation models twice of all 135 plots based on UAV RGB-images. Canopy changes will be mapped for all plots based on UAV data. Structural richness and complexity will be recorded and validated at the stand level with the help of corresponding indices.

The detection of individual structures (e.g. tree-related microhabitats) will be explored at the tree level. The detection of tree-related microhabitats is challenging from a remote sensing perspective. The structures to be detected are sometimes very fine and small, and the environment is extremely unstructured and difficult to access. Furthermore, various microhabitats in the tree crowns cannot be optimally detected from the ground because they may be covered or the viewing angle may be unfavorable.

We are therefore currently striving to classify these structures automatically utilizing small drones, which are flown directly in the stand using 3D reconstruction from camera images and machine learning. In regards, to terrestrial stans structure assessment we will test a hand-held laser measurement system, which allows to cover entire plots more easily due to the mobility of the system. All landscape and stand information is linked to the ecologically working B-projects to ensure the best possible usability of the information for further research or management instructions. The description of single tree structures is carried out in close cooperation with the A2 project.



In the first phase, we optimized the flight planning for the acquisition of forest structures by UAV and were able to describe the effects of different flight parameters on the geometric reconstruction. Various indices for the structural description of stands were subsequently evaluated in cooperation with the A2 project. We were able to show that there is a correlation, albeit weak, between biodiversity relevant tree microhabitats and such structural indices. In cooperation with the C2 project, we could show that structural indices can be an objective measure for structural description, which is reflected in expert assessments.

However, it was also shown that expert assessments in the same stand can be very heterogeneous. In the second phase we investigated UAV flight parameters for 3D reconstructions of the plots with the Structure for Motion method. We compared the image information from UAV to the information measured by a static TLS system. Furthermore, we tested small UAVs with RGB cameras for the assessment of small scaled structures along the tree stem.

First results describe the flight conditions for the use of such systems and its restrictions. With WorldView-3 satellite imagery standing deadwood has been mapped for various study plots. Additionally, numerous data sets for the ConFoBi overall project were created and processed in order to enable the entire group to work efficiently.


Future projects

For RTG phase II, A1 will consider an advanced monitoring concept for forest structures, which includes the wall-to-wall mapping of entire stands based on advanced remote sensing techniques. A regeneration and regrowth assessment will be tested with the dynamic hand-held laser system to investigate the influence of habitat trees on ecosystem processes together with the project A2. The Our findings will be tailored towards ecological monitoring and management perspectives. The significance of the structural assessments based on remote sensing devices compared to conventional methods of structure assessment on ecological parameters will be investigated.

Next PhD project (starting 1 July 2022).

In the third phase (PhD3), the previously collected UAV data will be combined and analyzed with newly assessed data. The tasks of PhD3 will be:

  • Assess and classify the stand structure from the ground with a terrestrial handheld laser scanner. The functionality and data quality of our new handheld laser scanner needs to assessed and described for the use on our study sites. Therefore, different algorithms have to be developed, tested and verified to classify the stand structure.
  • The structure classification based on terrestrial methods shall be linked to above ground UAV based measurements.
  • Relationship between abundancy and diversity of the investigated species and the structural classification from hand held TLS data as well as airborne data shall be investigated. In addition, multivariate analyses shall be implemented to show the relationship between different remote sensing-based structure classes on abundancy and diversity assessments on the plot (joint paper).

The candidate should have experience in the analyses of remote sensing data, basic programming knowledge (e.g. Python, R), 3D point cloud processing and basic statistical knowledge. In the master thesis the candidate should have worked with either multispectral data or 3D laser data. Close interdisciplinary collaboration, especially with A2 and B1 is expected.


ConFoBi publications by A1

Asbeck, Thomas; Pyttel, Patrick; Frey, Julian & Bauhus, Jürgen (2019). Predicting abundance and diversity of tree-related microhabitats in Central European montane forests from common forest attributes. Forest Ecology and Management, 432, 400–408.

Frey, Julian; Asbeck, Thomas & Bauhus, Jürgen (2020). Predicting Tree-Related Microhabitats by Multisensor Close-Range Remote Sensing Structural Parameters for the Selection of Retention Elements. Remote Sensing, 12, 867.

Frey, Julian; Joa, Bettina; Schraml, Ulrich & Koch, Barbara (2019). Same Viewpoint Different Perspectives—A Comparison of Expert Ratings with a TLS Derived Forest Stand Structural Complexity Index. Remote Sensing, 11, 1137.

Frey, Julian; Kovach, Kyle; Stemmler, Simon & Koch, Barbara (2018). UAV Photogrammetry of Forests as a Vulnerable Process. A Sensitivity Analysis for a Structure from Motion RGB-Image Pipeline. Remote Sensing, 10, 912.

Gustafsson, Lena; Bauhus, Jürgen; Asbeck, Thomas; Augustynczik, Andrey Lessa Derci; Basile, Marco & Frey, Julian et al. (2020). Retention as an integrated biodiversity conservation approach for continuous-cover forestry in Europe. Ambio, 49, 85–97.

Knuff, Anna K.; Staab, Michael; Frey, Julian; Helbach, Jan & Klein, Alexandra‐Maria (2019). Plant composition, not richness, drives occurrence of specialist herbivores. Ecol Entomol, 44, 833–843.

Knuff, Anna Katharina; Staab, Michael; Frey, Julian; Dormann, Carsten F.; Asbeck, Thomas & Klein, Alexandra-Maria (2020). Insect abundance in managed forests benefits from multi-layered vegetation. Basic and Applied Ecology, 48, 124–135.

Schiefer, Felix; Kattenborn, Teja; Frick, Annett; Frey, Julian; Schall, Peter & Koch, Barbara et al. (2020). Mapping forest tree species in high resolution UAV-based RGB-imagery by means of convolutional neural networks. ISPRS Journal of Photogrammetry and Remote Sensing, 170, 205–215.

Storch, Ilse; Penner, Johannes; Asbeck, Thomas; Basile, Marco; Bauhus, Jürgen & Braunisch, Veronika et al. (2020). Evaluating the effectiveness of retention forestry to enhance biodiversity in production forests of Central Europe using an interdisciplinary, multi-scale approach. Ecology and evolution, 10, 1489–1509.