Digital solutions for detecting and monitoring forest damage

Tuesday, March 28

09.30–10.00: Registration & coffee

10.00–10.10: Opening of the conference by the Dean of the Faculty of Technology, Jesper Andersson

10.10–10.30: Introduction and presentation of the program
Johan Fransson and Johanna Witzell

10.30–12.00: Session 1 - Early detection of forest damage

• 10.30–11.10: Keynote, Eija Honkavaara, Finnish Geospatial Research Institute, Finland, “Detecting and monitoring abiotic and biotic damage of forest using spectral drone remote sensing”
• 11.10–11.35: Langning Huo, Swedish University of Agricultural Sciences, Sweden, “Detectability of European spruce bark beetle green-attacks using multispectral drone images”
• 11.35–12.00: Endijs Baders, Latvian State Forest Research Institute Silava, Latvia, “Monitoring the health of Norway spruce forests using UAV imagery: Insights into responses to simulated damage and the importance of long-term monitoring”

12.00–13.30: Lunch at Teleborg castle

13.30–15.00: Session 2 - Monitoring of forest damage

• 13.30–14.10: Keynote, Allan Buras, Technical University of Munich, Germany, “Monitoring forest stress and damage in near real-time across Europe”
• 14.10–14.35: Jörgen Wallerman, Swedish University of Agricultural Sciences, Sweden, “Efficient storm damage inventory with drones”
• 14.35–15.00: Puzhao Zhang, KTH Royal Institute of Technology, Sweden, “Multi-source satellite remote sensing and deep learning for mapping forest damage induced by wildfire”

15.00–15.30: Coffee break with poster presentations

15.30–17.00: Session 3 - Ground-truthing experiments and data

• 15.30–15.55: Albert Monteith, Chalmers University of Technology, Sweden, “Effects of forest stress and damage on radar measurements”
• 15.55–16.20: Carl Barck, ForestX, Sweden, “Bringing research from academic to practical work – ForestX the company that help forest companies to realize there ambitions of digitalization”
• 16.20–16.40: Anton Holmström, Katam Technologies, Sweden, ”Video data for spatiotemporal analysis of biotic damages in different forest environments across the globe”
• 16.40–17.00: Levi Farrand, Deep Forestry, ”Autonomous forestry field work with 3D Artificial Intelligence”

19.00–19.30: Aperitif and mingle

19.30: Three course dinner downtown

Wednesday, March 29

09.00–11.30: Session 4 - Technical advances in digital solutions for detection and monitoring of forest damage

• 09.00–09.40: Keynote, Yunsheng Wang, Finnish Geospatial Research Institute, Finland, “Learning forest processes through long term hyper-temporal 3D observations”
• 09.40–10.05: Markus Drugge, Arboair, Sweden, “Arboair – For the power of trees”
• 10.05–11.30: Johan Fransson, Linnaeus University, Sweden, “Detecting wind-thrown forest damage in SAR images through backscatter signatures”

11.30-13.00: Lunch at Teleborg castle

13.00–15.00: Session 5 - Technical advances in digital solutions for detection and monitoring of forest damage (cont.)

• 13.00–13.25: Sanna Härkönen, Bitcomp, Finland, “AI-based forest vitality risk monitoring service for forest organizations”
• 13.25-13.50: Xinlian Liang, Wuhan University, China, "Building multi-view point clouds from close-range sensing for fine scale tree studies"
• 13.50–14.30: Anders Persson and Örjan Laneborg, Swedish Forest Agency, Sweden, “Geodata for forest damage”
• 14.30-15.00: Henrik Holmgren and Filip Jägervall, Södra, Sweden, ”Södra’s field App”

15.00-15.30: Coffee break with a short walk to the Enterprise demo site

15.30–16.50: Enterprise demo by Katam Technologies including autonomous drone flights (outdoors)

16.50–17.00: Closing of the conference

Eija Honkavaara

Dr. Eija Honkavaara is a Research Professor at the Finnish Geospatial Research Institute (FGI) in the National Land Survey of Finland and Docent at the Aalto University. She is leading the @dronefinland research team at FGI. She received her doctoral degree in Photogrammetry from the Aalto University in 2008 with a thesis “Calibrating digital photogrammetric airborne imaging systems using a test field”. Currently her research focuses on drone-based remote sensing, photogrammetry, and hyperspectral imaging particularly in forestry and agricultural applications. Together with the @dronefinland team she implements and studies performance of new drone remote sensing technologies for analysis of different phenomenon and develops and implements processing and analytics methods for near real-time and real-time applications. Her present research projects study the quantitative estimation of vegetation parameters, the forest decline and damage, and anomaly detection, for example, to enable detection of different stages of bark beetle outbreaks and to provide real-time situational awareness for wildfire management. She is co-leading the working group “Autonomous Sensing Systems and their Applications” in the International Society for Photogrammetry and Remote Sensing (ISPRS) and she is Associate Editor of the Open ISPRS Journal of Photogrammetry and Remote Sensing.

Detecting and monitoring abiotic and biotic damage of forest using spectral drone remote sensing

Drone remote sensing using multispectral and hyperspectral imaging, photogrammetry and LiDAR technologies offers unprecedentedly accurate and flexible means for monitoring of forest condition. Despite great promises, there exits knowledge gaps on how these technologies could be utilized efficiently, precisely, and autonomously. At FGI, the @dronefinland drone remote sensing research facility provides basis for studying these questions. It is equipped with advanced remote sensing equipment including hyper- and multispectral cameras in visible to shortwave-infrared-ranges (VIS-SWIR) and LiDARs, as well as with a heterogeneous fleet of drones. Furthermore, we are maintaining test areas to enable research on beyond visual line of sight remote sensing. In applications side, we study machine learning techniques for analysis of images having different spectral and spatial properties, particularly focusing on novel deep learning techniques. Detection of different stages of Ips typographus L. outbreaks in Norway spruce trees in Finland using different spectral remote sensing techniques has been our long-term research focus. A recent research topic is the potential of VIS-SWIR hyperspectral cameras for detecting a root-rot of spruce (Heterobasidion parviporum) based on crown symptoms. We are also studying holistic mapping solutions using drone swarms for creating real-time situational awareness, e.g. during forest fires. This presentation will show our recent research results on forest health monitoring.

Allan Buras

Dr. Allan Buras is a physical geographer (M.Sc.) and forest ecologist (Ph.D.) focusing on forests’ climate-change resilience by means of tree-ring analyses, remote sensing, and modelling. In context of this keynote, his work around monitoring most recent European forest decline using satellite-borne remote sensing deserves a specific mention. In 2019, he released the European Forest Condition Monitor which provides near real-time information on the canopy condition of European forests as well as corresponding data available for download.

Links:

http://waldzustandsmonitor.de/en/forest-condition-monitor/

http://interaktiv.waldzustandsmonitor.de

Monitoring forest stress and damage in near real-time across Europe

Over the last decade, European forests have experienced an increasing frequency of weather extremes resulting in successive forest decline and increasing rates of forest die-back. In order to sustainably manage and adapt forests to climate change, foresters and stakeholders need concurrent information on forest condition to identify hot-spots of forest decline and react accordingly. On the other hand, to more precisely quantify species-specific climate resilience in context of the ongoing forest conversion, forest research requires large quantities of empirical data to resemble various environmental conditions across gradients in space and time. Both requirements can be efficiently solved using satellite-borne remote sensing techniques. This keynote aims at reviewing various existing forest monitoring products with a specific emphasis on the European Forest Condition monitor. Existing challenges of satellite-borne forest monitoring are briefly outlined, providing insights into upcoming research avenues in context of monitoring the climate resilience of European forest ecosystems.    

Yunsheng Wang

Dr. Yunsheng Wang is a Senior Research Scientist from the Department of Remote Sensing and Photogrammetry, Finnish Geospatial Research Institute (FGI), National Land Survey of Finland (NLS), PI and coPI of several projects funded by the Academy of Finland and EU Horizon 2020. She is the Chair of the ISPRS WG I/7 Data Quality and Benchmark of Sensors, and the co-chair of EU COST Action 3DForEcoTech. Her current research works focus on remote sensing and computational technologies for digitization, reconstruction, and understanding of nature environment (e.g. forest), and studying functioning processes and dynamics in temporal and spatial domains in the context of ecology and climate. Dr.Wang has published 50+ peer reviewed publications, including six ESI highly cited papers, with 2000+ citations and an h-index of 23 on Web of Science, and with 3500+ citations and an h-index of 28 on Google Scholar. She currently serves as associate editor of Forest Ecosystems, Journal of Applied Remote Sensing.

Learning Forest Processes Through Long Term Hyper-Temporal 3D Observations

The maintenance of forest health and sustainability depends on the complex interactions among various forest processes such as the photosynthesis; the cycling of carbon, water, and nutrients; the succession of plants; and so on. Each forest process involves a wide range of interrelated biological, physical, and chemical reactions, which are difficult to understand without long-term detailed observations, and that is the reason why our understandings about the forest processes are still limited despite the progress made in the past few decades. The forest structure and its dynamics along with changing environments is a key access to reveal the functions of forest processes.

Recent developments in remote sensing techniques, especially LiDAR, allow the monitoring the 3D structure of the environment with both high temporal and spatial resolution. The LiPhe TLS forest station is digitizing the observed forest at an unprecedented level of detail in both spatial and temporal spaces. The LiPhe TLS forest station consists of a RIEGL VZ-2000i laser scanning installed near to the top of a 35-meter-high flux tower of the Hyytiälä Forest Research Station (SMEAR II), which delivered a unique hyper-temporal (once per hour) high-spatial (1cm 3D point spacing at a 100 m range) resolution time-series 3D point cloud dataset (PCTS) for an area of approximately 263 m × 169 m (ca. 4000 individual trees). In total 11 352 point cloud data during 4th April 2020 and 30th July 2021 were collected. In addition, the SMEAR II research station constantly provides various climate and atmosphere observations since 1995.

The dense 4D (3D + time) PCTS facilitated several new studies of forest processes such as the phenology timings, and the tree growth strategies. It was observed that sensor calibrated LiDAR backscattering intensity values of the point cloud showed annual tree-level patterns, and these patterns have high within-species homogeneity that separate them from individuals of the other tree species. Such species-specific intensity signatures suggested the phenological timing, such as sprouting, flourishing, decaying, and falling/hibernating of leaves of health trees, which can be used to identify the unhealthy trees that might present different intensity signature along time. Moreover, the growth strategy can be studied at a tree level. With clear mapping of tree crown growth on both horizontal and vertical directions over times, the growth strategy of a tree is possible to be explained and predicted based on the species and size composition of trees in its immediate neighborhood. Such detailed mapping / prediction of tree growth facilitated new possibility of early recognition of unhealthy trees.