Overview and details of the sessions and sub-session of this conference. Please select a date or session to show only sub-sessions at that day or location. Please select a single sub-session for detailed view (with abstracts and downloads if available).
E2: ID.10667 Forest Resources Research
10:30am - 11:30am
Session Chair: Christiana Cornelia Schmullius
Session Chair: Yong Pang
Workshop: Forest Mapping & Retrievals
Location: Sun Moon Room -1, 5.5 Floor, Junyi Dynasty Hotel
Finnish experiences of forest mapping based on satellite SAR derived elevation data
The land area of Finland is predominantly covered by forest and therefore the economy and well-being is greatly dependant on this ecosystem. One of the most important aspects from the economy point of view is the accurate and timely information about the forest resources. This information is collected in forest inventories, which gives accurate data so that forest operations can be planned in more optimal way effecting significantly to the economy of the forest owners. In this project, we have focused on the use of elevation data derived from SAR satellite images and how this elevation data can be used in forest resources mapping. Using SAR satellite data forest inventories could be carried more frequently, which could partially lead to a more cost-efficient update process of the national forest inventory. The main data source has been the German TerraSAR-X and TanDEM-X data. We created elevation models (3D model describing the forest canopy height) using stereo-radargrammetric and SAR interferometric techniques and the accuracy of forest inventory variables were checked using field measured ground truth data. In is important to note that SAR based canopy height models require the use of Airborne Laser Scanning (ALS) data as the terrain model. Finally, we compared the SAR based results with airborne techniques to have a better understanding of the usability of different techniques. In the presentation, we will show the first and preliminary comparison results of the forest resources maps in our test area, Evo, Finland.
3D Data from Different Remote Sensing Sensors in Forest Information Extraction and Change Detection
Finnish Geospatial Research Institute, NLS, Finland;
Forest resource maps are needed for forestry purposes as well as for global carbon cycle modelling related to climate change. Detailed and up-to-date forest information is required for allocation of forestry activities and also for national and international reporting obligations. The combination of increasing spatial resolution and availability of 3D techniques has increased the ability to obtain forest information and monitor the changes using remote sensing data. Image-based point clouds, coming from optical images or synthetic aperture radar (SAR) images, could be used for estimations of forest inventory attributes using methods similar to airborne laser scanning (ALS) data, i.e., point cloud processing.
A comparison study was carried out to study the explanatory power and information contents of several 3D remote sensing data sources on the retrieval of above ground biomass, stem volume, basal area, basal-area weighted mean diameter and Lorey’s mean height at the plot level. The remote sensing data sources used were: SAR interferometry (Tandem-X), SAR radargrammetry (TerraSAR-X), satellite-imagery having stereo viewing capability (WorldView-2), airborne laser scanning (ALS) with various densities and aerial stereo imagery. Point clouds representing forest height were extracted from each data set. The performance of these data sources in forest inventory attribute prediction was evaluated with 91 sample plots. The plots are located in in Evo, southern Finland within a boreal forest zone and were surveyed in 2014 for this comparison. The sample plot sizes were 32 m x 32 m and 16 m x 16 m. Preliminary forest change detection results based on 3D data from Tandem-X interferometric SAR data are also presented.
The prediction models were built using random forests technique. Features derived from each data source were used as independent variables and field measurements of forest attributes as response variable. Overall results showed that ALS data produced the most accurate estimates for forest inventory attributes when compared with other 3D point clouds. However, the point density of ALS data is higher than the point density of image-based data. The major advantage of space-borne images, compared with airborne images or ALS, has been their availability and larger coverage. The results of this study confirm that airborne and space-borne images in combination with 3D techniques can be used for developing methods for operational forest inventory and for monitoring of forest attributes of large areas at regional and country level.