Forest Modeling

A high-quality forest modelling is the basis for almost every application about the forest:

  • Where do I receive which timber assortment in which quantity at what price? (Who provides what?)
  • Automatic geo-referenced timber harvest
  • Asset valuation (tax office)
  • Simulation of forest growth (What is my forest in 30 years?)
  • Simulation of forest fire
  • Simulation and estimation of consequences of change in climate

The Virtual Forest in its final stage of extension will enable the description of the forest on the level of block of trees, single trees and sample data. Long-term declared aim of the Virtual Forest appropriately is the generation of a data basis, which lists all trees in North Rhine-Westphalia having a breast-height diameter bigger than 20 centimetres (approx. 240 million single trees) and knows their exact location with geographical coordinates. The fundamental feasibility of this plan has already been shown in the first phase of the project. However, in the second phase the generation of a new stand enumeration as a first milestone was the centre of attention. It divides the forest (part) automatically in blocks of trees and describes those again by summary attributes like tree species, share of area, stocking degree and store.

From the vantage point of the sensor data processing algorithm, both estimates are close to each other. For both, a powerful classification of tree species is necessary. The algorithm for recognition of stand as well as the algorithm for single tree recognition profit mutually from their results. Only the diversion of stock and tree described attributes is fundamentally different. The connections are worked out in cooperation with the Institute of Forest Inventory and Remote Sensing of Georg-August-University Göttingen (Prof. Kleinn). In summary, this proceeding leads to a mutual calibration of the processes for a spacious approach and, with it, to a quality enhancement on both sides.