Solutions

Following solutions have been realized and are available for industrial partners. We are happy to prove the applicabilities and modifications of the current methods for new industrial applications. Do not hesitate to contact us.


Large scale calculation of single trees

from ALS LiDAR data by a novel 3D segmentation technique

Software description



The new technology enables the precise and complete analysis of entire forest areas in all forest layers. Starting from dense point clouds, which are measured by means of laser beams or photogrammetrically from an aircraft or UAV, individual trees are completely captured in three dimensions. This not only allows an accurate determination of important tree parameters (position, height, crown base, volume), but also enables areal assessment of regeneration in the lower forest layers. Unlike the conventional forest inventory applied just in a few control points, the method provides a complete, comprehensive forest information, making it more accurate and more economical. The results can be imported into GIS systems for forest statistical calculations and visualizations. Even dead trees can be detected from the segmented tree point clouds.


Other features: - patented 3D technology - Platform: C / C ++, Windows or Linux:


Features:


  • Dominating trees as well as small trees in lower forest areas are extracted
  • Tree species classification
  • Fully automatic and seamless processing of large scale areas
  • Utilizes outcomes of full waveform decomposition (i.e. intensity, pulse width)
  • Optionally: Watershed-based segmentation of canopy surface only
  • Results are stored in a shape file for each tree
  • Platform: C/C++, Windows or Linux, stand-alone application or as a library for integration
  • Patented 3D technology


Input Data:


  • DTM (either raster or random, – ASCII)
  • LiDAR data (XYZ, optionally intensity and pulse width, ASCII or LAS 1.3)


Output data:


  • Tree position
  • Tree height
  • Crown base height
  • Crown volume
  • Tree species
  • Optionally: LiDAR points of segmented trees


Service and Product:


The software can be


  • installed as a stand-alone tool for service calculations, or
  • provided as a library for integration in customer tools
  • Service (offline calculation) can also be provided
  • We have also experience in extracting other forest objects like dead trees (fallen, standing). If interested, please ask.


Dieser Inhalt benötigt Flash / This content needs flash Adobe Flash Player.

Optimal rule-based classification of solder joints

Hidden solder joints can be detected by X-Rays. A classification was developed which

  • finds optimally decision rules by data analysis
  • works without parameters and , hence robustly
  • takes into account customer requirements
  • leads to reasonable decisions



Please find more informations here. (in German only)


Prof. Dr. Peter Krzystek
Room: F 303

Tel.: 089 1265-2617
Fax: 089 1265-2698

Profile >