• D
  • A
  • B
  • C
  • Sötåsa village


We develop tools for semi-automatic and automatic information extraction from digital image data. Our algoritms are built on methods making systematic use of local spatial contextual information.Our state-of-the-art algorithms segment and delineate colored and textured areas in digital images. The core of our technology is unique and several patents have been granted.

The software is a generic framework that is useful in a number of applications such as medical imaging, material sciences and remote sensing. So far our methods have been implemented into a stand-alone product,  to make it possible to easily analyze and process large geo coded satellite images. We also have expertise in ArcGIS, Matlab, and can integrate our technology with your current software. We also offer consulting services.

Here are some examples of applications of our technology:

Finding trees in aerial photographs:

Segmenting land use into milk production and sheep farming

In this image, qualities coming from milk production is marked in green and sheep farming in yellow. The raw data comes from the 20250 x 20250 pixel Landsat TM scene 194/20, 7 July 2001. The supervised segmentation into land use qualities was trained on the two delineated areas within the black rectangle.