Automated estimation of forest parameters for Sweden using Landsat data and the kNN algorithm

Sammanfattning

Abstract – The project “kNN-Sweden” has mapped forest parameters, such as wood volume, age, and height over Sweden. Landsat ETM satellite data from 2000, digital map data, and forest inventory data were combined to produce continuous estimates of forest parameters. The method for estimating the forest parameters was a ”k-Nearest Neighbor” algorithm. Reference data were obtained from the Swedish National Forest Inventory. The project was completed through use of an automated production-line, written in-house. The production-line includes steps such as haze reduction and topographic correction of the satellite data, as well as updating of the inventory data. The end product results in several raster files including total wood volume; volume for Norway spruce, Scots pine, birch, lodgepole pine, beech, and oak; height; and, age. Spin-off products, such as dominant tree species, stand delineation through generalisation of the data, and base information for property taxation are made. Future directions are estimation using SPOT data and neural network implementation

Nyckelord

kNN; estimation; forest parameters; automation

Publicerad i

Utgivare: Proceedings, ISRSE

Konferens

31st International Symposium on Remote Sensing of Environment

SLU författare

• Reese, Heather

  • Institutionen för skoglig resurshushållning, Sveriges lantbruksuniversitet
    

• Granqvist Pahlen, Tina

  • Institutionen för skoglig resurshushållning, Sveriges lantbruksuniversitet
    

• Egberth, Mikael

  • Institutionen för skoglig resurshushållning, Sveriges lantbruksuniversitet
    

• Nilsson, Mats

  • Institutionen för skoglig resurshushållning, Sveriges lantbruksuniversitet
    

• Olsson, Håkan

  • Institutionen för skoglig resurshushållning, Sveriges lantbruksuniversitet
    

UKÄ forskningsämne

Miljö- och naturvårdsvetenskap
Skogsvetenskap
Landskapsarkitektur

Permanent länk till denna sida (URI)

https://res.slu.se/id/publ/6209