Press release 25.7.2006
A combination of satellite images and aerial photograps brings better results in forest mapping than either of these image generation methods alone. According to a Metla study, accuracy was improved by as much as 17 percent compared to using mere satellite image interpretation.
Using multisource forest inventory techinique, developed for the National Forest Inventory (NFI) of Finland, estimates of forest parameters (such as stand volume, mean height, mean age and basal area) can be presented in the form of thematic maps or forest statistics. The use of satellite images as well as high-altitude and false-colour aerial photographs was examined to assess stand parameters by applying the multisource forest inventory technology developed by Metla.
Both the high-altitude panchromatic and colour-infrared aerial photographs added new information to the satellite imagery. The ground resolution of both aerial photograph types is high in comparison with satellite images. High resolution makes it possible to utilize textural features of the photograph. It should, however, be remembered that processing a single 180x180 satellite image is more convenient than handling an aerial photograph, since the area covered by an aerial photograph is smaller. Aerial photographs must be combined into larger mosaics to cover the necessary number of field plots.
The purpose of the study was to produce maps with different stand parameters and to create forest statistics of various-sized areas. To assess canopy coverage, statistical models were developed on the basis of measured canopy coverage materials. Other variables being studied were derived from the sample plot and tree measurements made during the NFI. The project focused on assessing forest parameters that could be derived from normal NFI field data or minor additional measurements made in the field. Studied parameters included canopy coverage, number of stems, standard deviation of diameter distribution, and mean diameter.The resolution used for satellite imagery was 25 m, and for high-altitude and false-colour aerial photography 1 m and 0.5 m, respectively. The study was started in 2003, and it was completed this spring. The “Estimating canopy coverage and spatial pattern parameters from remote sensing materials” project was funded by the Marjatta and Eino Kolli Foundation.