Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1049631 | Landscape and Urban Planning | 2010 | 16 Pages |
Significant parts of Fennoscandia have been used by Saami reindeer husbandry for many centuries, and this has altered the vegetation cover. Cultural remains like abandoned reindeer pens (giedtieh) and milking pens (buhtjeme-aevieh) are found in connection to abandoned seasonal dwellings (låavth-gåetieh and derhvie-gåetieh) for the reindeer Saami groups. The abandoned pens still appear as oval shaped areas of luxuriant grass- and herb-rich vegetation surrounded by meager mountain vegetation, due to slow growth of vegetation. This paper presents a study on using visual analysis of aerial and satellite imagery, multispectral classification (hyper-clustering) and linear spectral unmixing (SAM) of satellite imagery to detect Saami cultural remains. Visual analysis of high spatial resolution satellite imagery (QuickBird-2) and aerial photography enabled detection of maintained and abandoned reindeer pens, milking pens and old individual dwelling sites. Signs of them can also be detected by middle resolution imagery (Landsat TM/ETM+ and IRS-P6) but are to coarse to use since the spatial extent of these remains is to small to detect without prior knowledge of their location in the landscape. Using the classification methods on the fused Landsat ETM+ PAN imagery, we were able to detect the pens substantial to almost perfect (70–95% accuracy). The hyper-clustering method was slightly better than the SAM-method, but the first was more labor demanding. The results from this study show that use of satellite imagery is a helpful tool for the detection of even subtle traces in vegetation of cultural activities in large-scale mountain landscapes.