Detection of reindeer pens in the Swedish mountain areas using remote sensing
A comparative study on color infrared aerial photography and very high-resolution satellite imagery
According to the Swedish Environmental Quality Objectives a Magnificent Mountain Landscape, protection should be provided for the majority of mountain areas with representative and significant natural and cultural assets by 2010. Whether this objective will be achieved is highly conditioned by the availability of knowledge. Today only a portion of the representative and significant cultural assets are known and only a subset have been systematically invented and documented.
Some of the most significant results of the human interaction with the mountain landscape are the remains of historic reindeer pens (sve: renvallar). A reindeer pen is a restricted area where reindeer herders, until the middle of the 1900s, used to gather domesticated reindeers for milking, slaughtering, marking of calves, training and protection. Because of the slow growth of vegetation in the mountain area, the deserted reindeer pens still appears as more or less distinct oval shaped areas of luxuriant grass and herb-rich vegetation surrounded by meager heath vegetation. However the difficulties in detecting these remains are increasing each year due to the gradual deterioration of their characteristic features.
The idea of the project is to develop and evaluate methods for detection of the cultural heritage in the mountain area, on the basis of remote sensing. Due to the rapid development in the field of very high-resolution satellite imagery in recent year’s new possibilities have emerged making it feasible to apply automated detection procedures of small objects in satellite data. Still, the costs for purchasing very high-resolution data are very high regarding the need for extensive geographical coverage. In addition to this, it is not fully proven to what degree automated classification algorithms may substitute the hitherto superior method of manual interpretation of aerial imagery concerning the detection of fine objects. In order to establish a suitable method for further implementation we want to evaluate the different approaches with regard to cost, accuracy and work effort.
The aim of the project is to develop and evaluate cost-effective methods for the detection and monitoring of the cultural heritage in the mountain area, especially the remains of reindeer pens. The immediate needs concern methods for detection and mapping of such areas but eventually there will also be a need for implementation of such methods in monitoring activities. From a user point-of-view it is crucial to develop a method which meets the following requirement:
• It should be applicable on a vast geographical area
• It should be repeatable
• preferably, it should be based on automated classification or detection procedures
Jerker Moström, Riksantikvarieämbetet, 08-5191 8534, jerker.mostrom(a)raa.se
Rikard Sohlenius, Riksantikvarieämbetet, 08-5191 8598, rikard.sohlenius(a)raa.se
Maj-Liz Nordberg, Institutionen för naturgeografi och kvartärgeologi, Stockholms universitet, 08-16 4962, majliz.nordberg(a)natgeo.su.se
25 maj 2009