Forest cover type classification at Mt. Asama using ALOS fully Polarimetric PALSAR data, in Nagano prefecture, central Japan
スポンサーリンク
概要
- 論文の詳細を見る
PALSAR (Phased Array L-band Synthetic Aperture Radar) was installed on the Japanese ALOS (Advanced Land Observing Satellite) that was launched on 24 January 2006. This is the first satellite regularly operated with fully polarimetric SAR in the world, and is considered useful for monitoring forest and biomass, as well as topography and land use. Forest cover types and stand volume classifications were produced for Mt. Asama using unsupervised and supervised classifications from the backscatter matrix using polarimetric PALSAR data. The Mt. Asama region was selected for study because of its gentle geographic features and large number of representative Japanese larch (Larix kaempferi), red pine (Pinus densifolia), and sub-alpine conifer (Abies mariesii and Tsuga diversifolia) forests. The unsupervised classification data allowed the forest region to be distinguished from fields, rice fields, pastures, residential areas, roads, and bare ground. We also produced forest cover type and volume classification images using a supervised classifier with GIS and field data, with a classification accuracy of 11.0-56.5%. The accuracy of highest volume classes was 301-450 m3/ha (56.5%) and 451-800 m3/ha (49.6%). This may have been because the influence of geographic features such as slope azimuth and ridges were larger than that of forest cover types. Therefore, polarimetric PALSAR data are suitable for monitoring deforestation and detecting changes in forest cover types in a homogenous, large forested region without complex mountainous geographic features.
- 社団法人 日本写真測量学会の論文
社団法人 日本写真測量学会 | 論文
- Landform monitoring of Tottori Sand Dune using aerial photographs.
- One Projector-One Camera System on Short Range Photogrammetry
- タイトル無し
- A Study on Environmental Evaluation Method Using the Satellite Image-Radiation quantity adjustment of multi-temporal data-.:-Radiation quantity adjustment of multi-temporal data-
- タイトル無し