| 講演抄録/キーワード |
| 講演名 |
2014-10-23 15:00
Characterization of agricultural land abandonment based on Landsat and ALOS PALSAR multitemporal measurements ○Farrah Melissa Muharam(Univ. Putra Malaysia)・Wataru Takeuchi・Soni Darmawan(Univ. of Tokyo)・Muhamad Hafiz Abd Razak(FELCRA) SANE2014-87 |
| 抄録 |
(和) |
Abandoned land could be considered as land that holds potential for agricultural purposes but is not cultivated for three consecutive years or more. These lands are commonly scattered; causes a difficulty in monitoring them. The monitoring of land abandonment is vital due to food security issues, consequences on environment such as soil and water quality, carbon sequestration, nutrient cycling, and wasteful resources. Active and passive remote sensing sensors could be very powerful to identify and to monitor land use/land cover and agricultural land. This study is thus conducted to investigate the characteristics of agricultural land abandonment in the state of Perak, Malaysia based on ALOS PALSAR and LANDSAT measurements. The methodology of this study consists of collecting ALOS-PALSAR and LANDSAT multitemporal data, preprocessing, ground surveying, and characterizing the changes in spectral signature over times. 22 ALOS-PALSAR and 9 LANDSAT multitemporal images have been collected. PALSAR preprocessing includes converting Digital Number (DN) to backscatter HH and HV on decibel (dB) units, while LANDSAT preprocessing involves atmospheric correction, image normalization, and vegetation index generation. From the ground survey, 40 training samples have been collected in different classes of abandoned agricultural land: oil palm, rubber, paddy field, and forest, and non-abandoned classes of oil palm, rubber, paddy field, and forest. The result of this research shows that ALOS-PALSAR and LANDSAT could be utilized to characterize agricultural land abandonment due to different phonologies of agricultural crops. |
| (英) |
Abandoned land could be considered as land that holds potential for agricultural purposes but is not cultivated for three consecutive years or more. These lands are commonly scattered; causes a difficulty in monitoring them. The monitoring of land abandonment is vital due to food security issues, consequences on environment such as soil and water quality, carbon sequestration, nutrient cycling, and wasteful resources. Active and passive remote sensing sensors could be very powerful to identify and to monitor land use/land cover and agricultural land. This study is thus conducted to investigate the characteristics of agricultural land abandonment in the state of Perak, Malaysia based on ALOS PALSAR and LANDSAT measurements. The methodology of this study consists of collecting ALOS-PALSAR and LANDSAT multitemporal data, preprocessing, ground surveying, and characterizing the changes in spectral signature over times. 22 ALOS-PALSAR and 9 LANDSAT multitemporal images have been collected. PALSAR preprocessing includes converting Digital Number (DN) to backscatter HH and HV on decibel (dB) units, while LANDSAT preprocessing involves atmospheric correction, image normalization, and vegetation index generation. From the ground survey, 40 training samples have been collected in different classes of abandoned agricultural land: oil palm, rubber, paddy field, and forest, and non-abandoned classes of oil palm, rubber, paddy field, and forest. The result of this research shows that ALOS-PALSAR and LANDSAT could be utilized to characterize agricultural land abandonment due to different phonologies of agricultural crops. |
| キーワード |
(和) |
polarimetric / HH / HV / spectral signature / phenology / separation / / |
| (英) |
polarimetric / HH / HV / spectral signature / phenology / separation / / |
| 文献情報 |
信学技報, vol. 114, no. 264, SANE2014-87, pp. 129-133, 2014年10月. |
| 資料番号 |
SANE2014-87 |
| 発行日 |
2014-10-15 (SANE) |
| ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
| PDFダウンロード |
SANE2014-87 |
|