@article{oai:jaxa.repo.nii.ac.jp:00022383,
 author = {Thapa, Rajesh Bahadur and Itoh, Takuya and 島田, 政信 and 渡邉, 学 and 本岡, 毅 and Shiraishi, Tomohiro and Thapa, Rajesh Bahadur and Itoh, Takuya and Shimada, Masanobu and Watanabe, Manabu and Motohka, Takeshi and Shiraishi, Tomohiro},
 journal = {Remote Sensing of Environment},
 month = {May},
 note = {The transformation of natural forest to non-forest cover is a dominant phenomenon in tropical regions; this creates unprecedented pressure to climate, biodiversity, and ecosystem services. Owing to persistent clouds and other atmospheric effects in the region, forest managers are facing difficulties in mapping and monitoring the forest cover changes consistently. This article presents an automated mapping method and examines the potentials of ALOS PALSAR data for characterizing natural forest cover in the tropics. Several scenes of high-resolution PALSAR data with HH and HV polarizations were processed to cover Sumatra Island. The mapping method applied image segmentation and threshold techniques to discriminate land covers. Non-forest land covers were separated using HH backscatter thresholds, while forest cover was based on HV thresholds. Twenty-two thematic maps derived at different HV backscatter thresholds were evaluated by comparing the land cover classes of the reference data. Sampling theory was used to determine the required sample size and geographic locations of the reference data. Among the resulting maps, a map with a threshold of - 11.5 dB HV backscatter was found to be more sensitive for portraying the spatial patterns of land covers in the study area. The overall mapping accuracy at this threshold was 79.34%. Regardless of the data size, this automated mapping approach yielded decent spatial patterns of land covers and proved to be applicable for high-resolution wall-to-wall mapping and monitoring of natural forest cover in wider tropical areas., 資料番号: PA1410016000},
 title = {Evaluation of ALOS PALSAR sensitivity for characterizing natural forest cover in wider tropical areas},
 year = {2014}
}