著者

イクサン ムハマッド 中越 信和

出版者

日本森林学会

雑誌

日本森林学会大会発表データベース

巻号頁・発行日

vol.124, 2013

---

Precise forest maps are needed for forest planning. This study compared an accuracy of visual and automated forest boundary delineations using a Root Mean Square Error (RMSE) assessment in the Patenggang Conservation Area. On-screen digitization was utilized as the visual delineation method and Maximum Likelihood was used as the automated delineation method. The results showed that visual was more accurate than automated delineation. The RMSE of visual delineation on World View-2 image were 6.0 m and 9.3 m, while on Landsat ETM+7 image were 21.3 m and 21.5 m. The RMSE of automated delineation on World view-2 image were 16.05 m and 20.75 m, while on Landsat ETM+7 image were 32.45 m and 37.30 m. The study clarified that visual delineation using World View-2 image is the most accurate method for forest delineation.