Estimating Percent Tree Cover Using Regression Tree Method with Very-High-Resolution QuickBird Images as Training Data

DOI Web Site 1 Citations 35 References
  • ROKHMATULOH
    Center for Environmental Remote Sensing (CEReS), Chiba University
  • NITTO Daisuke
    Center for Environmental Remote Sensing (CEReS), Chiba University
  • Al BILBISI Hussam
    Center for Environmental Remote Sensing (CEReS), Chiba University
  • ARIHARA Kota
    Center for Environmental Remote Sensing (CEReS), Chiba University
  • TATEISHI Ryutaro
    Center for Environmental Remote Sensing (CEReS), Chiba University

Bibliographic Information

Other Title
  • クイックバード画像をトレーニングデータに用いた回帰ツリー法による樹木被覆率の推定 (英文)
  • Estimating percent of tree cover using regression tree method with very-high-resolution QuickBird images as training data

Search this article

Abstract

The estimation of tree cover area at continental scale is becoming more important than before due to the needs to improve our understanding of carbon dynamics. The estimation of percent tree cover of a large area using MODIS data by regression tree method is a promising method. New points of this study are the use of QuickBird images for the collection of training data and the use of the Stepwise Linear Regression (SLR) for selecting the best subset of predictor variables. The estimation of percent tree cover of African continent was tried using 11 QuickBird images to get 195 cells as training data and 32-day composite MODIS 2003 data as predictor variables. The predictor variables consist of surface reflectance, normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), normalized difference soil index (NDSI) and land surface temperature (LST). The result shows that NDVI and surface reflectance bands are effective to estimate percent tree cover and this method is acceptable with the prediction error of 5.17%.

Journal

Citations (1)*help

See more

References(35)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top