BROWSE

Related Researcher

Author

Im, Jungho
Intelligent Remote sensing and geospatial Information Systems (IRIS) Lab
Research Interests
  • Remote sensing, Geospatial modeling, Climate change

ITEM VIEW & DOWNLOAD

Effects of national forest inventory plot location error on forest carbon stock estimation using k-nearest neighbor algorithm

Cited 0 times inthomson ciCited 0 times inthomson ci
Title
Effects of national forest inventory plot location error on forest carbon stock estimation using k-nearest neighbor algorithm
Author
Jung, JaehoonKim, SangpilHong, SungchulKim, KyoungminKim, EunsookIm, JunghoHeo, Joon
Keywords
Forest carbons; K-nearest neighbors; Location errors; National forest inventories; Uncertainty
Issue Date
201307
Publisher
ELSEVIER SCIENCE BV
Citation
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING, v.81, no., pp.82 - 92
Abstract
This paper suggested simulation approaches for quantifying and reducing the effects of National Forest Inventory (NFI) plot location error on aboveground forest biomass and carbon stock estimation using the k-Nearest Neighbor (kNN) algorithm. Additionally, the effects of plot location error in pre-GPS and GPS NFI plots were compared. Two South Korean cities, Sejong and Daejeon, were chosen to represent the study area, for which four Landsat TM images were collected together with two NFI datasets established in both the pre-GPS and GPS eras. The effects of plot location error were investigated in two ways: systematic error simulation, and random error simulation. Systematic error simulation was conducted to determine the effect of plot location error due to mis-registration. All of the NFI plots were successively moved against the satellite image in 360° directions, and the systematic error patterns were analyzed on the basis of the changes of the Root Mean Square Error (RMSE) of kNN estimation. In the random error simulation, the inherent random location errors in NFI plots were quantified by Monte Carlo simulation. After removal of both the estimated systematic and random location errors from the NFI plots, the RMSE% were reduced by 11.7% and 17.7% for the two pre-GPS-era datasets, and by 5.5% and 8.0% for the two GPS-era datasets. The experimental results showed that the pre-GPS NFI plots were more subject to plot location error than were the GPS NFI plots. This study's findings demonstrate a potential remedy for reducing NFI plot location errors which may improve the accuracy of carbon stock estimation in a practical manner, particularly in the case of pre-GPS NFI data.
URI
Go to Link
DOI
http://dx.doi.org/10.1016/j.isprsjprs.2013.04.008
ISSN
0924-2716
Appears in Collections:
UEE_Journal Papers

find_unist can give you direct access to the published full text of this article. (UNISTARs only)

Show full item record

qr_code

  • mendeley

    citeulike

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

MENU