Details
| Original language | English |
|---|---|
| Article number | 6767102 |
| Pages (from-to) | 917-928 |
| Number of pages | 12 |
| Journal | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Volume | 7 |
| Issue number | 3 |
| Publication status | Published - Mar 2014 |
Abstract
For many ecological applications of remote sensing, traditional multispectral data with moderate spatial and spectral resolution have to be used. Typical examples are land-use change or deforestation assessments. The study sites are frequently too large and the timespan covered too long assumes the availability of modern datasets such as very high resolution or hyperspectral data. However, in traditional datasets such as Landsat data, separability of the relevant classes is limited. A promising approach is to describe the landscape context pixels that are integrated. For this purpose, multiscale context features are computed. Then, spectral-spatial classification is employed. However, such approaches require sophisticated processing techniques. This study exemplifies these issues by designing an entire framework for exploiting context features. The framework uses kernel-based classifiers which are unified by a multiple classifier system and further improved by conditional random fields. Accuracy on three scenarios is raised between 19.0%pts and 26.6%pts. Although the framework is designed, focusing an application in Chile, it is generally enough to be applied to similar scenarios.
Keywords
- Conditional random fields (CRFs), extended morphological profiles (EMPs), import vector machines (IVM), kernel composition, support vector machines (SVMs)
ASJC Scopus subject areas
- Earth and Planetary Sciences(all)
- Computers in Earth Sciences
- Earth and Planetary Sciences(all)
- Atmospheric Science
Sustainable Development Goals
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 7, No. 3, 6767102, 03.2014, p. 917-928.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Design of a spectral-spatial pattern recognition framework for risk assessments using landsat data
T2 - A case study in Chile
AU - Braun, Andreas Christian
AU - Rojas, Carolina
AU - Echeverri, Cristian
AU - Rottensteiner, Franz
AU - Bähr, Hans Peter
AU - Niemeyer, Joachim
AU - Arias, Mauricio Aguayo
AU - Kosov, Sergey
AU - Hinz, Stefan
AU - Weidner, Uwe
PY - 2014/3
Y1 - 2014/3
N2 - For many ecological applications of remote sensing, traditional multispectral data with moderate spatial and spectral resolution have to be used. Typical examples are land-use change or deforestation assessments. The study sites are frequently too large and the timespan covered too long assumes the availability of modern datasets such as very high resolution or hyperspectral data. However, in traditional datasets such as Landsat data, separability of the relevant classes is limited. A promising approach is to describe the landscape context pixels that are integrated. For this purpose, multiscale context features are computed. Then, spectral-spatial classification is employed. However, such approaches require sophisticated processing techniques. This study exemplifies these issues by designing an entire framework for exploiting context features. The framework uses kernel-based classifiers which are unified by a multiple classifier system and further improved by conditional random fields. Accuracy on three scenarios is raised between 19.0%pts and 26.6%pts. Although the framework is designed, focusing an application in Chile, it is generally enough to be applied to similar scenarios.
AB - For many ecological applications of remote sensing, traditional multispectral data with moderate spatial and spectral resolution have to be used. Typical examples are land-use change or deforestation assessments. The study sites are frequently too large and the timespan covered too long assumes the availability of modern datasets such as very high resolution or hyperspectral data. However, in traditional datasets such as Landsat data, separability of the relevant classes is limited. A promising approach is to describe the landscape context pixels that are integrated. For this purpose, multiscale context features are computed. Then, spectral-spatial classification is employed. However, such approaches require sophisticated processing techniques. This study exemplifies these issues by designing an entire framework for exploiting context features. The framework uses kernel-based classifiers which are unified by a multiple classifier system and further improved by conditional random fields. Accuracy on three scenarios is raised between 19.0%pts and 26.6%pts. Although the framework is designed, focusing an application in Chile, it is generally enough to be applied to similar scenarios.
KW - Conditional random fields (CRFs)
KW - extended morphological profiles (EMPs)
KW - import vector machines (IVM)
KW - kernel composition
KW - support vector machines (SVMs)
UR - http://www.scopus.com/inward/record.url?scp=84897016847&partnerID=8YFLogxK
U2 - 10.1109/JSTARS.2013.2293421
DO - 10.1109/JSTARS.2013.2293421
M3 - Article
AN - SCOPUS:84897016847
VL - 7
SP - 917
EP - 928
JO - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
JF - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
SN - 1939-1404
IS - 3
M1 - 6767102
ER -