Chapter 13
Support Vector Machines for Land Cover
Mapping from Remote Sensor Imagery
Dee Shi and Xiaojun Yang
Abstract Land cover mapping is an important activity leading to the generation of
various thematic products essential for numerous environmental monitoring and
resources management applications at local, regional, and global levels. Over the
years, various pattern recognition techniques have been developed to automate this
process from remote sensor imagery. Support vector machines (SVM) as a group of
relatively novel statistical learning algorithms have demonstrated their robustness
in classifying homogeneous and heterogeneous land cover types. In this chapter, we
review the status and potential challenges in the SVM implementation for land
cover classification. The chapter is organized into two major parts. The first part
reviews the research status of using SVM for land cover classification, focusing on
some comparative studies that demonstrated the algorithm effectiveness over other
conventional classifiers. We identify several areas for additional work, which are
mostly related to appropriate treatments of some parametric and non-parametric
factors in order to achieve improved mapping accuracies particularly for working
over heterogeneous landscapes. Then, we implement the support vector machine
technique to map various land cover types from a satellite image covering an urban
area, and demonstrate the robustness of this pattern recognition technique for
mapping heterogeneous landscapes.
Keywords Land cover • Image classification • Support vector machines •
Heterogeneous landscapes • Thematic accuracy assessment
13.1 Introduction
Land cover is the pattern of ecological resources and human activities dominating
different areas of Earth’s surface (Turner and Meyer 1994). It is a critical type of
data source essential for many environmental monitoring and natural resources
management applications at local, regional, and global scales (Foley et al. 2005;
D. Shi (*) • X. Yang
Department of Geography, Florida State University, Tallahassee, FL 32306, USA
e-mail: ds10f@my.fsu.edu; xyang@fsu.edu
© Springer Science+Business Media Dordrecht 2015
J. Li, X. Yang (eds.), Monitoring and Modeling of Global Changes:
A Geomatics Perspective, Springer Remote Sensing/Photogrammetry,
DOI 10.1007/978-94-017-9813-6_13
265
Support Vector Machines for Land Cover
Mapping from Remote Sensor Imagery
Dee Shi and Xiaojun Yang
Abstract Land cover mapping is an important activity leading to the generation of
various thematic products essential for numerous environmental monitoring and
resources management applications at local, regional, and global levels. Over the
years, various pattern recognition techniques have been developed to automate this
process from remote sensor imagery. Support vector machines (SVM) as a group of
relatively novel statistical learning algorithms have demonstrated their robustness
in classifying homogeneous and heterogeneous land cover types. In this chapter, we
review the status and potential challenges in the SVM implementation for land
cover classification. The chapter is organized into two major parts. The first part
reviews the research status of using SVM for land cover classification, focusing on
some comparative studies that demonstrated the algorithm effectiveness over other
conventional classifiers. We identify several areas for additional work, which are
mostly related to appropriate treatments of some parametric and non-parametric
factors in order to achieve improved mapping accuracies particularly for working
over heterogeneous landscapes. Then, we implement the support vector machine
technique to map various land cover types from a satellite image covering an urban
area, and demonstrate the robustness of this pattern recognition technique for
mapping heterogeneous landscapes.
Keywords Land cover • Image classification • Support vector machines •
Heterogeneous landscapes • Thematic accuracy assessment
13.1 Introduction
Land cover is the pattern of ecological resources and human activities dominating
different areas of Earth’s surface (Turner and Meyer 1994). It is a critical type of
data source essential for many environmental monitoring and natural resources
management applications at local, regional, and global scales (Foley et al. 2005;
D. Shi (*) • X. Yang
Department of Geography, Florida State University, Tallahassee, FL 32306, USA
e-mail: ds10f@my.fsu.edu; xyang@fsu.edu
© Springer Science+Business Media Dordrecht 2015
J. Li, X. Yang (eds.), Monitoring and Modeling of Global Changes:
A Geomatics Perspective, Springer Remote Sensing/Photogrammetry,
DOI 10.1007/978-94-017-9813-6_13
265