Support vector machine fusion of multisensor imagery in tropical ecosystems

One of the major stakeholders of image fusion is being able to process the most complex images at the finest possible integration level and with the most reliable accuracy. The use of support vector machine (SVM) fusion for the classification of multisensors images representing a complex tropical ecosystem is investigated. First, SVM are trained individually on a set of complementary sources: multispectral, synthetic aperture radar (SAR) images and a digital elevation model (DEM). Then a SVM-based decision fusion is performed on the three sources. SVM fusion outperforms all monosource classifications outputting results with the same accuracy as the majority of other comparable studies on cultural landscapes. SVM-based hybrid consensus classification does not only balance successful and misclassified results, it also uses misclassification patterns as information. Such a successful approach is partially due to the integration of DEM-extracted indices which are relevant to land cover mapping in non-cultural and topographically complex landscapes.

[1]  Johannes R. Sveinsson,et al.  Support vector machines in multisource classification , 2003, IGARSS 2003. 2003 IEEE International Geoscience and Remote Sensing Symposium. Proceedings (IEEE Cat. No.03CH37477).

[2]  Luis Marcelo Tavares de Carvalho,et al.  Performance evaluation of several adaptive speckle filters for SAR imaging , 2008 .

[3]  Jon Atli Benediktsson,et al.  Fusion of Support Vector Machines for Classification of Multisensor Data , 2007, IEEE Transactions on Geoscience and Remote Sensing.

[4]  John A. Richards,et al.  Analysis of remotely sensed data: the formative decades and the future , 2005, IEEE Transactions on Geoscience and Remote Sensing.

[5]  G. Mroz,et al.  Microclimate in Forest Ecosystem and Landscape Ecology , 1999 .

[6]  Giles M. Foody,et al.  The use of small training sets containing mixed pixels for accurate hard image classification: Training on mixed spectral responses for classification by a SVM , 2006 .

[7]  Johannes R. Sveinsson,et al.  Classifying Remote Sensing Data with Support Vector Machines and Imbalanced Training Data , 2009, MCS.

[8]  Zhenghao Shi,et al.  A comparison of digital speckle filters , 1994, Proceedings of IGARSS '94 - 1994 IEEE International Geoscience and Remote Sensing Symposium.

[9]  Bernhard Schölkopf,et al.  Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond , 2005, IEEE Transactions on Neural Networks.

[10]  Yichun Xie,et al.  Remote sensing imagery in vegetation mapping: a review , 2008 .

[11]  Jean-Claude Souyris,et al.  Support Vector Machine for Multifrequency SAR Polarimetric Data Classification , 2009, IEEE Transactions on Geoscience and Remote Sensing.

[12]  Chih-Jen Lin,et al.  A Practical Guide to Support Vector Classication , 2008 .

[13]  L. Nagy.  Alpine Biodiversity in Europe , 2003, Ecological Studies.

[14]  P. Swain,et al.  Neural Network Approaches Versus Statistical Methods In Classification Of Multisource Remote Sensing Data , 1990 .

[15]  G. Salvucci,et al.  Eagleson’s optimality theory of an ecohydrological equilibrium: quo vadis? , 1997 .

[16]  Vladimir Vapnik,et al.  Statistical learning theory , 1998 .

[17]  Russell G. Congalton,et al.  Assessing the accuracy of remotely sensed data : principles and practices , 1998 .

[18]  D. Ducrot,et al.  Land cover discrimination potential of radar multitemporal series and optical multispectral images in a Mediterranean cultural landscape , 2004 .

[19]  Stefanie Eberhardt.  Support Vector Machines For Pattern Recognition , 2006 .

[20]  Xavier Blaes,et al.  Efficiency of crop identification based on optical and SAR image time series , 2005 .

[21]  Alexander J. Smola,et al.  Learning with kernels , 1998 .

[22]  Jon Atli Benediktsson,et al.  Classification of multisource and hyperspectral data based on decision fusion , 1999, IEEE Trans. Geosci. Remote. Sens..

[23]  D. Michelson.  Comparison of Algorithms for Classifying Swedish Landcover Using Landsat TM and ERS-1 SAR Data , 2000 .

[24]  M. Fladeland,et al.  Remote sensing for biodiversity science and conservation , 2003 .