Spectral Derivative Features for Classification of Hyperspectral Remote Sensing Images: Experimental Evaluation

Derivatives of spectral reflectance signatures can capture salient features of different land-cover classes. Such information has been used for supervised classification of remote sensing data along with spectral reflectance. In the paper, we study how supervised classification of hyperspectral remote sensing data can benefit from the use of derivatives of spectral reflectance without the aid of other techniques, such as dimensionality reduction and data fusion. An empirical conclusion is given based on a large amount of experimental evaluations carried out on three real hyperspectral remote sensing data sets. The experimental results show that when a training data set is of a small size or the quality of the data is poor, the use of additional first order derivatives can significantly improve classification accuracies along with original spectral features when using classifiers which can avoid the “curse of dimensionality,” such as the SVM algorithm.

[1]  Robert E. Wolfe,et al.  An Enhanced TIMESAT Algorithm for Estimating Vegetation Phenology Metrics From MODIS Data , 2011, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[2]  Peter E. Hart,et al.  Nearest neighbor pattern classification , 1967, IEEE Trans. Inf. Theory.

[3]  Benoit Rivard,et al.  Derivative spectral unmixing of hyperspectral data applied to mixtures of lichen and rock , 2004, IEEE Transactions on Geoscience and Remote Sensing.

[4]  Jon Atli Benediktsson,et al.  Spectral derivative features for supervised classification of remote sensing data: An experimental evaluation , 2012, 2012 4th Workshop on Hyperspectral Image and Signal Processing (WHISPERS).

[5]  Joydeep Ghosh,et al.  Investigation of the random forest framework for classification of hyperspectral data , 2005, IEEE Transactions on Geoscience and Remote Sensing.

[6]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[7]  Begüm Demir,et al.  Spectral Magnitude and Spectral Derivative Feature Fusion for Improved Classification of Hyperspectral Images , 2008, IGARSS 2008 - 2008 IEEE International Geoscience and Remote Sensing Symposium.

[8]  Eyal Ben-Dor,et al.  Use of Derivative Calculations and Minimum Noise Fraction Transform for Detecting and Correcting the Spectral Curvature Effect (Smile) in Hyperion Images , 2010, IEEE Transactions on Geoscience and Remote Sensing.

[9]  William Philpot,et al.  The derivative ratio algorithm: avoiding atmospheric effects in remote sensing , 1991, IEEE Trans. Geosci. Remote. Sens..

[10]  Saurabh Prasad,et al.  Decision-Level Fusion of Spectral Reflectance and Derivative Information for Robust Hyperspectral Land Cover Classification , 2010, IEEE Transactions on Geoscience and Remote Sensing.

[11]  Fuan Tsai,et al.  Derivative analysis of hyperspectral data for detecting spectral features , 1997, IGARSS'97. 1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings. Remote Sensing - A Scientific Vision for Sustainable Development.

[12]  Ye Zhang,et al.  An improved spectral reflectance and derivative feature fusion for hyperspectral image classification , 2011, 2011 IEEE International Geoscience and Remote Sensing Symposium.