A model on achieving higher performance in the classification of hyperspectral satellite data: a case study on Hyperion
- PDF / 1,380,263 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 111 Downloads / 183 Views
ORIGINAL PAPER
A model on achieving higher performance in the classification of hyperspectral satellite data: a case study on Hyperion data Dibyajyoti Chutia & Dhruba Kumar Bhattacharyya & Ranjan Kalita & Jonali Goswami & Puyam S. Singh & S. Sudhakar
Received: 15 August 2013 / Accepted: 7 July 2014 / Published online: 20 July 2014 # Società Italiana di Fotogrammetria e Topografia (SIFET) 2014
Abstract Hyperspectral remote sensing data is characterized by the large number of contiguous spectral bands with narrow bandwidth. Enormous information available in hyperspectral data is quite challenging for classification as compared to multispectral remote sensing data. Most of the widely used conventional ‘hard’ classifiers are producing inconsistent classification results while employed in classification of hyperspectral data. In this paper, we present an effective hyperspectral classification model for achieving higher accuracy. The proposed model is characterized by three major components: dimensionality reduction using principal component analysis (PCA), multiresolution segmentation, and fuzzy membership-based nearest neighbor (NN)-classification. Here, the bands of the dimensionality-reduced images are represented by the first principal component (PC) of each of the spectral region covered by the hyperspectral sensor. Then, multiresolution segmentation is carried out on these PC composite images based on color and shape homogeneity criterion. The conventional NN-classifier is effectively used D. Chutia (*) : R. Kalita : J. Goswami : P. S. Singh : S. Sudhakar North Eastern Space Applications Centre, Department of Space, Government of India, Umiam, Meghalaya 793103, India e-mail: [email protected] R. Kalita e-mail: [email protected] J. Goswami e-mail: [email protected] P. S. Singh e-mail: [email protected] S. Sudhakar e-mail: [email protected] D. K. Bhattacharyya Department of Computer Science and Engineering, Tezpur University, Napaam, Assam 784028, India e-mail: [email protected]
by appropriate utilization of fuzzy membership function defined on a set of optimal features derived from the segmented image objects. We demonstrate a case study on Hyperion sensor data of Earth Observing-1 (EO-1) satellite. A comparative assessment is carried out with other competing techniques such as spectral angle mapper (SAM), artificial neural network (ANN), and support vector machine (SVM) on a set of images with different land cover surfaces. The proposed classification model outperforms the existing classification approaches investigated here. Keywords Hyperspectral . PC composite images . Fuzzy membership function . Multiresolution . Optimal features
Introduction A latest advancement in remote sensing is addition of hyperspectral sensors which is able to acquire enormous amount of data simultaneously in hundreds of bands with narrow bandwidths. Hyperspectral sensors can provide detailed contiguous spectral curves and makes possible detail assessment of earth surfaces which would be otherwise not
Data Loading...
