Volume XLII-4/W18
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 41–44, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-41-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 41–44, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-41-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

  18 Oct 2019

18 Oct 2019

SPECTRAL-SPATIAL CLASSIFICATION OF HYPERSPECTRAL IMAGERY USING A HYBRID FRAMEWORK

D. Akbari1 and M. Moradizadeh2 D. Akbari and M. Moradizadeh
  • 1Dept. of Surveying and Geomatics Engineering, College of Engineering, University of Zabol, Zabol, Iran
  • 2Dept. of Geomatics, Faculty of Civil and Transportation Engineering, University of Isfahan, Isfahan, Iran

Keywords: Remote sensing, Hyperspectral image, neural network, Marker selection

Abstract. Hyperspectral Images are worthwhile data for many processing algorithms (e.g. Dimensionality Reduction, Target Detection, Change Detection, Classification and Unmixing). Classification is a key issue in processing hyperspectral images. Spectral-identification-based algorithms are sensitive to spectral variability and noise in acquisition. There are many algorithms for classification. This paper describes a new framework for classification of hyperspectral images, based on both spectral and spatial information. The spatial information is obtained by an enhanced Marker-based Hierarchical Segmentation (MHS) algorithm. The hyperspectral data is first fed into the Multi-Layer Perceptron (MLP) neural network classification algorithm. Then, the MHS algorithm is applied in order to increase the accuracy of less-accurately classified land-cover types. In the proposed approach, the markers are extracted from the classification maps obtained by MLP and Support Vector Machines (SVM) classifiers. Experimental results on Quebec City hyperspectral dataset, demonstrate that the proposed approach achieves approximately 9% and 5% better overall accuracy than the MLP and the original MHS algorithms respectively.