Hyperspectral Image Segmentation Using Normalized Cuts

Torun O., YÜKSEL ERDEM S. E.

24th Signal Processing and Communication Application Conference (SIU), Zonguldak, Turkey, 16 - 19 May 2016, pp.1717-1720 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/siu.2016.7496090
  • City: Zonguldak
  • Country: Turkey
  • Page Numbers: pp.1717-1720
  • Hacettepe University Affiliated: Yes

Abstract

The segmentation of hyperspectral images based on spatial and spectral information is being used in many fields from target detection to classification. In this paper, the normalized cuts algoritm is extended for the segmentation of hyperspectral images. The affinity matrix of the normalized cuts algorithm is modified such that it uses both the spatial and spectral information. The effect of the parameters on the segmentation results is investigated, and it is shown that with proper parameter selection, the proposed method can segment the spectrally similar materials, and also extract the targets that we are after.