Target Detection in Hyperspectral Imagery Combining Spectral and Spatial Features


Creative Commons License

Saltürk S.

International Conference on Future Computational Technologies, İstanbul, Turkey, 30 - 31 May 2016, vol.1, no.1, pp.236-239

  • Publication Type: Conference Paper / Full Text
  • Volume: 1
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.236-239

Abstract

—Hyperspectral images are high dimensional data compared to RGB and gray level images. As well as these images have lots of decisive feature in spectral domain, they have plenty of information about the target also in spatial domain. Hence, performance of target detection algorithms can be improved when spatial and spectral features are employed together. In this paper, Covariance Descriptor method which enable combining of spectral and spatial features are used for target detection. In addition, Kernel Covariance Descriptor technique is developed utilizing kernel approach. Efficient spectral and spatial features are determined by experimental studies. The target detection performance is compared with each other. The experiments on two real hyperspectral data sets demonstrate that the proposed scheme is effective for target detection problems