Target Detection in Hyperspectral Images

Karakaya A., YÜKSEL ERDEM S. E.

24th Signal Processing and Communication Application Conference (SIU), Zonguldak, Türkiye, 16 - 19 Mayıs 2016, ss.1501-1504 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası:
  • Doi Numarası: 10.1109/siu.2016.7496036
  • Basıldığı Şehir: Zonguldak
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.1501-1504
  • Hacettepe Üniversitesi Adresli: Evet

Özet

Target detection in hyperspectral images is important in many applications including search and rescue operations, defense systems, mineral exploration, mine detection and border security. In this study, the goal is to detect the nine sub-pixel targets, from seven different materials, that are placed around the town. For this purpose, eight hyperspectral target detection algorithms are compared and the three most successful algorithms are fused together. The results are compared with ROC curves, and it is found that the fusion of signed ACE, CEM and AMSD algorithms can achieve very successfull results in comparison to the other algorithms.