A Study on Linear Spectral Mixing Model for Hyperspectral Imagery with Geometric Method

기하학적 기법을 이용한 하이퍼스펙트럴 영상의 Linear Spectral Mixing모델에 관한 연구

  • 장은석 (서울대학교 지구환경시스템공학부 공간정보연구실) ;
  • 김대성 (서울대학교 지구환경시스템공학부 공간정보연구실) ;
  • 김용일 (서울대학교 지구환경시스템공학부 공간정보연구실)
  • Published : 2003.11.01


Detection in remotely sensed images can be conducted spatially, spectrally or both [2]. If the images have high spatial resolution, materials can be detected by using spatial and spectral information, unless we can't see the object embedded in a pixel. In this paper, we intend to solve the limit of spatial resolution by using the hyperspectral image which has high spectral resolution. Therefore, the Linear Spectral Mixing(LSM) Model which is sub-pixel detection algorithm is used to solve this problem. To find class Endmembers, we applied Geometric Model with MNF(Minimum Noise Fraction) transformation. From the result of sub-pixel detection algorithm, we can see the detection of water is satisfied and the object shape cannot be extracted but the possibility of material existence can be identified.