A novel unsupervised classification approach for hyperspectral imagery based on spectral mixture model and MARKOV random field.
Yuan FangLinlin XuXiao SunLongshan YangYujia ChenJunhuan PengPublished in: IGARSS (2016)
Keyphrases
- mixture model
- markov random field
- hyperspectral imagery
- unsupervised classification
- hyperspectral images
- unsupervised learning
- hyperspectral
- em algorithm
- multispectral
- higher order
- remote sensing
- probabilistic model
- pairwise
- expectation maximization
- graph cuts
- image segmentation
- generative model
- target detection
- maximum a posteriori
- remote sensing images
- model selection
- multispectral images
- maximum likelihood
- language model
- spatial resolution
- supervised classification
- infrared
- high resolution
- high quality
- data clustering
- object detection
- clustering ensemble
- pattern recognition
- co occurrence
- change detection
- computer vision