A Novel Approach to the Selection of Spatially Invariant Features for Classification of Hyperspectral Images.
Claudio PerselloLorenzo BruzzonePublished in: IGARSS (2) (2009)
Keyphrases
- hyperspectral images
- invariant features
- hyperspectral data
- hyperspectral
- remote sensing
- hyperspectral imagery
- feature space
- multispectral
- image classification
- target detection
- shape features
- feature vectors
- feature extraction
- feature selection
- classification accuracy
- image matching
- image analysis
- training set
- pattern recognition
- support vector
- multiscale