Improvements on Classification by Tolerating NoData Values - Application to a Hybrid Classifier to Discriminate Mediterranean Vegetation with a Detailed Legend Using Multitemporal Series of Images.
Gerard MoréXavier PonsPere SerraPublished in: IGARSS (2006)
Keyphrases
- image classification
- satellite images
- multiple classifiers
- classification process
- remotely sensed images
- extracted features
- classification method
- svm classifier
- image registration
- classification scheme
- class labels
- remotely sensed
- decision trees
- classification rate
- classification algorithm
- change detection
- supervised classification
- image data
- input image
- image database
- image analysis
- classification accuracy
- remote sensing
- support vector machine
- training set
- feature space
- learning vector quantization
- feature extraction
- machine learning
- class conditional densities
- pattern recognition
- training samples
- image retrieval
- image features
- multispectral
- discriminative classifiers
- training data
- hyperspectral images
- pixel values
- fisher linear discriminant
- classification models
- classifier combination
- supervised classifiers
- k nearest neighbour
- feature selection
- feature values
- support vector
- multiple classifier systems
- individual classifiers
- text classification
- object recognition
- feature set
- remote sensing images
- naive bayes
- minimum distance classifier
- higher classification accuracy
- learning algorithm
- remotely sensed data
- land cover
- classifier ensemble
- ensemble classifier
- decision boundary
- feature vectors
- supervised learning
- hyperspectral
- feature subset
- image set