High-Resolution Remote Sensing Data Classification over Urban Areas Using Random Forest Ensemble and Fully Connected Conditional Random Field.
Xiaofeng SunXiangguo LinShuhan ShenZhanyi HuPublished in: ISPRS Int. J. Geo Inf. (2017)
Keyphrases
- random forest
- conditional random fields
- fully connected
- decision trees
- urban areas
- feature set
- remote sensing
- remote sensing images
- ensemble methods
- satellite images
- high resolution
- hidden markov models
- probabilistic model
- land cover
- classification accuracy
- aerial images
- machine learning methods
- multispectral
- graphical models
- feature extraction
- information extraction
- markov random field
- pairwise
- higher order
- base classifiers
- feature vectors
- generative model
- segmentation method
- text classification
- multi label
- pattern recognition
- class labels
- training data
- feature space
- image processing
- feature selection
- machine learning
- change detection
- support vector machine svm
- computer vision
- training set
- classification algorithm
- neural network
- multi class