Optimal segmentation of high spatial resolution images for the classification of buildings using random forests.
James BialasThomas OommenTimothy C. HavensPublished in: Int. J. Appl. Earth Obs. Geoinformation (2019)
Keyphrases
- high spatial resolution
- random forests
- decision trees
- image classification
- remote sensing
- multispectral
- spatial and temporal resolution
- image data
- multispectral images
- random forest
- high resolution
- satellite images
- image features
- machine learning algorithms
- spatial resolution
- aerial images
- video camera
- image analysis
- input image
- text classification
- remote sensing images
- image registration
- hyperspectral data
- support vector
- optimal segmentation
- supervised learning
- change detection
- logistic regression
- feature selection
- hyperspectral images
- low spatial resolution
- feature set
- training set
- ensemble methods
- spatial information
- svm classifier
- low resolution
- urban areas
- active learning
- satellite imagery
- pattern recognition
- high quality
- image segmentation
- learning algorithm