A fast and effective deep learning approach for road extraction from historical maps by automatically generating training data with symbol reconstruction.
Chenjing JiaoMagnus HeitzlerLorenz HurniPublished in: Int. J. Appl. Earth Obs. Geoinformation (2022)
Keyphrases
- automatically generating
- deep learning
- training data
- road extraction
- unsupervised learning
- automatically generated
- machine learning
- learning algorithm
- training set
- aerial images
- unsupervised feature learning
- data sets
- training examples
- supervised learning
- decision trees
- text classification
- learning strategies
- spatial resolution
- high resolution
- high dimensional
- pattern recognition
- image processing