Swin transformers make strong contextual encoders for VHR image road extraction.
Tao ChenDaguang JiangRuirui LiPublished in: CoRR (2022)
Keyphrases
- aerial images
- road extraction
- remotely sensed images
- change detection
- remote sensing
- satellite imagery
- remote sensing images
- image features
- input image
- urban areas
- image representation
- image data
- image content
- image classification
- single image
- image retrieval
- segmentation method
- line segments
- multispectral
- image regions
- data sets
- denoising
- high resolution
- multispectral images
- image analysis
- data analysis
- image processing