Very High Resolution Land Cover Mapping of Urban Areas at Global Scale with Convolutional Neural Networks
This work addresses the need for detailed urban land cover mapping, which is incremental as it applies existing convolutional neural network architectures to a specific domain with refined data processing.
The paper tackles the problem of producing a 7-class land cover map for urban areas at a global scale using very high-resolution images and limited noisy labeled data, achieving a segmentation map for a large French department with classes such as asphalt, buildings, and water.
This paper describes a methodology to produce a 7-classes land cover map of urban areas from very high resolution images and limited noisy labeled data. The objective is to make a segmentation map of a large area (a french department) with the following classes: asphalt, bare soil, building, grassland, mineral material (permeable artificialized areas), forest and water from 20cm aerial images and Digital Height Model. We created a training dataset on a few areas of interest aggregating databases, semi-automatic classification, and manual annotation to get a complete ground truth in each class. A comparative study of different encoder-decoder architectures (U-Net, U-Net with Resnet encoders, Deeplab v3+) is presented with different loss functions. The final product is a highly valuable land cover map computed from model predictions stitched together, binarized, and refined before vectorization.