Building extraction from high spatial resolution remote sensing images is a hot spot in the field of remote sensing applications and computer vision. This paper presents a semantic segmentation model, which is a supervised method, named Pyramid Self-Attention Network (PISANet). Its structure is simple, because it contains only two parts: one is the backbone of the network, which is used to learn the local features (short distance context information around the pixel) of buildings from the image; the other part is the pyramid self-attention module, which is used to obtain the global features (long distance context information with other pixels in the image) and the comprehensive features (includes color, texture, geometric and high-level semantic feature) of the building. The network is an end-to-end approach. In the training stage, the input is the remote sensing image and corresponding label, and the output is probability map (the probability that each pixel is or is not building). In the prediction stage, the input is the remote sensing image, and the output is the extraction result of the building. The complexity of the network structure was reduced so that it is easy to implement. The proposed PISANet was tested on two datasets. The result shows that the overall accuracy reached 94.50 and 96.15%, the intersection-over-union reached 77.45 and 87.97%, and F1 index reached 87.27 and 93.55%, respectively. In experiments on different datasets, PISANet obtained high overall accuracy, low error rate and improved integrity of individual buildings.
Keywords: building extraction; deep learning; high resolution image; semantic segmentation.