U-Net-Id, an Instance Segmentation Model for Building Extraction from Satellite Images—Case Study in the Joanópolis City, Brazil

Abstract

Currently, there exists a growing demand for individual building mapping in regions of rapid urban growth in less-developed countries. Most existing methods can segment buildings but cannot discriminate adjacent buildings. Here, we present a new convolutional neural network architecture (CNN) called U-net-id that performs building instance segmentation. The proposed network is trained with WorldView-3 satellite RGB images (0.3 m) and three different labeled masks. The first is the building mask; the second is the border mask, which is the border of the building segment with 4 pixels added outside and 3 pixels inside; and the third is the inner segment mask, which is the segment of the building diminished by 2 pixels. The architecture consists of three parallel paths, one for each mask, all starting with a U-net model. To accurately capture the overlap between the masks, all activation layers of the U-nets are copied and concatenated on each path and sent to two additional convolutional layers before the output activation layers. The method was tested with a dataset of 7563 manually delineated individual buildings of the city of Joanópolis-SP, Brazil. On this dataset, the semantic segmentation showed an overall accuracy of 97.67% and an F1-Score of 0.937 and the building individual instance segmentation showed good performance with a mean intersection over union (IoU) of 0.582 (median IoU = 0.694).

Semantic Segmentation Deep Learning for Extracting Surface Mine Extents from Historic Topographic Maps

“…Reinforcing the findings of Uhl et al [25] specifically, our study highlights the value of large, quality training datasets for training DL semantic segmentation algorithms to recognize features in topographic maps. Lastly, our study reinforces the documented strong performance of the UNet semantic segmentation method for extracting features and classifying pixels from a wide variety of data sources to support varying mapping tasks (for example, [54,55,[62][63][64][65][66][67]69,70,[91][92][93]). Such techniques, including future advancements and modifications, may eventually replace traditional ML methods, such as random forests (RF) and support vector machines (SVM), as operational standards in the field [46,[52][53][54][55].…”

“…UNet and other semantic segmentation methods have been applied to a variety of feature extraction and classification problems and have also been applied to a variety of geospatial and remotely sensed data. For example, modifications of UNet have been applied to the mapping of general land cover change [62], coastal wetlands [63], palm trees [64], cloud and cloud shadows [65], urban buildings and change detection [66][67][68], roads [69], and landslides [70]. Generally, UNet and other FCNs have shown great promise due to their ability to model complex spatial patterns and context while generating data abstractions that generalize well to new data [54,55].…”

Joint Learning of Contour and Structure for Boundary-Preserved Building Extraction

“…With the development of DCNNs in recent years, many algorithms have been proposed for processing remote sensing images [25][26][27][28][29][30][31][32]. The fully convolutional network [33] (FCN) replaces the fully connected layers with convolutional layers, making it possible for large-scale dense prediction.…”

A Deep Learning Approach to an Enhanced Building Footprint and Road Detection in High-Resolution Satellite Imagery

“…Promising building footprint detection approaches have been proposed in the literature. Wagner et al [40] presented a modified U-Net capable of discriminating between adjacent buildings. To incorporate the structure information of buildings, Hui et al [41] opted for a multi-task learning strategy, replacing the vanilla U-Net encoder with an Xception module.…”