Automated Extraction of Human Settlement Patterns From Historical Topographic Map Series Using Weakly Supervised Convolutional Neural Networks

Abstract

Information extraction from historical maps represents a persistent challenge due to inferior graphical quality and the large data volume of digital map archives, which can hold thousands of digitized map sheets. Traditional map processing techniques typically rely on manually collected templates of the symbol of interest, and thus are not suitable for large-scale information extraction. In order to digitally preserve such large amounts of valuable retrospective geographic information, high levels of automation are required. Herein, we propose an automated machine-learning based framework to extract human settlement symbols, such as buildings and urban areas from historical topographic maps in the absence of training data, employing contemporary geospatial data as ancillary data to guide the collection of training samples. These samples are then used to train a convolutional neural network for semantic image segmentation, allowing for the extraction of human settlement patterns in an analysis-ready geospatial vector data format. We test our method on United States Geological Survey historical topographic maps published between 1893 and 1954. The results are promising, indicating high degrees of completeness in the extracted settlement features (i.e., recall of up to 0.96, F-measure of up to 0.79) and will guide the next steps to provide a fully automated operational approach for large-scale geographic feature extraction from a variety of historical map series. Moreover, the proposed framework provides a robust approach for the recognition of objects which are small in size, generalizable to many kinds of visual documents.INDEX TERMS Convolutional neural networks, digital humanities, digital preservation, document analysis, geospatial analysis, geospatial artificial intelligence, human settlement patterns, image analysis, weakly supervised learning.

Automatic Detection of Road Types From the Third Military Mapping Survey of Austria-Hungary Historical Map Series With Deep Convolutional Neural Networks

“…Uhl et al used USGS maps for human settlement footprint retrieval in their studies [3], [2]. By applying a weakly supervised CNN, they achieved promising results (i.e., recall of up to 0.96, F_measure of up to 0.79).…”

“…As far as the complexity is concerned, our model has 32.8M parameters. When we searched the literature for historical map processing, the training times and complexity of models are not generally reported but Uhl et al [2] reported that 138 million parameters used in their VGGNet-16 models. Our model has relatively lower number of parameters, training time and complexity because of the pretrained architectures.…”

Automatic Road Extraction from Historical Maps Using Deep Learning Techniques: A Regional Case Study of Turkey in a German World War II Map

“…Most of the accessible historical maps are digitally available only as scanned images; therefore, it is not possible to conduct quantitative and geometrical analysis from these maps without further processing [6]. However, multi-date historical maps or integrated usage of historical maps, aerial photographs, and satellite images can be used to extract ISPRS Int.…”

“…In particular, they explained the CNN architectures used in semantic segmentation such as U-Net, SegNet, and DeepLab. Moreover, deep learning-based methods have started to be implemented in historical map-related tasks recently [2,6,9,15].…”

“…Uhl et al [6] proposed a weakly supervised CNN-based framework for the extraction of buildings and urban areas from the USGS historical topographic maps published between 1893 and 1954. They compared the results of the VGGNet-16, LeNet, and AlexNet architectures, and the best classification accuracy results were obtained with VGGNet-16, whereas the lowest accuracy was obtained with the shallow LeNet architecture.…”