A fast and effective deep learning approach for road extraction from historical maps by automatically generating training data with symbol reconstruction

Jiao, Chenjing; Heitzler, Magnus; Hurni, Lorenz

doi:10.1016/j.jag.2022.102980

Cited by 16 publications

(7 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To verify the building footprint extraction with a contemporary map, the historical map was georeferenced by determining multiple control points visible in both maps. Automatic segmentation of historical maps is considered a difficult challenge due to vast changes in appearance between maps of different epochs [164,288]. While one approach may be successful for the 1936 map, it may easily fail for another map.…”

Section: Data Processing 421 Building Footprint Extraction From Histo...mentioning

confidence: 99%

A Digital 4D Information System on the World Scale: Research Challenges, Approaches, and Preliminary Results

Münster,

Maiwald,

Bruschke

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

Numerous digital media repositories have been set up during recent decades, each containing plenty of data about historic cityscapes. In contrast, digital 3D reconstructions of no longer extant cityscapes have so far almost solely been performed for specific cases and via manual modelling techniques. Within the Jena4D research group, we are investigating and developing methods and technologies for transferring historical media and their contextual information into 4D models. The next step will be to automate this process and extend it to the world scale. Our team are working on different facets of that endeavor: retrieving images, text, and object data from public sources; crowdsourcing and collecting qualitative data from contests and pupil labs; processing historical plans and photographs to extract spatial data; location-based mobile visualization; and collection-browsing interfaces. This article is intended to highlight (1) current challenges, (2) the state of the art, (3) our approach to 4D modelling and visualization on the world scale, and (4) results from testing demo cases in Dresden, Jena, and Amsterdam.

show abstract

Section: Data Processing 421 Building Footprint Extraction From Histo...mentioning

confidence: 99%

A Digital 4D Information System on the World Scale: Research Challenges, Approaches, and Preliminary Results

Münster,

Maiwald,

Bruschke

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…Furthermore, there are a few state-of-the-art studies using machine learning methods for map digitization. For example, text detection in historic maps 38 , segmentation and digitization of historic maps 39 , 40 , feature recognition 41 , 42 , road extraction 43 , detecting road types 44 , and cadastral boundary extraction 45 , 46 using neural network, deep neural network, and convolutional neural network models.…”

Section: Background and Summarymentioning

confidence: 99%

Creating a dataset of historic roads in Sydney from scanned maps

Turner,

Lahoorpoor,

Levinson

2023

Sci Data

View full text Add to dashboard Cite

This study creates a historic dataset of road opening dates in Sydney. A method was developed for map digitization to extract spatial data from historic maps and place them in a collective vector layer. The method includes extensive georeferencing of the maps, as well as editing and cleaning the maps through raster and vector analysis. Preferred methods for map digitization used in the project were identified. For a considerable area of Sydney, in which approximately 52000 road links were included, almost half of the links were identified with an open date by the start of the twentieth century. A further half of these links were confined to opening within a thirty-year period. The project has established a strong foundation for a historic road dataset for Sydney. It has also outlined methods and procedures that can be followed to progress the dataset further.

show abstract

“…Traditionally, the logit models such as the MNL model, the nested logit model, and the mixed logit model, are probably one of the most commonly-used travel mode choice models (Zhao et al, 2020). In recent years, machine learning has been popularized and pervasive in many fields, including but not limited to transportation, such as transportation mode recognition (Jahangiri and Rakha, 2015), traffic flow prediction (Pun et al, 2019), road extraction (Jiao et al, 2022), etc. A series of recent studies have indicated that machine learning can outperform logit models in travel mode choice modeling.…”

Section: Travel Mode Choicementioning

confidence: 99%

Modeling the choice of shared micro-mobility services using XGBoost machine learning algorithm

Ren¹,

Zhao²,

Mansourian³

2023

AGILE GIScience Ser.

View full text Add to dashboard Cite

Abstract. In recent years, shared micro-mobility services (e.g., bikes, e-bikes, and e-scooters) have been popularized at a rapid pace worldwide, which provide more choices for people’s short and medium-distance travel. Accurately modeling the choice of these shared micro-mobility services is important for their regulation and management. However, little attention has been paid to modeling their choice, especially with machine learning. In this paper, we explore the potential of the XGBoost model to model the three types of shared micro-mobility services, including docked bike, docked e-bike, and dockless e-scooter, in Zurich, Switzerland. The model achieves an accuracy of 72.6%. Moreover, the permutation feature importance is implemented to interpret the model prediction. It is found that trip duration, trip distance, and difference in elevation present higher feature importance in the prediction. The findings are beneficial for urban planners and operators to further improve the shared micro-mobility services toward sustainable urban mobility.

show abstract

A fast and effective deep learning approach for road extraction from historical maps by automatically generating training data with symbol reconstruction

Cited by 16 publications

References 37 publications

A Digital 4D Information System on the World Scale: Research Challenges, Approaches, and Preliminary Results

A Digital 4D Information System on the World Scale: Research Challenges, Approaches, and Preliminary Results

Creating a dataset of historic roads in Sydney from scanned maps

Modeling the choice of shared micro-mobility services using XGBoost machine learning algorithm

Contact Info

Product

Resources

About