OpenStreetMap: Challenges and Opportunities in Machine Learning and Remote Sensing

  title={OpenStreetMap: Challenges and Opportunities in Machine Learning and Remote Sensing},
  author={John E. Vargas-Mu{\~n}oz and Shivangi Srivastava and Devis Tuia and Alexandre Xavier Falc{\~a}o},
  journal={IEEE Geoscience and Remote Sensing Magazine},
OpenStreetMap (OSM) is a community-based, freely available, editable map service created as an alternative to authoritative sources. Given that it is edited mainly by volunteers with different mapping skills, the completeness and quality of its annotations are heterogeneous across different geographical locations. Despite that, OSM has been widely used in several applications in geosciences, Earth observation, and environmental sciences. In this article, we review recent methods based on… 

Figures from this paper


  • H. LiA. Zipf
  • Computer Science
    The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
  • 2022
A conceptual model consisting of three major components: historical OSM and external datasets, quality indicators, and GeoAI models is proposed to fill the research gap in harvesting OSM contribution as quality-aware training data for different GeoAI tasks.


A tool that provides access to currently 32 attributes or indicators at the level of single OpenStreetMap objects, and can be seamlessly applied to any polygonal Open StreetMap data and also supports linear and point data is presented.

GeoVectors: A Linked Open Corpus of OpenStreetMap Embeddings on World Scale

The GeoVectors corpus captures semantic and geographic dimensions of OSM entities and makes these entities directly accessible to machine learning algorithms and semantic applications, and provides a SPARQL endpoint that offers direct access to the semantic and latent representations of geographic entities in OSM.

De/colonizing OpenStreetMap? Local mappers, humanitarian and commercial actors and the changing modes of collaborative mapping

In its early days, the geodata and mapping project OpenStreetMap (OSM) was widely celebrated for opening up and “democratizing” the production of geographic knowledge. However, critical research

Towards an unsupervised large-scale 2D and 3D building mapping with airborne LiDAR data

A 2D and 3D building map provides invaluable information for understanding human activities and their impacts on the Earth and its environment. Despite enormous efforts to improve the quality of

Urban land-use analysis using proximate sensing imagery: a survey

This paper reviews and summarizes the state-of-the-art methods and publicly available data sets from proximate sensing to support land-use analysis and identifies several research problems in the perspective of examples to support the training of models and means of integrating diverse data sets.

An Automatic Approach to Extracting Large-Scale Three-Dimensional Road Networks Using Open-Source Data

3D road networks are amongst the indispensable elements of a smart city, which has been explored in various ways. However, researchers still faces challenges extracting 3D networks on a large scale.

A Tag Recommendation Method for OpenStreetMap Based on FP-Growth and Improved Markov Process

The aim of the algorithm is to improve the quality of OSM objects by recommending some tags when volunteers contribute to the platform, and the results show that the method based on FP-Growth and Improved Markov Process can effectively recommend tag-keys for different feature classes.



Joint Learning from Earth Observation and OpenStreetMap Data to Get Faster Better Semantic Maps

  • N. AudebertB. L. SauxS. Lefèvre
  • Computer Science, Environmental Science
    2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
  • 2017
It is shown that OpenStreetMap data can efficiently be integrated into the vision-based deep learning models and that it significantly improves both the accuracy performance and the convergence speed of the networks.

Automated highway tag assessment of OpenStreetMap road networks

A machine learning model is proposed that learns the geometrical and topological characteristics of different semantic classes of streets and is subsequently used to accurately determine if a street has been assigned a correct/incorrect semantic class.

Multi-granular Street Network Representation towards Quality Assessment of OpenStreetMap Data

A simple and intuitive representation for street networks where information like street name and category is easily available and allows for an efficient application of machine learning algorithms towards quality assessment of street networks datsets is presented.

Quality assessment of building footprint data using a deep autoencoder network

The matched results between OSM building footprints and official data are used as the samples for training an autoencoder network, which encodes and reconstructs the sample populations according to unknown complex multivariate probability distributions.

Using Crowdsourced Trajectories for Automated OSM Data Entry Approach

This paper suggests an automatic mechanism, which uses raw spatial data (trajectories of movements contributed by contributors to OSM) to minimise the uncertainty and impact of the above-mentioned issues and takes the raw trajectory datasets as input and analyses them using data mining techniques.

Exploiting deep learning and volunteered geographic information for mapping buildings in Kano, Nigeria

This work presents a new approach, coupling deep convolutional neural networks (CNNs) with VGI for automating building map generation from high-resolution satellite images for Kano state, Nigeria, and demonstrates potential advancements in current mapping capabilities in resource constrained countries.

Quality Assessment of the French OpenStreetMap Dataset

The quality of French OpenStreetMap data is studied to provide a larger set of spatial data quality element assessments, and raises questions such as the heterogeneity of processes, scales of production, and the compliance to standardized and accepted specifications.

The world’s user-generated road map is more than 80% complete

Two complementary, independent methods are used to assess the completeness of OSM road data in each country in the world and find that globally, OSM is ∼83% complete, and more than 40% of countries—including several in the developing world—have a fully mapped street network.

Ambiguity and plausibility: managing classification quality in volunteered geographic information

This study uses the rich data set of OSM to analyze the properties of geographic entities with respect to their implicit characteristics in order to develop classifiers based on them and shows that classification-based approaches can be a valuable tool for managing and improving VGI data.