How To Download Building Footprint Information From Google Map
How was this dataset created?
A deep learning model was trained to make up one's mind the footprints of buildings from high resolution satellite imagery. Our accompanying technical report describes the methodology.
How is the data licensed?
The data is shared under the Creative Commons Attribution (CC Past-iv.0) license and the Open up Data Commons Open Database License (ODbL) v1.0 license. As the user, you can pick which of the two licenses you adopt and use the information under the terms of that license.
Why two licenses?
Nosotros wanted to make the data compatible for ingestion by those working with ODbL-licensed datasets (namely the OpenStreetMap community) while enabling people who don't utilise ODbL licensing to utilize information technology under the terms of the CC BY-iv.0 license. We hoped to take away the burden of figuring out whether the two licenses were compatible and simply release the data set under both licenses.
Can these buildings be added to OpenStreetMap?
Yes – still, to maintain the quality of OSM, please be mindful of the need for human review when adding car-generated features, and where possible to practise this with the benefit of local knowledge. Errors in the data to look out for include faux detections and inaccurate shapes (see more near accuracy below). Nosotros also recommend starting by filtering out edifice detections that have a confidence score beneath the estimated ninety% precision threshold.
Are these the same buildings as on Google Maps?
The buildings on Google Maps come from a diversity of sources, including the model used to generate this dataset. Then there is some overlap, merely the sets of footprints are not exactly the same.
Why are buildings offset from satellite imagery in some areas?
Every bit the imagery in Google Maps is updated over time, the specific images used to identify these buildings are not necessarily the same images that are currently published in Google Maps. If there is a misalignment between these two sets of imagery, buildings displayed in the information explorer map may appear to exist offset from the underlying imagery.
You can view a timeline of the imagery for a specific area using the Historical Imagery characteristic in Google Earth Pro which may testify this imagery showtime between different images and dates. To learn a niggling more about satellite imagery offset come across these sites (one, 2). Also see the technical written report for details about data limitations and quality.
Why are the edifice detections less accurate in certain urban, arid or rural areas?
Despite having a diverse prepare of preparation data, some scenarios are challenging for the building detection pipeline, including: 1) geological or vegetation features which can be dislocated with congenital structures; 2) settlements with many contiguous buildings non having articulate delineations; 3) areas characterised by small buildings, which can announced just a few pixels wide at the given image resolution; and iv) rural or desert areas, where buildings constructed with natural materials tend to visually blend into the surrounding area. Come across the technical written report for more details.
How accurate is the data?
The information is subject area to both omission and committee errors, of these types:
Imagery completeness errors: for some areas, up-to-date satellite imagery may not have been bachelor, or in that location were buildings on the basis that were not visible from the satellite image, or there was cloud cover.
Detection errors: estimated precision and remember curves for our detection model, based on a held-out test ready, are every bit shown below. The tradeoff betwixt false positives and simulated negatives varies between geographical regions.
Our model sometimes wrongly detects buildings where at that place are actually rocks or vegetation features, for example.
By choosing the confidence score threshold at which buildings are filtered out, the tradeoff between precision and recollect can be controlled. We provide suggested thresholds with each download tile to obtain guess lxxx% and ninety% precision levels.
Does dataset quality vary per location?
Yes, see plots below. To address this nosotros provide a CSV file with suggested score thresholds to obtain specific precision levels for each download tile.
How complete and up-to-date is it?
The dataset freshness is determined past the availability of the high-resolution source imagery which we use to notice buildings. While nosotros have tried to include the near recent images possible, specially in populated areas, in some cases, the most recent paradigm for some location was several years former or not available to us at all. To look at freshness for a particular surface area, the Historical Imagery part in Google World Pro shows the specific dates and imagery (check for imagery before the inference date given in the version history below). Furthermore, nosotros accept non processed imagery for the unabridged continent: to check whether a particular region has been included, the dataset explorer map above visualises all buildings in the dataset.
What is the distribution of confidence scores and building sizes?
We filtered detections to include only those with confidence score 0.6 or greater. Depending on the application, information technology is likely necessary to filter at a higher threshold (e.g. with the score thresholds above to achieve ninety% precision).
How tin can I access the data?
We currently provide 3 options:
- CSV files. See the Data format and Download sections.
- Earth Engine asset. See this catalog page or this script.
- Explorer embedded in this website. Run across Explore department.
How can I download data for a specific country or region?
The data is organised into tiles that tin be directly downloaded. Alternatively, the instance Colab notebook shows how to download data for a specific region, given the geometry of the area of involvement.
Is the imagery also available?
The underlying satellite imagery is non part of this dataset. Even so, the source imagery used for detections can be viewed in Google Earth Pro. Unlike time frames can be viewed using the Historical Imagery function.
Are there examples of resources for analysing the data?
In this Colab notebook , we demonstrate some analysis methods on the information for a specific country or region:
- Generating heatmaps of edifice density and size, and how to salvage this information in GeoTIFF format.
- Analysis of accessibility to wellness sites.
Will the data be updated or extended?
We hope to continue improving this dataset, by both refreshing it using new source imagery, and by refining the detection model to improve accuracy. Based on community feedback, we may extend the dataset to new areas or add together additional features: please let u.s.a. know, using the contact details below, whatsoever queries or requests.
What is the commendation for this dataset?
If this dataset is useful, delight consider citing our technical report:
W. Sirko, Due south. Kashubin, One thousand. Ritter, A. Annkah, Y.S.Eastward. Bouchareb, Y. Dauphin, D. Keysers, M. Neumann, M. Cisse, J.A. Quinn. Continental-scale building detection from high resolution satellite imagery. arXiv:2107.12283, 2021.
Who can I contact about bug with the dataset?
Please contact open-buildings-dataset@google.com with whatsoever feedback.
DOWNLOAD HERE
Posted by: rojasyourall.blogspot.com
0 Komentar
Post a Comment