Geo-localization by Image Retrieval
- Project
- 15026 PS-CRIMSON
- Type
- New standard
- Description
- High geo-localization accuracy due to of advanced feature extraction by neural network
- Robust to varying illumination conditions
- Contact
- Egor Bondarev
- e.bondarev@tue.nl
- Technical features
Input(s):
- Database of geotagged city/rural images
- Query image for geolocalization
Main feature(s):
- The component is able to detect the latitude and longitude of a taken photo with a mean accuracy of 7 meters
Output(s):
- Latitude and longitude of the place from which the image was taken
- Integration constraints
- Requires geo-tagged image database of the region
- Hardware: no
- Software: no
- Targeted customer(s)
CycloMedia, Police, emergency services, Research community (via open source).
- Conditions for reuse
Licensing
- Confidentiality
- Public
- Publication date
- 13-01-2020
- Involved partners
- Eindhoven University of Technology (NLD)