A review of Mapillary Traffic Sign Data Quality and OSM Coverage
summary
Mapillary extracts stop signs from GoPro Max street imagery using semantic segmentation and Structure from Motion, then compares results to survey ground truth and OSM in San Francisco and Redmond. Results show 95%+ recall for stop signs with positional errors about 2.5 m avg (down to 2.16 m with denser processing), but relatively low precision driven by temporary signs; OSM misses 70–90% of stop signs in the study area and over 200 intersections were added.
AI-written summary · about
related talks
18mRecognition and Localization of Parking Signs using Deep Learning
22mIsland Mapping: Measuring Urban Footprints with Mapillary & Open Source Tools
21mFIXME: Documenting the Uneven Distribution of OSM Data Quality Issues
26mSelf Driving cars with Opencv
16mAutomated Calibration of Traffic Demand and Traffic Lights in SUMO Using Real-World Observations
13mUtilizing Vehicle Trajectory Data to Characterize Performance Measures of a Highway Corridors
07mUrban Map Inference by Pervasive Vehicular Sensing Systems with Complementary Mobility
15mThe turf wars on data - can be a matter of life and death