Context Aware Object Geotagging

TL;DR

This paper presents a novel method for improving street object geotagging by combining image metadata enhancement with contextual geographic information, validated specifically for traffic lights.

Contribution

It introduces a pipeline that enhances image metadata with Structure from Motion and refines geolocation using OpenStreetMap data, advancing street object localization techniques.

Findings

Outperforms existing methods in traffic light geotagging accuracy

Effective integration of image metadata and geographic context

Validated on real-world street view datasets

Abstract

Localization of street objects from images has gained a lot of attention in recent years. We propose an approach to improve asset geolocation from street view imagery by enhancing the quality of the metadata associated with the images using Structure from Motion. The predicted object geolocation is further refined by imposing contextual geographic information extracted from OpenStreetMap. Our pipeline is validated experimentally against the state of the art approaches for geotagging traffic lights.