A simple content-based strategy for estimating the geographical location of a webcam
PCM'10 Proceedings of the 11th Pacific Rim conference on Advances in multimedia information processing: Part I
UIC'10 Proceedings of the 7th international conference on Ubiquitous intelligence and computing
| Hi-index | 0.01 |
Geo-spatial image tagging is increasingly used for improving image storage, retrieval and search. Currently, geo-spatial tagging of images relies on GPS-enabled cameras, where the geographical location of the photographer is embedded in the image files. However, problems with the GPS approach is that it takes time for most GPS equipment to lock onto overhead satellites, GPS-enabled cameras are expensive and large image collections without geo-spatial information are common. This study explores a simple strategy for determining the approximate geographical location of the photographer from a series of images without considering the image contents. Two models are explored. The basic model is based on an image time-stamp profile alone, and the enhanced model improves the accuracy by incorporating the optical camera settings for the images and compares these with the elevation path of the sun. The strategy is applied to three image collections whereupon the approximate locations, i.e., the latitude and longitude, of the photographer were determined.