Location Aware Image Database (LAID) provides dense high quality images annotated with information such as GPS location and time.
LAID is a testbed for researchers to study various problems associated with image processing, data mining, modelling, networking, simulation, and navigation.
Current LAID research includes 3D city modelling, visual based navigation, object location determination from images and scene prediction
In order to provide high-quality image sequences for navigation, we mount two
3.2 (2048 × 1536) million pixels (MP) cameras on the roof of a sedan (at height of 1.4 meters) to acquire
images. Digital cameras are used due to their higher resolution as most digital camcorders provide only
340K pixels. In the continuous-acquiring mode, each camera can obtain one image approximately every 1.5
seconds. Each image is associated with the GPS location of the camera, the time the image is taken, the
date, and the weather condition.
Images in the database are currently of the city of West Lafayette, IN. We plan to extend to 8 cameras for a 360 degrees panoramic view,
2 camcorders for stereo vision and add more cities to the database by 2006.