Webcam geo-localization using aggregate light levels

Abstract

We consider the problem of geo-locating static cameras from long-term time-lapse imagery. This problem has received significant attention recently, with most methods making strong assumptions on the geometric structure of the scene. We explore a simple, robust cue that relates overall image intensity to the zenith angle of the sun (which need not be visible). We characterize the accuracy of geolocation based on this cue as a function of different models of the zenith-intensity relationship and the amount of imagery available. We evaluate our algorithm on a dataset of more than 60 million images captured from outdoor webcams located around the globe. We find that using our algorithm with images sampled every 30 minutes, yields localization errors of less than 100km for the majority of cameras.

Incorporating Cloud Distribution in Sky Representation

“…In this paper, we use the AMOS data set [8] to evaluate our proposed model. Among 633 cameras with the ground truth location data in the AMOS data set, we choose 198 cameras which take pictures where the portion of the sky is about one-third of the entire image or more.…”

“…Figure 7 shows some typical failure cases from the AMOS data set [8]. The intensities in the areas marked by the red rectangles in column (d) should be darker than the solved values.…”

“…The reconstructed images using the uniform sky index model and our proposed model (one example per row). Column (a) is the input sky images taken from the AMOS data set [8]. Column (b) and (c) are the reconstructed sky images (brighter pixels indicate higher sky intensity) with the uniform sky index model and our model respectively.…”

“…The failure cases from the AMOS data set [8] using our proposed model. The ordering of (a)∼(d) is the same as that in Figure 6.…”

“…The Figure 8. Sample outdoor images of the target data set derived from the AMOS data set [8]. Figure 9.…”