Euclidean Auto Calibration of Camera Networks: Baseline Constraint Removes Scale Ambiguity

TL;DR

This paper introduces a method for auto calibrating camera networks to recover scale and shape accurately, using a stereo pair with known baseline, which is validated experimentally and compares favorably with existing methods.

Contribution

It proposes a novel calibration approach that includes a known-baseline stereo pair, enabling Euclidean auto calibration and scale recovery in camera networks.

Findings

Successfully recovers scale in 3D reconstructions

Experimental validation with a four-camera network

Outperforms Zhang and Pollefeys methods in shape accuracy

Abstract

Metric auto calibration of a camera network from multiple views has been reported by several authors. Resulting 3D reconstruction recovers shape faithfully, but not scale. However, preservation of scale becomes critical in applications, such as multi-party telepresence, where multiple 3D scenes need to be fused into a single coordinate system. In this context, we propose a camera network configuration that includes a stereo pair with known baseline separation, and analytically demonstrate Euclidean auto calibration of such network under mild conditions. Further, we experimentally validate our theory using a four-camera network. Importantly, our method not only recovers scale, but also compares favorably with the well known Zhang and Pollefeys methods in terms of shape recovery.