Real-time Low-Cost Omni-directional Stereo Vision via Bi-Polar Spherical Cameras

Abstract

With the rise of consumer-grade spherical cameras, offering full omni-directional 360∘ image capture, the potential for low-cost omni-directional stereo vision is ever present. Whilst this potentially offers novel low-cost omni-directional depth sensing without the need for active range sensing, it presents a number of challenges associated with the realization of contemporary dense stereo matching across this spherical imaging format. In this work, we outline an approach for omni-directional dense stereo vision, based on imagery from two consumer-grade spherical cameras mounted in a bi-polar configuration, offering 360∘ depth recovery at 5.5 fps. Specifically we outline the required aspects of inter-camera calibration and equirectangular image correction before detailing the required angular disparity correction that is unique to this stereo sensing formulation. Furthermore we illustrate the disparity and synchronization error achievable with the use of such consumer-grade spherical camera units, in addition to the quality of disparity (depth) available, within the context of on-road sensing for future vehicle autonomy.