The Open Vision Computer (OVC) was designed to support high speed, vision guided autonomous drone flight. In particular, our aim was to develop a system that would be suitable for relatively small-scale flying platforms where size, weight, power consumption and computational performance were all important considerations. Our submitted manuscript describes the primary features of our OVC system and explains how they are used to support fully autonomous indoor and outdoor exploration and navigation operations on our Falcon 250 quadrotor platform. This manuscript can be found on arXiv.
Since the inception of the project, there has been a shift to a more generalized compute module. With the introduction of OVC4, cameras are now modular to allow for arbitrary selection and combination. This enables rapid prototyping for applications such as robot heads and autonomous vehicles without the need to re-fabricate the compute board with every sensor adjustment.
We have made many of our software repositories publicly available, and have been successfully used on our OVC platforms. We currently have many vision related tools such as stereo camera rectification, GPU accelerated depth estimation, deep learning based object segmentation etc.
All resources required to manufacture and assemble these devices is publicly available on our GitHub repository.