Event-based vision for Robotics

Abstract

Event cameras, such as the Dynamic Vision Sensor (DVS), are bio-inspired vision sensors that output pixel-level brightness changes instead of standard intensity frames. They offer significant advantages over standard cameras, namely a very high dynamic range, no motion blur, and a latency in the order of microseconds. However, because the output is composed of a sequence of asynchronous events rather than actual intensity images, traditional vision algorithms cannot be applied, so that new algorithms that exploit the high temporal resolution and the asynchronous nature of the sensor are required. In this research seminar at the Escuela Politécnica Superior of the Universidad Autonoma de Madrid I provide a broad introduction to the topic. Then, I discuss two recent works on the use of event-based camera for motion estimation: event-based, 6-DOF camera tracking from photometric depth maps, and a unifying contrast maximization framework that allows us to solve multiple computer vision problems. Finally, I present an outlook of the future, learning-based approaches and the challenges that remain ahead of us.