Bio-inspired looming object detection using event-based cameras
Sánchez Attolini, Guillermo
University of Lethbridge. Faculty of Arts and Science
Lethbridge, Alta. : Universtiy of Lethbridge, Department of Mathematics and Computer Science
We present a novel looming object detection method for event-based cameras. Event-based cameras capture a scene without generating redundant information thus reducing the needed transmission power, bandwidth, and computations to process the redundant information generated by conventional frame-based cameras. Regular computer vision algorithms cannot be directly applied to them. Conversely, existing event-based algorithms that detect looming objects have some limitations. In this thesis, an existing bio-inspired looming object detection algorithm for frame-based cameras was adapted for an event-based camera. The adapted algorithm was then used as part of a novel looming object detection algorithm that is fast and capable of detecting multiple looming objects in a scene. We tested our approach on the Davis Dynamic Vision Sensor event-based camera.
Computer vision , Computer algorithms , Optics , Digital images , Deconvolution , Dissertations, Academic