Event-Driven Particle Filter for Tracking Irregularly Moving Objects
Yuki Kawasaki, Masahiro Ohtani, Shinsuke Yasukawa · Proceedings of International Conference on Artificial Life and Robotics · 2023
Conventional object tracking techniques that use general-purpose cameras and particle filters find it difficult to track irregularly and rapidly moving objects.To track an irregularly moving object without losing sight, quickly measuring the position of the object is necessary.In this study, we used a fast-response event-based camera, which is a bioinspired camera that produces a spiking output.We propose an event-driven particle filter that performs processing in response to the input from an event-based camera.Our proposed method was evaluated by presenting an eventbased camera with a rectangular motion pattern that combines periodic and constant-velocity motions at various speeds.The experimental results demonstrated that our proposed method could track objects in a test video.