Testing Bio-inspired Motion Detection Models


David C O'Carroll, Mandyam V Srinivasan;
Post Doctoral Research Fellow: Russell SA Brinkworth.

Introduction: This project will test bio-inspired models of the insect visual system under realistic conditions. It will take a recently developed model of rotational motion processing by the fly and apply it to conditions involving both translation and rotation in order to produce a full motion detection model based solely on passive optical inputs.
Significance: As a consequence of this project a number of high resolution, high dynamic range (HDR), full panorama movies will be recorded. This, together with the motion trajectory of the recording platform, will be a useful base set for researchers looking to develop and test other models of motion detection.
Applications: The successful demonstration of these algorithms will have applications in the development of miniature autonomous systems in defence and civilian roles, including robotics, miniature unmanned aerial vehicles, intruder detection and collision avoidance sensors.
top of page