The relevance of honeybee vision to the design of seeing systems

Abstract

The first chapter in the application of insect vision to seeing systems, piloting and control, has almost been completed. For many years it was thought that the vision of the fly, as illustrated in the optomotor response of pinned down animals. represented insect vision. Fixed insects respond to the passing of edges by the spatiotemporal correlation between adjacent visual axes and therefore to the temporal frequency of passing edges, irrespective of pattern or the angular velocity across the eye. But a visual system that behaves like this is unsuitable for guiding in free flight, preventing crashes or steering towards a target, never mind pattern perception. Around 15 years ago, we realized that freely flying bees can measure the range and size of a nearby target [1]. They can do this because they can measure the angular velocity of contrasts that move across the eye as a result of their own motion, and they can measure their own velocity relative to their surroundings as they fly. Bees can also integrate the angular velocity over time to give the distance travelled [2]. This is active vision. As a result, we have been able to copy these principles into freely flying vehicles that can manoeuvre themselves in the air at high speed.