ABSTRACT
As seen in Chapter 3, the main sensory cue for flight control in insects is visual motion, also called optic flow (OF). In this Chapter, the formal description of OF enables us to gain an insight of global motion fields generated by particular movements of the observer and the 3D structure of the scene. It is then possible to analyse them and develop the control strategies presented in the following Chapter. In practice, lightweight robots cannot afford high-resolution, omnidirectional cameras and computationally intensive algorithms. OF must thus be estimated with limited resources in terms of processing power and vision sensors. In the second part of this Chapter, an algorithm for OF detection that meets the constraints imposed by the 8-bit microcontroller equipping our robots is described. Combining this algorithm with a 1D camera results in what we call an optic-flow detector (OFD). Such an OFD is capable of measuring OF in real-time along one direction in a selectable part of the field of view. Several of these OFDs, spanning one or more cameras are implemented on the robots to serve as image preprocessors for navigation control.
