The mantis head camera (why the praying mantis is so good at catching its prey)

Abstract

Inspired by the abilities of the praying mantis to judge the distance to its prey before striking by use of motion-based visually mediated odometry, we create a miniature model for depth estimation that is similar to the head movements of the praying mantis. We develop mathematical models of the praying mantis behavior and describe our implementations and experimental environment. We investigate the structure-from-motion problem when images are taken from a camera whose focal point is translating according to the biological model. This motion is reminiscent of a praying mantis peering its head left and right, apparently to obtain depth perception, hence "mantis head camera". We present the performance of the mantis head camera model and provide experimental results and error analysis of the algorithm. The precision of our mathematical model and its implementation is consistent with experimental facts obtained from various biological experiments.