High-Speed Object Recognition Based on a Neuromorphic System

Neuromorphic systems are bio-inspired and have the potential to break through bottleneck of existing intelligent systems. This paper proposes a neuromorphic high-speed object recognition method based on DVS SpiNNaker implements system in which an OR logic aggregation algorithm is used acquire sufficient effective information asynchronous sparse computing mechanism SNNs exploited reduce computation. The experiment’s results show that detection rate designed more than 99% at rotating speed 900~2300 rpm; its response time within 2.5 ms; it requires 96.3% less computation traditional using same scaled ANN.