Neuromorphic visual sensor target detection method based on spiking neural network

The invention discloses a neuromorphic visual sensor target detection method based on a pulse neural network. The method comprises the following steps: S1, training a clipped artificial neural network; s2, converting the artificial neural network into a pulse neural network; s3, the neuromorphic visual sensor collects picture information of the dynamic object, and inputs a space-time pulse signal into the pulse neural network after image reconstruction is carried out on the picture information; s4, the pulse neural network calculates, detects and decodes the space-time pulse signal and then outputs an image, marks the image according to a detection result, marks the position and the type of the mark in each target by using a rectangular frame, and judges a detection and identification result; and S5, verification is carried out on identification and detection results. The method is friendly to hardware and can process detection images of high-speed dynamic objects. 本发明公开了一种基于脉冲神经网络的神经形态视觉传感器目标检测方法，包括以下步骤：S1，训