Method and system for generating physical true random number based on convolutional neural network

The invention provides a method and a system for generating a physical true random number based on a convolutional neural network. The method comprises the following steps: S1, setting a target identification object; s2, turning over the target identification object to change the state of the target identification object; s3, obtaining the state of the target identification object through a video or a video; s4, recognizing the current state of the target recognition object through the convolutional neural network; and S5, generating a group of random numbers according to the feature information of the current state of the target identification object. The method has the advantages of being simple, rapid, high in anti-interference performance, completely unpredictable, large in generation and the like, and tens of thousands and even hundreds of thousands of true random numbers (depending on the neural recognition speed of the convolutional network, the size of a container and the conversion mode of a data mat