Design and management of license plate recognition system based on BP neural network principle

The BP neural network is used for character recognition. Gradient descent is used to train the network in order to reduce the oscillatory tendency of the neural network learning process. BP neural networks can converge much faster. We construct a neural network structure with 88 input layer nodes, 15 hidden layer nodes, and 5 output layer nodes. In this way, a license plate recognition algorithm based on BP neural networks is implemented. The trial had a 90 percent success rate.