License Plate Image Restoration Based on Fuzzy Neural Network

Automatic license plate recognition system is an intelligent surveillance system in traffic management and toll. Various image restoration methods have been proposed, but they are deficient in function approximation. Neural network has its unique advantages for its large-scale nonlinear dynamic characteristic, parallelism calculation, high robustness, strong capacity of self-adaptive, self-organization and self-learning. A novel license plate preprocessing technique based on fuzzy neural network has been proposed here, which is verified to be effective by the experiments.