Double-Row License Plate Segmentation with Convolutional Neural Networks

With the fast development of intelligent traffic,the license plate recognition technology progressively improves.Most of existing license plate recognition techniques can well recognize character information for singlerow license plates but the recognition accuracies for double-row license plates are not ideal and even less algorithms support Chinese characters.This paper introduces a doublerow license plate segmentation method with CNN,enabling efficient double-row license plate recognition for originally single-row recognition algorithms.First,this method trains a multi-label classification model with the image features extracted using CNN.Then,we use the model to automatically segment a double-row license plate into two single-row license plates.In addition,we have constructed a training and validation dataset containing more than 200 000 Chinese license plate images.The experimental results show that the proposed method has a higher accuracy in automatic segmentation of double-row license plate,thus effectively improving the accuracy of double-row license plate recognition.