基于BP神经网络和安卓的车牌识别系统设计

Design of Vehicle License Plate Recognition System Based on BP Neural Network and Android

提出了一种基于B P神经网络和安卓的车牌识别方法。以BP神经网络的模式识别技术作为主要识别方法,选取尺度不变特征转换(SIFT)算法探测牌照中局部特征字符特征作为神经网络的输入,进行神经网络训练,实现了对字符的识别,然后将车牌识别方法移植到安卓平台,实验结果表明该识别方法可行,满足了车牌识别基本需求。

To enhance the vehicle license plate recognition speed and convenience, a method based on Neural Network and Android is proposed. It selects scale-invariant feature transform to detect features of the licence based on BP neural network, trains and realizates the license plate characters. The experiment indicates that the mehod can reduce the error effectively with rapid response.