基于改进BP神经网络的夜间车牌识别算法研究

Research on Night Vehicle License Plate Recognition Algorithm Based on Improved BP Neural Network

为了提高夜间条件下车牌识别准确率,提出了一种基于改进BP神经网络的车牌识别算法。为了改善夜间环境下车牌图像的质量和清晰度,在图像预处理过程中采用了图像平滑处理增强技术;利用图像边缘检测技术实现了对图像正确定位,然后通过统计车牌图像白色像素个数的方法对字符分割;在此基础上,使用基于附加动量法和自适应学习速率改进的BP神经网络方法精确识别车牌。实验结果表明,该方法对夜间车牌的分割和识别是有效的。

In order to improve the accuracy of license plate recognition at night, a license plate recognition algorithm based on improved BP neural network is proposed. The image smoothing enhancement technology was used for improving the quality and clarity of the license plate images at night. The license plate was located by image edge detection technology and each single character was segmented by counting the number of white pixels in the license plate images. The additional momentum method and adaptive learning rate for improving the good classification performance of BP neural network were also discussed. Experimental results show that this method can effectively segment and identify license plates at night.