针对视频监控系统采集的车牌图像分辨率低和模糊不清的问题，结合视频监控系统相对固定和车牌字符集有限的特点，提出了基于神经网络和先验知识的低分辨率车牌复原方法。该方法利用误差反向传播神经网络对监控系统采集的不同分辨率的车牌样本图像进行学习和训练，充分融合了低分辨率序列图像中的冗余信息和车牌字符先验知识。实验和实际应用验证了本文算法的有效性。“,”In this paper, a novel prior knowledge based license plate image restoration method is proposed to improve the quality of the low resolution degraded images captured by the video surveillance system. An error back propagation algorithm is used to learn and train different spatial resolution sample license plate images in order to combine the redundancy information of the captured images sequences and prior knowledge of the license plate characters. Experiments with synthetic and real image sequences are presented to demonstrate the effectiveness and robustness of the proposed method.