离散型Hopfield神经网络算法在车牌识别中的应用

The Application of Discrete Hopfield Neural Network Algorithm in License Plate Character Recognition

在车牌识别的过程中存在一定的噪声干扰情况,传统的模板匹配法识别率大大降低。利用离散型Hopfield神经网络具有联想记忆功能,能够为受噪声干扰的图像去噪,与模板匹配法相结合可提高模板匹配法的抗干扰性和识别的准确率。通过实验表明,该方法具有较强的有效性和可行性。

In the process of license plate recognition,there is some noise interference,and the recognition rate of traditional template matching method is greatly reduced.The discrete Hopfield neural network has associative memory function,which can denoise the image disturbed by noise.Combining with template matching method,the anti-interference and recognition accuracy of template matching method can be improved.Experiments show that the method is effective and feasible.