BP神经网络在路标识别上的应用研究

Research on Application of BP Neural Network in Road Sign Recognition

借助matlab中BP神经网络工具对生活中的路面导流标志进行检测识别,对输入到BP神经网络中的原始图像依次进行了灰度化和二值化图像处理后,BP神经网络的识别结果表现出很大的不稳定性,在此基础上,引入了Hough变换对输入图像进行了旋转校正处理,试验结果表明Hough变换的引入对于BP神经网络识别准确率的提升较大,同时,识别结果与其它几个路面导流标志的识别区分度较高。

This paper uses the BP neural network tool in matlab to detect and recognize the road diversion signs in life.After the original image input into the BP neural network is grayed and binarized image processing,the recognition result of the BP neural network showed great instability.On this basis,the Hough transform is introduced to perform rotation correcting processing on the input image.The test results show that the introduction of Hough transform greatly improves the recognition accuracy of the BP neural network.At the same time,the recognition result is highly distinguishable from other road diversion signs.