基于卷积神经网络的轮胎花纹噪声值预测  

Prediction of Tire Pattern Noise Value Based on Convolutional Neural Network

作　　者：李志伟 苏宇 张舜 王青春[1] LI Zhiwei;SU Yu;ZHANG Shun;WANG Qingchun(Beijing Forestry University,Beijing 100091,China;Anhui Lupital Iot Co.,Ltd,Hefei 230031,China)

机构地区：[1]北京林业大学工学院,北京100091 [2]安徽路必达智能科技有限公司,安徽合肥230031

出　　处：《轮胎工业》2023年第12期756-761,共6页Tire Industry

摘　　要：利用图像处理和卷积神经网络(CNN)搭建轮胎花纹结构与轮胎花纹噪声值之间的数学模型,分别采用CNN模型和BP神经网络对轮胎花纹噪声值进行预测,并对比预测精度。结果表明:采用CNN模型,轮胎花纹噪声的预测值与实测值的平均绝对误差为0.591 dB,平均相对误差为0.81%;采用BP神经网络,轮胎花纹噪声的预测值与实测值的平均绝对误差为0.713 dB,平均相对误差为0.95%;相较于BP神经网络,CNN模型对轮胎花纹噪声值的预测精度更高。The image processing and convolutional neural networks(CNN)were used to establish a mathematical model between the tire pattern structure and the tire pattern noise values.The CNN model and BP neural network were used to predict the tire pattern noise values respectively,and the prediction accuracies of those two methods were compared.The results showed that,using the CNN model,the average absolute error between the predicted and measured tire pattern noise values was 0.591 dB,and the average relative error was 0.81%.Using the BP neural network,the average absolute error between the predicted value and the measured tire pattern noise value was 0.713 dB,and the average relative error was 0.95%.Compared with the BP neural networks,the CNN models had a higher prediction accuracy for tire pattern noise values.

关 键 词：轮胎花纹 图像处理 卷积神经网络 BP神经网络 噪声值预测 

分 类 号：TQ336.1[化学工程—橡胶工业]