| 项大兵,张春生,王红彦,侯丹丹.基于深度学习的轮胎花纹噪声预报方法研究[J].轮胎工业,2019,39(8):0503-0506 |
二维码（扫一下试试看！） |
| 基于深度学习的轮胎花纹噪声预报方法研究 |
| Tire Pattern Noise Prediction based on Deep Learning |
| 投稿时间：2018-10-08 修订日期：2018-10-08 |
| DOI：10.12135/j.issn.1006-8171.2019.08.0503 |
| 中文关键词: 轮胎噪声预报 花纹识别 深度学习 神经网络 |
| 英文关键词: tire noise prediction pattern image recognition deep learning neural network |
| 基金项目: |
| 中图分类号: |
|
| 摘要点击次数: 1134 |
| 全文下载次数: 392 |
| 中文摘要: |
| 在进行大量不同花纹轮胎通过噪声试验基础上，利用图像识别和深度学习方法建立了轮胎花纹几何结构与轮胎通过噪声试验结果之间的数学模型，并采用该模型对新设计的轮胎花纹进行通过噪声预测。结果表明：以1 dB声压级为预测精度时，模型预测准确度达到83%；利用花纹图像识别和深度学习理论所建立的轮胎噪声预报数学模型方便可靠，而且随着试验数据的不断积累和算法的改进，基于该方法的预报模型精度会进一步提高。 |
| 英文摘要: |
| Based on a lot of different pattern tire noise testing，a mathematical model between the tire pattern geometry and the tire noise level was developed based on the image recognition and deep learning technology.What’s more，the noise level of a new tire pattern was predicted by this model.The results showed that the model prediction accuracy reaches at 83% when the resolution was 1dB.The mathematical model of tire noise prediction developed by image recognition and deep learning theory was reliable and convenient.With more testing data accumulated and the improvement of algorithm the prediction accuracy would be constantly improved. |
|
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
