基于深度学习算法的织物风格评价模型与纤维力学性能分析  

Test and model study of protective plate under spherical explosion load

作　　者：申粉粉 SHEN Fenfen(School of mechanical and automotive engineering,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学机械与汽车工程学院,上海201620

出　　处：《智能计算机与应用》2021年第8期167-172,共6页Intelligent Computer and Applications

摘　　要：为了将图像处理技术与卷积神经网络(CNN)在深度学习中相结合,提高异形纤维的识别准确率,提出一种基于力学性能分析和反向传播神经网络(BPNN)的纺织品质量预测模型。此外,利用传统的BP神经网络来模拟原纤维质量与纺织品质量之间的相关性,从而构建了纺织品质量预测模型。希望通过对针织复合材料力学性能的分析,为针织复合材料的应用提供依据。To combine the image processing technology with the convolutional neural network(CNN)in deep learning,improve the recognition accuracy of shaped fibers,and propose a model for textile quality prediction based on mechanical property analysis and back propagation neural network(BPNN),a deep learning-based CNN is trained to recognize the shaped fibers based on fiber image processing.In addition,the traditional BPNN is utilized to simulate the correlation between the quality of raw fibers and textile products,thereby constructing a model for textile quality prediction.It is hoped to provide a basis for the application of knitted composite materials by analyzing the mechanical properties of the material.

关 键 词：深度学习 神经网络 形状的纤维识别 力学性能 纺织品质量预测 

分 类 号：U471.1[机械工程—车辆工程]