基于深度学习的电池陶瓷复合材料成分检测

使用传统方法进行电池等电子材料检测时,往往由于检测时间较长,和受到温湿度等条件的限制,造成检测的精准度较低的问题。为提高检测的精准度,提出一种利用深度学习中的卷积神经网络算法进行电池电子陶瓷复合材料成分检测方法。首先,收集电子陶瓷复合材料无损区域的近红外图像并对图像进行预处理;其次,进行图像特征的提取,提取出不同材料成分的主要特征;再次,将图像特征输入卷积神经网络的特征层;最后应用卷积神经网络算法对图像特征进行训练和深度学习,最终分析出稳定的电子陶瓷复合材料的成分。并且通过实验结果的对比分析,验证该方法对于提高复合材料成分的检测精准度的有效性,并且可以缩短检测时间。

When using traditional methods to detect electronic materials such as batteries, the detection accuracy is often low because of the long testing time and the limitation of temperature and humidity. In order to improve the accuracy of detection, a convolution neural network algorithm in depth learning is proposed to detect the composition of battery electronic ceramic composites. First, the near infrared images of the lossless region of electronic ceramic composites are collected and preprocessed;second, the image features are extracted to extract the main features of different material components;third, the image features are input into the feature layer of the convolution neural network;finally, the convolution neural network algorithm is used to train and learn the image features. In the and the stable composition of electronic ceramic composites is analyzed. Through the comparative analysis of the experimental results, it is verified that the method is effective to improve the detection accuracy of composite components, and can shorten the detection time.