Scholarlink

<p>As a traditional handicraft with a long history and exquisite skills, the manufacturing process of twisted tire porcelain involves complex steps and high energy consumption. Because twisted tire porcelain has many unique patterns, it is necessary to classify porcelain according to the patterns in the production process. However, the current manual sorting method is inefficient, time-consuming, labour-intensive and costly, so exploring an intelligent sorting scheme is urgent. Artificial intelligence technology is changing with each passing day, and image classification algorithms in deep learning are emerging one after another with high efficiency and accuracy. It has become possible to realize the intelligent classification algorithm of twisted tire porcelain patterns. Taking twisted tire porcelain in Dangyangyu, Jiaozuo City as the research object, combining with the needs of twisted tire porcelain manufacturers in Dangyangyu, twisted tire porcelain is divided into four kinds of patterns: natural pattern, geometric pattern, ancient pattern and bionic pattern, and a large number of twisted tire porcelain pattern pictures are collected, and the classification data set of twisted tire porcelain pattern is constructed by manual labelling and labelling. At the same time, based on the convolutional neural network model Alexnet, the cross-entropy loss function is selected as the training loss, and the twisted tire porcelain pattern classification algorithm under the deep learning framework is designed. After training and optimization of the hardware equipment, the optimal model is obtained. After the algorithm is trained on GPU many times, the obtained optimal model is tested many times in the above data set, with an accuracy rate of 90.2% and a speed of 66 frames per second. Through qualitative analysis, the algorithm performs well in actual classification. The experimental results show that the proposed pattern classification algorithm can realize the automatic classification of different patterns of twisted tire porcelain with high accuracy, which saves workforce, improves efficiency and meets the needs of industrial intelligence.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Construction and scheduling optimization of renewable energy consumption forecasting system for twisted tire porcelain manufacturing industry based on deep learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年5月25日

查看全文
  • Fangrong Liao,
  • Runlin Ran,
  • Jiayi Zhang

摘要

As a traditional handicraft with a long history and exquisite skills, the manufacturing process of twisted tire porcelain involves complex steps and high energy consumption. Because twisted tire porcelain has many unique patterns, it is necessary to classify porcelain according to the patterns in the production process. However, the current manual sorting method is inefficient, time-consuming, labour-intensive and costly, so exploring an intelligent sorting scheme is urgent. Artificial intelligence technology is changing with each passing day, and image classification algorithms in deep learning are emerging one after another with high efficiency and accuracy. It has become possible to realize the intelligent classification algorithm of twisted tire porcelain patterns. Taking twisted tire porcelain in Dangyangyu, Jiaozuo City as the research object, combining with the needs of twisted tire porcelain manufacturers in Dangyangyu, twisted tire porcelain is divided into four kinds of patterns: natural pattern, geometric pattern, ancient pattern and bionic pattern, and a large number of twisted tire porcelain pattern pictures are collected, and the classification data set of twisted tire porcelain pattern is constructed by manual labelling and labelling. At the same time, based on the convolutional neural network model Alexnet, the cross-entropy loss function is selected as the training loss, and the twisted tire porcelain pattern classification algorithm under the deep learning framework is designed. After training and optimization of the hardware equipment, the optimal model is obtained. After the algorithm is trained on GPU many times, the obtained optimal model is tested many times in the above data set, with an accuracy rate of 90.2% and a speed of 66 frames per second. Through qualitative analysis, the algorithm performs well in actual classification. The experimental results show that the proposed pattern classification algorithm can realize the automatic classification of different patterns of twisted tire porcelain with high accuracy, which saves workforce, improves efficiency and meets the needs of industrial intelligence.