摘要
农机设备备件管理是农业领域的一个关键问题,准确的备件分类可以确定更优的库存管理策略。改进深度残差网络的备件分类方法,可以通过建立多维度的分类准则和对备件信息数据预处理,得到具有时间维度和备件属性图像的备件标识方法。提取特征时,为保证相似属性间的特征联系,在网络模型中加入挤压和激励网络(squeeze-and-excitation networks,SENet),得到改进的残差网络模型。为验证模型的分类效果,选用某拖拉机集团的大型设备的备件作为案例分析。结果表明,该方法对大型生产机器的备件具有很好的分类效果。
Agricultural machinery equipment component management is a critical issue in the agriculture field,and a better management strategy can be determined by accurate spare parts classification.This paper presents a method based on an improved depth residual network proposed to solve the problem of spare parts classification.Establishing multi-criteria classification and spare part information data were preprocessed to obtain identification methods with time dimensions and attribute characteristics.The attention mechanism based on the Squeeze and Excitation Network(SENet)structure was introduced into the residual neural network residual block to ensure the characteristic relationship between similar attributes.The effectiveness of the classification method was verified by a practical case in the large equipment of a tractor group.The results show that spare parts of large equipment can be accurately classified by this method.
作者
刘梦飞
周伟
李西兴
LIU Mengfei;ZHOU Wei;LI Xixing(School of Mechanical Engineering,Hubei Univ.of Tech.,Wuhan 430068,China)
出处
《湖北工业大学学报》
2023年第5期34-39,共6页
Journal of Hubei University of Technology
基金
国家自然科学基金(51805152)。
关键词
农机设备
备件分类
多准则
深度残差网络
挤压与激励结构
agricultural machinery equipment
spare parts classification
multi-criteria
deep residual network
extrusion and excitation network
