摘要
为了保证农副产品的质量和安全,基于多传感器的远程监控技术已经广泛应用于农产品冷链运输行业。传统的冷链运输环境监测与预测技术主要结合各环境指标的分析,并没有对异构和非均衡数据进行有效融合和拟合预测。基于预训练卷积自编码器、注意力机制与长短记忆网络(LSTM)提出了K-LSTM融合与预测算法模型,实验结果表明K-LSTM算法模型融合精度达到了96%,相较于文献研究算法指标结果提高了20%~70%,因此提出的K-LSTM能够对冷藏车厢内部温度和湿度提供准确预测,为冷链的智能管理提供了有效支持。
In order to ensure the quality and safety of agricultural and sideline products,remote monitoring technology based on multi-sensor has been widely used in agricultural cold chain transportation industry.The traditional cold chain transportation environmental monitoring and prediction technology mainly combines the analysis of various environmental indicators,and does not effectively integrate and fit the heterogeneous and unbalanced data.In this paper,a K-LSTM fusion and prediction algorithm model is proposed based on the pre-trained convolution encoder,attention mechanism and long and short memory network(LSTM).The experimental results show that the fusion accuracy of the K-LSTM algorithm reaches 96%,which is 20%~70% higher than the index results of the literatures.Therefore,the K-LSTM proposed in this paper can accurately predict the temperature and humidity inside the refrigerated carriage,which provides effective support for the intelligent management of the cold chain.
作者
陈林
丁士杰
CHEN Lin;DING Shi-jie(Anhui Vocational College of Finance and Trade,Hefei 230601,Anhui;Suzhou University,Suzhou 234000,Anhui)
出处
《陇东学院学报》
2024年第2期7-12,共6页
Journal of Longdong University
基金
安徽省高校自然科学重点项目(KJ2018A0908)。
关键词
冷链运输
数据融合
LSTM
预测
cold chain transportation
data fusion
LSTM
forecast
作者简介
陈林(1980-),男,安徽石台人,讲师,主要从事物联网技术研究。
