Scholarlink

<p>For dairy supply chains to be safer, more transparent, and better managed, food traceability is essential. Nevertheless, current traceability solutions include drawbacks such as inadequate quality assessment, ineffective data sharing, and a lack of security. In order to address these issues, this article proposes a blockchain-assisted food quality assessment and traceability framework for milk and dairy products using hybrid deep learning models. In order to guarantee data integrity and privacy, the proposed system first tags dairy products with RFID. The acquired data is then sent via Internet of Things devices and safely stored in a blockchain network. The four stages of the traceability process are manufacturer, distributor, transporter, and consumer. The Hybrid randomized sine cosine chaotic assisted whale optimization (Hy-RasinWop) algorithm is used for feature selection. A pruning assisted multi-spatiotemporal attention based LSTM is used for classification, and pre-processing techniques like min–max normalization and data cleaning are used for quality evaluation. A milk grading dataset is used to assess the performance of the proposed model and compare it with other methods. The findings show that the proposed framework performs better in classification with an accuracy of 98.54%, precision of 97.84%, recall of 97.02%, and F1-score of 97.37%. Furthermore, the model's low error rates (MSE of 0.1724 and RMSE of 0.1954) guarantee accurate forecasts. From a system standpoint, the proposed method highlights its computational efficiency by lowering response time to 20.414&#xa0;s and energy consumption to 21.82&#xa0;J. The results' robustness is further supported by statistical analysis, which shows consistent performance over several runs and a considerable improvement (p = 0.0032 &lt; 0.05) over baseline models.</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

Blockchain assisted food traceability and quality evaluation for IoT driven data with hybrid model

  • 期刊论文

  • 发表日期：

    2026年7月13日

查看全文
  • Ch. IndiraPriyadarsini,
  • K. Raja Sravan Kumar,
  • Arudra Sandeep Kumar,
  • Vunnava Dinesh Babu,
  • Chinta Venata Murali Krishna,
  • Popuri Srinivasarao

摘要

For dairy supply chains to be safer, more transparent, and better managed, food traceability is essential. Nevertheless, current traceability solutions include drawbacks such as inadequate quality assessment, ineffective data sharing, and a lack of security. In order to address these issues, this article proposes a blockchain-assisted food quality assessment and traceability framework for milk and dairy products using hybrid deep learning models. In order to guarantee data integrity and privacy, the proposed system first tags dairy products with RFID. The acquired data is then sent via Internet of Things devices and safely stored in a blockchain network. The four stages of the traceability process are manufacturer, distributor, transporter, and consumer. The Hybrid randomized sine cosine chaotic assisted whale optimization (Hy-RasinWop) algorithm is used for feature selection. A pruning assisted multi-spatiotemporal attention based LSTM is used for classification, and pre-processing techniques like min–max normalization and data cleaning are used for quality evaluation. A milk grading dataset is used to assess the performance of the proposed model and compare it with other methods. The findings show that the proposed framework performs better in classification with an accuracy of 98.54%, precision of 97.84%, recall of 97.02%, and F1-score of 97.37%. Furthermore, the model's low error rates (MSE of 0.1724 and RMSE of 0.1954) guarantee accurate forecasts. From a system standpoint, the proposed method highlights its computational efficiency by lowering response time to 20.414 s and energy consumption to 21.82 J. The results' robustness is further supported by statistical analysis, which shows consistent performance over several runs and a considerable improvement (p = 0.0032 < 0.05) over baseline models.