Scholarlink

<p>Aflatoxin contamination in cereals poses a persistent public health, food safety, and economic challenge in the East and sub-Saharan Africa, contributing to hepatocellular carcinoma, immunosuppression, and post-harvest losses. Conventional detection methods, including enzyme-linked immunosorbent assays (ELISA) and high-performance liquid chromatography (HPLC), provide analytical accuracy but are limited to centralized laboratories, are costly, and offer delayed results, making them unsuitable for real-time monitoring in low-resource storage environments. This study systematically reviews existing literature to identify technologies and components suitable for the design and implementation of an artificial intelligence (AI)-enabled aflatoxin detection and prediction platform. Using PRISMA guidelines, 65 peer-reviewed studies published between 2010 and 2025 were analyzed across three domains: biosensor technologies for aflatoxin detection, environmental monitoring systems relevant to fungal growth, and Artificial Intelligence (AI) models for detection, data fusion, and short-term risk prediction. This review highlights that electrochemical biosensors provide a favourable balance of sensitivity, cost, and field deployability considering the different environmental factors, while hybrid AI-mechanistic models demonstrate promising predictive capability and improved generalization potential under limited-data conditions, although empirical validation in low-resource settings remains limited, although this advantage is supported by a limited number of studies that explicitly evaluated low-data scenarios. Critical integration challenges, including sensor calibration, data fusion, and contextual adaptation, are identified. The findings establish an evidence-based foundation for the subsequent design, simulation, and prototyping of an integrated hardware–software system, enabling a transition from reactive testing to continuous and predictive aflatoxin risk surveillance of grain storage systems.</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

Artificial intelligence based aflatoxin risk detection and prediction in stored foods a systematic literature review

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月9日

查看全文
  • Bonny Matsiko,
  • Johnes Obungoloch,
  • William Wasswa

摘要

Aflatoxin contamination in cereals poses a persistent public health, food safety, and economic challenge in the East and sub-Saharan Africa, contributing to hepatocellular carcinoma, immunosuppression, and post-harvest losses. Conventional detection methods, including enzyme-linked immunosorbent assays (ELISA) and high-performance liquid chromatography (HPLC), provide analytical accuracy but are limited to centralized laboratories, are costly, and offer delayed results, making them unsuitable for real-time monitoring in low-resource storage environments. This study systematically reviews existing literature to identify technologies and components suitable for the design and implementation of an artificial intelligence (AI)-enabled aflatoxin detection and prediction platform. Using PRISMA guidelines, 65 peer-reviewed studies published between 2010 and 2025 were analyzed across three domains: biosensor technologies for aflatoxin detection, environmental monitoring systems relevant to fungal growth, and Artificial Intelligence (AI) models for detection, data fusion, and short-term risk prediction. This review highlights that electrochemical biosensors provide a favourable balance of sensitivity, cost, and field deployability considering the different environmental factors, while hybrid AI-mechanistic models demonstrate promising predictive capability and improved generalization potential under limited-data conditions, although empirical validation in low-resource settings remains limited, although this advantage is supported by a limited number of studies that explicitly evaluated low-data scenarios. Critical integration challenges, including sensor calibration, data fusion, and contextual adaptation, are identified. The findings establish an evidence-based foundation for the subsequent design, simulation, and prototyping of an integrated hardware–software system, enabling a transition from reactive testing to continuous and predictive aflatoxin risk surveillance of grain storage systems.