{"title":"A comprehensive review of AI-driven plant stress monitoring and embedded sensor technology: Agriculture 5.0","authors":"Biplob Dey , Romel Ahmed","doi":"10.1016/j.jii.2025.100931","DOIUrl":null,"url":null,"abstract":"<div><div>To address the future demand for food, fiber, and fuel, crop production must double by 2050. It requires reshaping Agriculture 5.0 by overcoming typical limitations in crop yields caused by biotic and abiotic stresses while integrating computational power, artificial intelligence (AI), and sensor technology. To date, many studies have explored AI applications in plant stress monitoring and detection, and a critical synthesis of algorithm suitability, sensor integration, and critical assessment of industrial scalability is still lacking. Thus, this review systematically evaluates 175 peer-reviewed articles (from an initial pool of 687, published between 2010 and 2024) to identify trends, performance benchmarks, and integration challenges in AI-driven plant stress detection. The trend analysis revealed a substantial increase in AI and sensor applications for plant stress monitoring, specifically after COVID-19, although ∼67 % of studies limit classification tasks to 5 or fewer classes, often lacking field validation. CNN-based classification models (e.g., VGG16, VGG19, and ResNet50) consistently perform well across stress types, whereas detection-focused models such as YOLO and lightweight architectures such as MobileNet show greater variability, particularly in biotic stress identification tasks. Traditional machine learning methods, such as support vector machines, decision trees, and k-nearest neighbors remain relevant for structured, low-resolution data, especially under constrained conditions. Optimization algorithms such as stochastic gradient descent (biotic stress) and Adam (abiotic stress) are widely used. Advancements in sensor technologies, including hyperspectral imaging, volatile organic compound (VOC) detection via electronic noses, and real-time monitoring systems enable noninvasive and continuous stress detection. In parallel, the deployment of industrial robots equipped with embedded AI and multimodal sensors allows for automated, high-frequency stress surveillance and precision intervention in commercial-scale crop systems. Aligning AI models with specific sensor modalities is pivotal for developing scalable, interoperable, and industrial-grade monitoring systems.</div></div>","PeriodicalId":55975,"journal":{"name":"Journal of Industrial Information Integration","volume":"47 ","pages":"Article 100931"},"PeriodicalIF":10.4000,"publicationDate":"2025-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Industrial Information Integration","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2452414X25001542","RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS","Score":null,"Total":0}
引用次数: 0
Abstract
To address the future demand for food, fiber, and fuel, crop production must double by 2050. It requires reshaping Agriculture 5.0 by overcoming typical limitations in crop yields caused by biotic and abiotic stresses while integrating computational power, artificial intelligence (AI), and sensor technology. To date, many studies have explored AI applications in plant stress monitoring and detection, and a critical synthesis of algorithm suitability, sensor integration, and critical assessment of industrial scalability is still lacking. Thus, this review systematically evaluates 175 peer-reviewed articles (from an initial pool of 687, published between 2010 and 2024) to identify trends, performance benchmarks, and integration challenges in AI-driven plant stress detection. The trend analysis revealed a substantial increase in AI and sensor applications for plant stress monitoring, specifically after COVID-19, although ∼67 % of studies limit classification tasks to 5 or fewer classes, often lacking field validation. CNN-based classification models (e.g., VGG16, VGG19, and ResNet50) consistently perform well across stress types, whereas detection-focused models such as YOLO and lightweight architectures such as MobileNet show greater variability, particularly in biotic stress identification tasks. Traditional machine learning methods, such as support vector machines, decision trees, and k-nearest neighbors remain relevant for structured, low-resolution data, especially under constrained conditions. Optimization algorithms such as stochastic gradient descent (biotic stress) and Adam (abiotic stress) are widely used. Advancements in sensor technologies, including hyperspectral imaging, volatile organic compound (VOC) detection via electronic noses, and real-time monitoring systems enable noninvasive and continuous stress detection. In parallel, the deployment of industrial robots equipped with embedded AI and multimodal sensors allows for automated, high-frequency stress surveillance and precision intervention in commercial-scale crop systems. Aligning AI models with specific sensor modalities is pivotal for developing scalable, interoperable, and industrial-grade monitoring systems.
期刊介绍:
The Journal of Industrial Information Integration focuses on the industry's transition towards industrial integration and informatization, covering not only hardware and software but also information integration. It serves as a platform for promoting advances in industrial information integration, addressing challenges, issues, and solutions in an interdisciplinary forum for researchers, practitioners, and policy makers.
The Journal of Industrial Information Integration welcomes papers on foundational, technical, and practical aspects of industrial information integration, emphasizing the complex and cross-disciplinary topics that arise in industrial integration. Techniques from mathematical science, computer science, computer engineering, electrical and electronic engineering, manufacturing engineering, and engineering management are crucial in this context.