In tomato farms, the variation in ecological circumstances causes a disease known as leaf curl. in this paper, we have used a combination of deep learning (DL) and machine learning (ML) classifiers for the accurate detection and prevention of leaf curl disease. Initially, we designed a Smart Farm Monitoring System (SFMS), an energy-efficient prototype for monitoring and collection of real-time data from the tomato farms. Each sensing nodes senses temperature, humidity and soil moisture data from the tomato farms, aggregate them, and forward only a single packet towards the edge server. The energy-efficient data aggregation at the node level ensures that the highly-refined and fused data is available at the edge server. The use of ML and DL allows various computational models at the edge server to represent data with multiple levels of abstraction for the classification of tomato farm\('s\) data. The use of energy-efficient data aggregation at the node level and computational models at the edge server ensures that highly refined and redundant free data facilitate accurate disease prediction. Our proposed SFMS prototype shows better results as compared to the traditional approaches. For disease classification, the results obtained with different ML/DL algorithms were: ANN accuracy 92%, PNN 42% , logistic regression 97.46%, and GRNN 60%, respectively.

在番茄农场，生态环境的变化会导致一种称为卷叶的疾病。在本文中，我们结合使用深度学习 (DL) 和机器学习 (ML) 分类器来准确检测和预防卷叶病。最初，我们设计了一个智能农场监控系统 (SFMS)，这是一个用于监控和收集番茄农场实时数据的节能原型。每个传感节点从番茄农场感知温度、湿度和土壤水分数据，将它们聚合起来，然后仅将单个数据包转发到边缘服务器。节点级别的高能效数据聚合确保了高度精细化和融合的数据在边缘服务器上可用。\('s\)数据。在节点级别使用节能数据聚合和边缘服务器的计算模型可确保高度精细和冗余的免费数据有助于准确的疾病预测。与传统方法相比，我们提出的 SFMS 原型显示出更好的结果。对于疾病分类，使用不同 ML/DL 算法获得的结果分别为：ANN 准确度 92%、PNN 42%、逻辑回归 97.46% 和 GRNN 60%。