Multimodal Deep Learning – IoT Systems for Tomato Crops: A Systematic Review
Multimodal Deep Learning – IoT Systems for Tomato Crops: A Systematic Review

Summary/Abstract: The paper aims to provide a systematic synthesis of recent literature on multimodal deep learning–IoT systems applied to tomato crops, placing them in the broader context of precision agriculture and rural development. The digital transformation of agriculture, marked by the intensive use of sensors, IoT platforms and deep learning models, creates important premises for improving productivity, risk management and economic sustainability of farms, but, at the same time, generates a fragmented and heterogeneous solution landscape. Against this background, the study pursues three main objectives: (O1) identifying and classifying the main multimodal deep learning–IoT systems for tomato crops, based on the types of data used (images, environmental sensors, agronomic data), the proposed architectures and the implementation contexts; (O2) to critically review the technical approaches and reported performances, by comparing deep learning solutions and multimodal fusion strategies from the perspective of robustness, scalability and feasibility in real farm conditions; (O3) to identify research gaps and formulate future directions, with a focus on the potential of these systems to support technical and economic decisions in agriculture and to contribute to rural development. Methodologically, the paper follows a systematic review approach, based on the query of the main international scientific databases and the application of explicit study inclusion criteria. The results show the existence of significant progress in the detection of foliar diseases and microclimate monitoring, but also important limitations related to the size and quality of the datasets, the lack of economic evaluations and long-term studies in commercial farms. The main conclusion is that multimodal deep learning–IoT systems for tomatoes represent a promising but insufficiently integrated field, and this paper provides a useful reference framework for both researchers and decision-makers interested in the digitalization of agriculture and strengthening the resilience of rural farms.

• Details
• Contents