Transforming tomato agriculture: a comparative analysis using transfer learning approach for tomato leaf disease detection

Abstract

The fundamental purpose of this research-based project explores the frontier of tomato leaf detection disease using deep learning, leveraging cutting-edge technologies in the deep learning to address the challenges posed by a myriad of diseases affecting global tomato crops. Timely diagnosis and management of diseases is critical, as the agricultural sector plays a pivotal role in maintaining both food security and economic stability. Prescription datasets that have been specially gathered are used to train and assess six advanced models: ResNet152, ResNet50, MobileNetV2, InceptionV3, DenseNet201 and DenseNet121. The model with the highest accuracy of 98% among them is DenseNet201. Rest of them are, Densenet121 is 96.35%, MobilenetV2 is 94.22%, InceptionV3 is 89.06%, Resnet50 is 47.11% and Resnet152 is 41.03%. This work centers on the development of automated systems that can quickly and accurately identify a varietyof diseases that affect tomato yield, quality, and overall crop health. The systems are designed to be integrated with artificial intelligence and image processing techniques. The investigation explores tomato leaf disease detection techniques, obstacles, and developments, emphasizing how machine learning algorithms are transforming precision agriculture. This research aims to minimize losses, reduce environmental impact through targeted interventions, and optimize resource utilization, thereby contributing to the principles of sustainable agriculture. Farmers will be empowered with proactive tools for early disease detection. The main goal of this work is to help the agricultural sector with the valuable knowledge and tools, promote the sustainability, elevate the productivity of tomato cultivation.