Optimizing Lettuce Cultivation: Nutrient and Disease Monitoring in Vertical Farms

Plant health management in vertical farming can undergo a revolution through the utilization of artificial intelligence (AI) and computer vision for real-time detection of nutrient deficiencies and diseases in lettuce plants. To tackle this challenge, this study delves into state-of-the-art convolutional neural network (CNN) models, encompassing VGG16, VGG19, ResNet50, EfficientNetB0, MobileNetV3, and Xception. These models underwent meticulous training and fine-tuning, harnessing transfer learning techniques to heighten accuracy and convergence despite limited data. The significance of this endeavor lies in its capacity to elevate and refine vertical farming practices. Manual assessment of plant health proves labor-intensive and error-prone, impinging on yield and resource efficiency. By automating diagnostics via AI-driven models, this work aspires to alleviate these hurdles and optimize crop production. This study's dataset encompasses an all-encompassing array of lettuce images, capturing diverse health conditions, nutrient scarcities, and disease indications. The methodological approach adopted here guarantees reproducibility by illuminating model selection, training protocols, and dataset curation. The study unveils findings that underscore the precision and resilience of AI-based diagnostics. The seamless integration of these models into vertical farming systems could potentially chart the course for sustainable and robust crop cultivation, curtailing losses and maximizing yields through well-timed interventions.