Mohammad Nasar, Mohammad Abu Kausar, Md. Abu Nayyer
As the global population is expected to expand from 7.3 to 9.7 billion during the next three decades, agriculture has been put under tremendous pressure to increase productivity while dealing with the challenges of climate change and resource limitations. In this chapter, a novel technique for real-time crop nutrition monitoring using Edge AI and IoT cameras is presented. These models are based on lightweight CNNs and ViTs, which can be run on low-cost edge hardware (e.g., Raspberry Pi or NVIDIA Jetson modules) for real-time plant imagery analysis during nutrient deficiency detection, in a remote offline setting. High-resolution IoT cameras capture detailed leaf images, which are analyzed on edge to detect macro- and micronutrient deficiencies with high accuracy. The use of robotic arms for precision fertilization and AI in rural areas is illustrated through a comprehensive case study on dragon fruit cultivation, backed by pilot studies conducted on coffee and maize farms to provide some practical insights. The performance indicators like accuracy, throughput, and energy efficiency validate the proposed system as an effective solution for farm-level resource-constrained settings. Field-validated calibration and explainable AI (XAI) tools provide transparent and reliable system outputs so that farmers can have trust in the performance of the technology. In line with Industry 5.0's vision of human-centered, sustainable technology, the system transforms traditional farming into a data-driven ecosystem that increases productivity, lowers costs, and upholds a respect for the planet. In this chapter, we share an in-depth, user-friendly, technically sophisticated introduction to the technology and its implications for terrestrial agriculture worldwide.