Utilizing Artificial Intelligence for Plant Phenotyping in Soilless Farming: An Innovative Deep Learning Approach on a Unique Dataset
Damodar Reddy, Krishna Priya R, J. Arun Kumar, P. Dharani Prasad, M. Saqib Nawaz, Nookala Venu
Abstract
In the dynamic landscape of agriculture, the convergence of technological advancements and sustainable practices has become imperative. The emergence of soilless farming, driven by hydroponics and aeroponics, presents a promising solution to address food security challenges while mitigating environmental concerns. Central to the success of soilless farming is efficient plant phenotyping, which traditionally relies on labour-intensive methodologies. However, this paradigm is shifting with the integration of Artificial Intelligence (AI) and Deep Learning techniques. This research endeavours to pioneer an innovative approach to plant phenotyping in soilless farming by harnessing the power of AI. Leveraging a unique dataset meticulously curated from controlled hydroponic and aeroponic environments, our study aims to redefine the boundaries of agricultural research and practice. By employing state-of-the-art Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), we dissect complex phenotypic traits encompassing leaf morphology, biomass accumulation, and physiological responses. Through iterative model refinement and validation, we strive to develop a robust framework capable of real-time phenotypic assessment across diverse plant species and growth stages. By synthesizing diverse environmental conditions and perturbations, we augment the original dataset, enhancing model generalization and adaptability. Moreover, through transfer learning techniques, Furthermore, the scalability and accessibility of AI technologies pave the way for democratizing agricultural innovation, fostering inclusive growth and resilience in the face of global challenges.