IoT based Low Cost Fertilization and Irrigation Monitoring System for Cotton Crop field
Loading...
Date
item.page.authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The escalating demand for sustainable agricultural practices coupled with the advent of
newlinemodern technologies like Machine Learning (ML) and Artificial Intelligence (AI) has spurred
newlinesignificant advancements in the domain of Smart Farming. This thesis endeavors to explore and
newlinecontribute to the intersection of ML, AI, and agriculture, particularly focusing on the
newlineenhancement of farming practices through data-driven approaches. The thesis commences with
newlinean empirical evaluation of fertilization and irrigation monitoring systems, addressing the
newlineintricate design and integration requirements. It highlights the plethora of components essential
newlinefor the efficient operation of such systems, encompassing agricultural sensors, actuation devices
newlineand deployments, communication interfaces, and fertilization mechanisms. These systems rely on
newlinediverse data inputs from soil sensors, crop quality sensors, and yield sensors, which undergo
newlineanalysis through various processing models and algorithms. However, the selection of an
newlineappropriate model often poses a challenge due to the multitude of available options, leading to
newlinedelays in system deployment and increased costs. To mitigate this ambiguity, the thesis presents
newlinea comprehensive review of fertilization and irrigation monitoring models, comparing them
newlineacross performance metrics such as accuracy, deployment cost, complexity, and scalability.
newlineThrough this analysis, agricultural stakeholders can discern best practices for system design and
newlinedeployment, thereby streamlining the implementation process. Additionally, the thesis proposes
newlineenhancements and model fusion techniques to improve system performance while reducing
newlinevalidation efforts, particularly emphasizing the optimization of deep learning models to mitigate
newlinecomputational complexities.
newlineSubsequently, the thesis introduces the quotMSMRBEFquot framework, a Multimodal framework for
newlineSoil Monitoring and Recommendation via Bioinspired Ensemble Feature processing operations.
newlineRecognizing soil monitoring as a multidomain p