Optimizing Crop Yield - An IoT and ML-based Soil Testing and Crop Recommendation System
S Prathosh, K. Veerasamy
Abstract
Agriculture is the primary source of income for a large segment of the Indian population, making this sector essential to achieving high crop yields for food security and economic stability. There is a growing concern that the misuse of fertilizers is disrupting nutrient balance, which negatively affects soil health and reduces productivity. Moreover, inderterminate climate patterns have affected crop yield leading to extensive farming losses. To address these challenges, a system leveraging the Internet of Things (IoT) and Machine Learning (ML) was developed for improved soil testing, real-time monitoring, and crop guidance. This innovative system employed a network of sensors that evaluated critical soil characteristics such as temperature, moisture, pH, and NPK levels. The data collected by these sensors was processed through a micro controller and analyzed using the Random Forest algorithm, resulting in precise nutrient recommendations. Additionally, the robustness of the model was further enhanced through Stacking and Adaptive Feature Fusion techniques. Additionally, the system included a built-in plant health monitoring module that utilized Integrated Convolutional Neural Networks (CNNs) to analyze images captured by integrated cameras, allowing for the early detection of plant diseases and timely interventions to prevent outbreaks. This solution refines agricultural practices through precise nutrient management and early disease identification, ultimately improving crop yields and promoting soil sustainability. In addition, this initiative has the potential to significantly transform the livelihoods of millions of farmers across India. The proposed approach outperformed existing algorithms achieving a 99.5 % accuracy.