International Journal of Scientific Research and Engineering Development

Soil fertility is a key determinant of sustainable agriculture, influencing nutrient absorption, crop performance, and long-term soil health. Yet, many farmers struggle to apply fertilizers correctly due to limited access to scientific soil assessment and reliance on guesswork. Conventional soil testing requires laboratory infrastructure, multiple sampling cycles, and delayed reporting, making it impractical for timely decision-making. As a result, excessive chemical usage remains common, contributing to nutrient imbalance, declining organic matter, and ecological damage. To address these issues, this paper presents an AI-Assisted Soil Condition Intelligence Framework that integrates multimodal data sources to evaluate soil health and nutrient adequacy. The system processes historical Soil Health Card (SHC) records along with current environmental parameters such as temperature, humidity, and rainfall using a sequence-learning model based on Long Short-Term Memory (LSTM) networks. In parallel, visual cues from crop leaves are analyzed using an image-efficient deep vision architecture derived from EfficientNet-B7 to detect early-stage nitrogen variation through Leaf Color Index (LCI) estimation. The predictions from the temporal soil model and the image-based nitrogen classifier are fused through a dynamic weighting mechanism that generates a unified Soil Vitality Index (SVI). This index provides farmers with real-time, crop-specific nutrient recommendations. By merging data-driven soil analytics with lightweight plant-visual assessment, the proposed system offers rapid, accessible, and field-ready guidance, enabling farmers to reduce fertilizer misuse, preserve soil ecosystems, and improve agricultural productivity.

P.Logaiyan, R.Suriya,"AI-Augmented Soil Fertility Index Generation Using Temporal and Visual Deep Learning Pipelines" International Journal of Scientific Research and Engineering Development, V9(3): Page(1304-1309) May-June 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.