Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{198547,
        author = {Yashwanth Gouda S and Dr. Chatrapathy K and Vinay R and Vishwas B A and Yashas S},
        title = {Krishi: An Integrated Artificial Intelligence Framework for Precision Agriculture and Smart Crop Advisory in Indian Smallholder Farming},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {11688-11694},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=198547},
        abstract = {Indian agriculture, which supports over 600 mil-lion people, suffers from persistent productivity gaps stemming from limited access to expert agronomic advice, delayed disease diagnosis, and poor market intelligence for smallholder and marginal farmers. This paper presents Krishi, a full-stack, cloud-deployed intelligent advisory platform that unifies five AI-driven modules—crop recommendation, fertilizer advisory, plant disease detection, market price forecasting, and real-time weather analytics—into a single multilingual interface. The disease detection subsystem employs a novel Ensemble Vision Engine comprising two MobileNetV2-based deep learning models fine-tuned on the Plant Village dataset (54,306 images, 38 disease classes), aggregated with confidence-weighted voting (60%/40%), achieving an empirical detection accuracy of 95–98%. A rule-based Plant Guard pre-processing layer filters non-leaf inputs through spectral greenness, texture variance, and skin/soil heuristics before invoking the ensemble, dramatically reducing false positives caused by non-plant objects. The crop recommendation module employs a Random Forest classifier trained on 1,200 agronomic samples covering 24 crop types under Indian climate regimes. Market price forecasting uses ARIMA time-series modelling for a 6-month forward outlook. A React progressive web application with Web Speech API integration delivers a barrier-free, voice-navigable interface in English, Hindi, Kannada, and Telugu. Experimental evaluation validates superior accuracy and user-experience outcomes compared to existing rule-based agricultural advisory tools.},
        keywords = {Smart farming, crop recommendation, plant disease detection, ensemble learning, MobileNetV2, precision agriculture, natural language interface, ARIMA forecasting, fertilizer advisory.},
        month = {April},
        }