A research team from the Department of Computer Science and the School of Agriculture, University of Jammu, has developed an Artificial Intelligence model capable of predicting crop yields in J&K’s complex hilly terrain with 87% accuracy. The research, published in the journal Computers and Electronics in Agriculture, could revolutionize agricultural planning in the region.
The AI model uses satellite imagery from ISRO’s Resourcesat-2, ground weather station data, and historical crop records to predict yield outcomes 6-8 weeks in advance. It was trained on 15 years of agricultural data from 1,200 farm plots across Jammu, Rajouri, and Poonch districts.
The model can predict yields for major J&K crops including Basmati rice (Jammu region), maize, wheat, and saffron (Kashmir). Early adoption by even 20% of farmers could increase regional food production by an estimated 12-15%, according to the research paper.
The J&K Agriculture Department has shown interest in scaling the technology for use in its e-Kisan portal. ISRO has offered to provide satellite data at no cost for a 2-year pilot program across all agricultural districts of J&K.
The research was funded by the Department of Science and Technology (DST) under the Technology Innovation Hub on AI for Agriculture. Contact the project PI: Prof. Deepak Sharma at [email protected].
