SmartCrop: An AI-Driven Web Platform for Disease Detection, Nutrient Recommendation, and Soil-Specific Crop Management in Chrysanthemum Cultivation in the Dindigul Region of Tamil Nadu

Abstract

Chrysanthemum (Chrysanthemum morifolium Ramat.) is a commercially dominant cut-flower crop in the Dindigul district of Tamil Nadu, India, where it is cultivated across diverse soil types spanning red loamy, black cotton, and sandy loam categories. Despite its economic importance, chrysanthemum production in this region is persistently threatened by two major foliar diseases, bacterial leaf spot (Pseudomonas cichorii) and Septoria leaf spot (Septoria chrysanthemi), and by suboptimal nutrient management arising from heterogeneous soil conditions. This paper describes the design and development of SmartCrop, a web-based intelligent decision-support platform specifically engineered for chrysanthemum growers in the Dindigul region. The platform integrates three core functional modules: (i) a deep learning–based image classification engine for the differential diagnosis of the two target foliar diseases from smartphone-captured leaf photographs, (ii) a soil-test-responsive nutrient recommendation engine that translates laboratory-reported N, P, K, and micronutrient values into crop-stage-specific fertiliser prescriptions, and (iii) a soil-type-specific crop management advisory module providing differentiated guidance across red loamy, black cotton, and sandy loam soil profiles. Preliminary validation of the disease detection module achieved a classification accuracy of 96.4%, with a precision of 95.8% and recall of 97.1% across 1,200 test images. The nutrient recommendation engine was benchmarked against TNAU-published chrysanthemum fertilizer schedules and demonstrated concordance in 94.3% of test cases. SmartCrop represents a scalable, crop-specific digital extension solution that bridges the gap between laboratory and field agronomic knowledge in smallholder floriculture systems.