MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202621099393 A) filed by Archit Dharmesh Khatri; Aditya Garad; Dhruv Gandhi; Harsh Doshi; and Prof. Nagaraju Bogiri on August 17, 2026, for System And Method For Non-Invasive Detection Of Sub-Surface Pathologies In Rhizomatous Crops Using Bi-Phasic Domain-Adaptive Inference.
Inventors include Archit Dharmesh Khatri; Aditya Garad; Dhruv Gandhi; Harsh Doshi; and Prof. Nagaraju Bogiri.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: Rhizo-Net is a domain-adaptive, bi-phasic deep-learning and computer-vision framework for non-invasive screening of sub-surface pathologies in rhizomatous crops such as Ginger and Turmeric. The invention receives a foliar image and processes it through two concurrent streams. A Spectral Pattern Analyzer uses MobileNetV2 with transfer learning to identify chlorosis, necrosis, and related spectral/texture anomalies, while a Morphological Geometry Tracker uses HSV masking, Gaussian smoothing, Canny edge detection, and edge-density calculation to quantify leaf rolling and structural deformation. A central Bi-Phasic Logic Gate combines these outputs and can incorporate rainfall/weather context to distinguish alternative sources of plant stress. The system further provides Grad-CAM visual explanations, diagnostic confidence, and actionable agricultural directives through a Streamlit interface. The approach is intended to provide an accessible, lightweight, and non- destructive diagnostic aid for early crop-health assessment.
Disclaimer: Curated by HT Syndication.
