MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202641121085 A) filed by Rajalakshmi Engineering College on October 09, 2026, for Ai Crop Disease Predictor Using Soil And Leaf Data.
Inventor includes Dr. M. Thangamani.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: This specification describes a portable apparatus for predicting crop diseases before visible symptoms appear by combining two sensing modalities processed through a purpose-built neural network. A soil acoustic sensing module (100) captures root-zone micro-vibrations using a piezoelectric contact 605 microphone probe (110) inserted at controlled depth into the root zone, where acoustic signatures differ measurably between healthy root systems and those under pathogenic attack. A hyperspectral leaf reflectance analysis module (200) measures leaf reflectance across at least four discrete wavelength bands from 450 nm to 1600 nm using a muhispectral camera unit (210) with controlled illumination (230), capturing spectral changes caused by early-stage infection before visible symptoms develop. A dual- 610 encoder fusion neural network processing unit (300) processes both data streams through separate encoder subnetworks (310, 320) joined by a cross-attention fusion layer (330) that learns diagnostic interactions between below-ground acoustic and above-ground spectral indicators. The network outputs disease class, severity estimate, and spread probability. The complete apparatus operates as a battery-powered handheld field instrument performing edge inference without network connectivity, with calibration systems (700) 615 for sustained field accuracy and cloud interfaces (600) for model improvement and regional disease surveillance.
Disclaimer: Curated by HT Syndication.
