Intellectual Property India Publishes Patent Application for Deep Learning Based System And Method For Real-Time Crop Monitoring And Automated Weed Detection Using Convolutional Neural Networks

MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202621048321 A) filed by Symbiosis International Deemed University on April 15, 2026, for Deep Learning Based System And Method For Real-Time Crop Monitoring And Automated Weed Detection Using Convolutional Neural Networks.

Inventors include Anushri Adapawar; Anushka Poshattiwar; Tanvi Bandebuche; and Dr. Priya Dasarwar.

The application for the patent was published on June 12, 2026, under issue no. 24/2026.

Abstract: ABSTRACT DEEP LEARNING BASED SYSTEM AND METHOD FOR REAL-TIME CROP MONITORING AND AUTOMATED WEED DETECTION USING CONVOLUTIONAL NEURAL NETWORKS The present invention provides a deep learning based system (100) and method for real-time crop monitoring and automated weed detection. The system comprises an image acquisition module (110), data preprocessing module (120) for image resizing and normalization, deep learning classification module (130) utilizing transfer learning with pre-trained InceptionV3 architecture having frozen base layers and custom classification layers including GlobalAveragePooling2D layer (132), Dense layer (134) with ReLU activation, BatchNormalization layer (136), Dropout layer (138), and output layer (140) with sigmoid activation for binary classification. The system further comprises Grad-CAM visualization module (150) for generating interpretable class activation maps and output interface module (160). The system achieves classification accuracy exceeding 98% and enables integration with automated agricultural machinery for precision weed management, reducing herbicide usage while improving crop yields. [

Disclaimer: Curated by HT Syndication.