MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641094836 A) filed by Dhanalakshmi Srinivasan College Of Engineering And Technology on August 05, 2026, for Deep Learning - Driven Change Detection Framework For Pre And Post Flood Impact Analysis.
Inventors include K. Senbagam; Dhanush S; Dilli Babu K; and Gopinath An S.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: The present invention relates to a deep learning-based system and method for automated flood detection and pre- and post-flood impact analysis using Synthetic Aperture Radar (SAR) satellite imagery. The system acquires pre-flood and post-flood SAR images from satellite sources and processes them through a preprocessing module that performs orbit correction, thermal noise removal, radiometric calibration, speckle noise filtering, terrain correction, and image normalization to generate high-quality data suitable for analysis. A deep learning-based feature extraction module then extracts meaningful spatial features from the processed images, while a change detection module compares temporal SAR images to identify newly flooded regions. The extracted features are further processed by a semantic segmentation model to classify each pixel as flooded or non-flooded, enabling accurate delineation of flood boundaries. The system generates flood extent maps, statistical impact reports, and temporal flood change analysis through a web-based visualization interlace. The proposed framework provides an automated, scalable, and weather- independent solution for flood monitoring, supporting disaster management authorities in rapid damage assessment, emergency response, resource allocation, and recovery planning.
Disclaimer: Curated by HT Syndication.
