MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641072751 A) filed by Vellore Institute Of Technology on June 11, 2026, for Hybrid Context-Aware Satellite Scene Understanding System And Method Using Semantic Segmentation And Object Detection.
Inventors include K Rajakumar; Tejas Chauhan; Charan Rajana; Kaviti Sravan; and Kasetti Sai Bhargav.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: HYBRID CONTEXT-AWARE SATELLITE SCENE UNDERSTANDING SYSTEM AND METHOD USING SEMANTIC SEGMENTATION AND OBJECT DETECTION A hybrid satellite scene understanding system (100) for automated interpretation and analysis of satellite and aerial imagery using integrated semantic segmentation, object detection, contextual validation, and scene intelligence generation. The system comprises an image acquisition module (101), a preprocessing module (102), a semantic segmentation module (103), an object detection module (104), a contextual validation module (105), a feature analysis module (106), a scene intelligence generation module (107), and an output generation module (108). The semantic segmentation module classifies terrain regions, while the object detection module identifies spatial objects within satellite imagery. The contextual validation module performs terrain-aware validation between detected objects and segmented terrain regions to improve detection reliability. The feature analysis module classifies environmental structures including water regions using geometric and spectral analysis. The system further generates infrastructure analytics, land-use analysis, environmental monitoring, and automated scene intelligence reports for real-time geospatial interpretation and remote sensing applications. [Figure 1]
Disclaimer: Curated by HT Syndication.
