MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641049615 A) filed by Srm Institute Of Science And Technology; R Harshavardhan; Manoj Kumar A; Soorej S; and Sabitha P, Chennai, Tamil Nadu, on April 18, for 'an event-driven sar change detection device utilizing multithreshold spiking u-net hardware acceleration.'
Inventor(s) include R Harshavardhan; Manoj Kumar A; Soorej S; and Sabitha P.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "The proposed invention presents a high-performance, energy-efficient architecture entitled ''An Event-Driven SAR Change Detection Device employing Multithreshold Spiking U-Net Hardware Acceleration''. The system is real time remote sensing capable and is designed to analyze SAR data in a neuromorphic fashion, overcoming the frame-based processing limitations. Using an event driven approach, this device minimizes both data redundancy and overall energy consumption making it an ideal candidate for edge computing in satellite and airborne observation systems. The primary innovation in this system is the multi-stage processing pipeline. The raw SAR signals are converted to temporal events which are subsequently processed through a multithreshold generator to create a multi-dimensional sensitivity spike train. These spike patterns are then processed through a hardware accelerated Spiking U-Net which has the ability to identify and track complex Geo-spatial changes, such as flooding, urban sprawl, deforestation and other changes, with a high degree of temporal accuracy. The results of the analysis are generated in real time to an integrated display as a color-coded Change Detection Map for rapid response. This invention combines robust hardware with the energy efficiency of spiking neural networks to create a scalable and viable solution for environmental protection and conservation, disaster response, infrastructure assessment and monitoring, representing a significant leap forward in intelligent low latency remote sensing systems."
Disclaimer: Curated by HT Syndication.
