MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541122260 A) filed by G. Narayanamma Institute Of Technology & Science (For Women), Hyderabad, Telangana, on Dec. 5, 2025, for 'automatic extraction of building rooftops from higher -resolution aerial imagery.'
Inventor(s) include Dr. Renuka Devi SM; Anjali Manchala; Sreya Doddi; Rishitha Sai Gattu; Sanjana Gowdiperu; P. Lavanya; G. Madhavi; Y. Prakash; Dr. P. Satyanarayana Goud; and Anagama Rajitha.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "The present disclosure relates to a computer-implemented system and method for automatically extracting building rooftops from higher-resolution aerial imagery using deep learning-based semantic segmentation. The system receives aerial or satellite images and performs pre-processing including resizing, normalization, augmentation, and mask generation. A deep-learning segmentation model, such as U-Net or a Multiple-Parallel Vision Mamba Network (MVM-Net), processes the pre-processed images through an encoder-bottleneck-decoder architecture to generate pixel-wise rooftop predictions. The model is trained using a hybrid Binary Cross-Entropy and Dice loss function to enhance boundary accuracy and segmentation robustness. Post-processing operations refine the predicted rooftop masks by removing noise and smoothing contours using morphological filtering. The refined masks are overlaid on the original aerial images to generate the final rooftop extraction output. The disclosure enables accurate, automated, and scalable rooftop segmentation suitable for applications in urban planning, GIS mapping, solar assessment, and disaster management."
Disclaimer: Curated by HT Syndication.
