Intellectual Property India Publishes Patent Application for 'Amphibious Drone System With Deep Learning-Based Underwater Object Detection Using Yolov8 And Autonomous Navigation For Marine Exploration' Filed by Nitte Meenakshi Institute Of Technology, Nitte; Dr. Nalini N; Dr. P. N Tengli; Chinmayi V Naik; Aditi Singh; Affaan Rafeeq M; and Karan Mohan

MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641051253 A) filed by Nitte Meenakshi Institute Of Technology, Nitte; Dr. Nalini N; Dr. P. N Tengli; Chinmayi V Naik; Aditi Singh; Affaan Rafeeq M; and Karan Mohan, Bangalore, Karnataka, on April 22, for 'amphibious drone system with deep learning-based underwater object detection using yolov8 and autonomous navigation for marine exploration.'

Inventor(s) include Nitte Meenakshi Institute Technology, Nitte; Dr. Nalini N; Dr. P. N Tengli; Chinmayi V Naik; Aditi Singh; Affaan Rafeeq M; and Karan Mohan.

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 present invention relates to an amphibious drone system (100) for underwater object detection and autonomous navigation employing the YOLOv8 deep neural network architecture. The system comprises a data acquisition module (110) with underwater cameras, environmental sensors, sonar, IMU, and positioning systems; an image preprocessing engine (120) performing depth-dependent color correction, CLAHE contrast enhancement, noise filtering, and data augmentation to address underwater optical degradation; a YOLOv8 detection engine (130) with CSPDarknet backbone, FPN-PAN multi-scale neck, and anchor-free detection heads trained on a custom dataset of more than 1350 annotated underwater images derived from real-world video footage, achieving mean average precision of 74.9 percent, precision of 69.8 percent, and recall of 76.5 percent; a post-processing module (140) with non-maximum suppression and object tracking; and a navigation decision engine (150) that generates obstacle maps, computes collision-free paths, and executes real-time collision avoidance for autonomous underwater drone operation. A feedback loop (190) continuously improves detection and navigation performance. The invention enables safe autonomous underwater operation of amphibious drones for marine research, environmental monitoring, infrastructure inspection, and underwater mapping applications."

Disclaimer: Curated by HT Syndication.