MUMBAI, India, May 29 -- Intellectual Property India has published a patent application (202641041788 A) filed by V. Bhavana; B. Harsha Vardhan Reddy; N. Akhil Yadav; Rajeev Gandhi Memorial College Of Engineering & Technology; and Mr. Y. P. Srinath Reddy, Nandyal, Andhra Pradesh, on April 1, for 'real-time traffic object detection and multi-object tracking using yolo-based deep learning.'
Inventor(s) include Mr. Y. P. Srinath Reddy; N Akhil Yadav; V. Bhavana; and B. Harsha Vardhan Reddy.
The application for the patent was published on May 29, under issue no. 22/2026.
According to the abstract released by the Intellectual Property India: "Existing object detection approaches in traffic video analysis mainly focus on frame-by-frame detection and classification of objects using deep learning models. These methods are effective in identifying vehicles and locating them within individual frames; however, they generally treat each frame independently and do not maintain object identity across time. As a result, such systems often lack multi-object tracking capability, temporal consistency, and reliable vehicle counting in dynamic traffic scenes. To address these limitations, the proposed work presents an AI-based intelligent traffic monitoring system that integrates deep learning-based object detection with multi-object tracking and stable vehicle counting. The system processes traffic video obtained from surveillance cameras or recorded video streams and converts it into sequential frames for analysis. A YOLO-based object detection model is used to identify vehicles and generate bounding boxes around them, after which a tracking mechanism assigns unique IDs to detected vehicles and follows their movement across consecutive frames. To improve reliability, the system incorporates a track stability verification stage that ensures objects are consistently detected over time, reducing false detections and identity switching. Based on these stable tracking identities, an adaptive vehicle counting method is used to estimate the number of vehicles present in the monitored area. The processed frames are then annotated and displayed through a monitoring interface for traffic observation. By combining detection, tracking, and stable counting within a single framework, the proposed system enables more reliable and continuous traffic monitoring while reducing manual supervision, making it suitable for deployment in intelligent transportation systems and smart city traffic management applications."
Disclaimer: Curated by HT Syndication.
