MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641100173 A) filed by Karpagam College Of Engineering; Karpagam Academy Of Higher Education; Dr. V. Rajeswari; and Anishkha H M on August 19, 2026, for A Hybrid Spatial Temporal Framework For Behavior Aware Wildlife Monitoring Using Deep Learning.
Inventors include Dr. V. Rajeswari; and Anishkha H M.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: The monitoring of wildlife in their natural habitats is a fundamental activity in the preservation of biodiversity, the study of ecosystems, and the mitigation of conflicts between humans and wild animals. The conventional camera trap approach to wildlife monitoring is exhausting due to the need to manually scan through thousands of images and videos to identify specific patterns of behavior. As a result, there are issues of delay, misidentification, and the likelihood of overlooking crucial behaviors. As a solution to these challenges, this project offers an AI-based wildlife monitoring system for real-time detection of the targeted subjects and their behavior based on the videos screened by the camera traps. The system improves the monitoring of wildlife through automation, boosting efficiency, accuracy, and efficacy in the accomplishment of wildlife monitoring activities. The project approach employs a YOLOv8 deep-learning object detection algorithm to identify various animal species in different ecosystems within the forests, such as low light, dense vegetation, or rapid movement conditions. An LSTM approach is then employed to examine the patterns of behavior based on the movement of the animals, such as walking, running, grazing, resting, being alert, and others. All detections, timestamps, and behavior patterns are structured in a database format and visualized using an intuitive dashboard interface. This project shows how the integration of real time object detection and temporal analysis of behavior leads to enhanced capabilities within the wildlife surveillance system. Experimental results reveal good performance from both the species recognition and behavior classification, which reduces workload by manual mechanisms and provides consistent and actionable insights. Conclusion: AI driven approaches offer a feasible and practical solution for wildlife conservation to support informed decision-making, early threat detection, and ecological study in the long term.
Disclaimer: Curated by HT Syndication.
