International Journal of Science, Engineering and TechnologyAuthors: Tummalapalli Jaswanth Nagasairam, Dasa Divya Santhoshi, S.Deepajothi
Abstract: Growing international security threats also mean that autonomous surveillance research systems are required to undertake real-time perception, adaptative assessments of threats and decentralized functioning within limited communication conditions. The traditional monitoring systems are dependent on the centralized processing and human monitoring that lead to slow detection of threats, high bandwidth usage, and inability to be responsive to complex terrain situations. To surmount such constraints, an autonomous defence surveillance research framework of border threat intelligence is provided. The architecture involves the integration of high resolution optical imaging, thermal sensing module and unmanned monitoring platform to make persistent observation of the perimeter. Streams of sensor data are processed at the local edge, and thus can be analyzed in a low-latency manner and still operated continuously in distant areas. An object recognition and multi-object tracking mechanism using deep learning allows identifying intruders, vehicles, wildlife, aerial objects and evaluating temporal trajectory allows behavior interpretation and analysis of intrusion patterns. A rule-based threat intelligence model combines classification result, movement patterns, and contextual limitations to determine the degree of threat severity and create takeable alerts. Edge-centric inference minimizes network dependency, maximizes the level of operational confidentiality, and speed up the execution of responses. An average detection accuracy of 96.4 was experimentally assessed and the proposed output implementation model was capable of constant real-time performance under a variety of conditions with illumination levels, weather and terrain changes. Effective situational awareness, credible threat classification and efficient decision making are confirmed by observed results to improve the effectiveness of autonomous border surveillance operations.
DOI: https://doi.org/10.5281/zenodo.18628269
