Modelling and Simulation of UAV Based Aerial Surveillance for Anomaly Detection

Anomaly Detection is one of the crucial components in real-time scenarios in various fields including surveillance, where identifying irregular patterns can address issues such as potential threats, crop field infestations, forest fires, etc. Advanced detection methods leverage sensor technologies to analyze environmental data, enhancing the accuracy and efficiency of identifying anomalies. This paper explores an Unmanned Aerial Vehicle (UAV) based aerial surveillance system that integrates these technologies for robust anomaly detection in various real-time environments. A virtual and realistic 3D environment using Blender is designed, which is then exported and simulated in Gazebo for UAV navigation and detection in the 3D environment. The UAV is equipped with sensors such as GPS, Inertial Measurement Units (IMU), Camera, and LIDAR which are configured using the SDF files. These sensors capture anomalies in the environment and then process them to identify them and mark their coordinates. The ROS2 nodes provide seamless communication with the UAV and its sensors. The autonomous mission planning for the UAV based aerial surveillance is achieved using PX4-Autopilot which provides a mighty framework for defining and executing UAV missions, which ensures efficient navigation and systematic coverage of large surveillance areas and enables the UAV to autonomously follow the pre-defined paths. OpenCV is utilized to identify anomalies within the simulated environment. This paper underscores the potential of autonomous UAV systems in enhancing surveillance capabilities across diverse applications.