Author(s): Sekou Hera
Mentor(s): Prosenjit Chatterjee
Institution SUU
Applications are growing steadily, from commercial planes to military drones and unmanned aerial systems, the demand is increasing fast for real-time, automated solutions to the assessment of threats. The approaches to threat assessment, most of them relay human observations that bring inefficiency and erroneous classifications. This is an ongoing research into the automation of threat detection and classification posed by an aerial object using image recognition techniques. The proposed model will classify the aerial vehicles as None, Low, Medium, and High according to crucial characteristics such as type, speed, and weaponry. The aim was to develop a framework that would enhance activities relating to security like surveillance, defense, and air traffic management by improving the time factor and accuracy in response. Initial testing has given positive feedback; however, more analysis and adjustments are needed for final verification. The goal of this research is to develop an automated system capable of assessing threat levels associated with aerial objects, ranging from benign (None) to critical (High). The system will use a dataset of labeled images of commercial aircraft, military drones, and other aerial objects, categorized by threat level. Threat classification is based on attributes such as vehicle type, observable weaponry, and speed, providing a basis for early and accurate threat detection.