Uncrewed aerial vehicles (UAVs), commonly known as drones, have become increasingly prevalent in various applications. However, the localization and payload detection of drones is crucial for ensuring safety and security. This thesis proposes a vision-based solution using deep learning techniques to address these challenges. Existing solutions like radars and acoustic sensors have limitations, including high costs, limited accuracy, and the need for prior knowledge of the drone's model. Normal radars lack angle estimation accuracy and rely on known micro-Doppler features for payload detection, while acoustic sensors are restricted to close ranges for payload analysis. In contrast, cameras offer a cost-effective alternative as they have become widely available and can capture visual data. In addition, due to resource constraints, mounting multiple sensors on the UAV along with the camera is impractical, making reliance on cameras alone essential for addressing the mentioned problems. Recent advancements in deep learning algorithms enable regression and classification tasks, making vision data a promising choice for solving drone localization and payload detection problems. The proposed solution leverages convolutional neural networks (CNNs) for regression tasks, estimating the distance of a drone from the captured image. The CNN takes a cropped image within the drone's bounding box as input and outputs the estimated distance. Additionally, the drone's azimuth and elevation angles have been estimated based on its position in the captured image using a simple pinhole model for the camera. Also, the ResNet and EfficientNet classifiers could accurately classify drones as loaded or unloaded, even without prior knowledge of their shape. Due to a scarcity of publicly available datasets, this study developed the first air-to-air simulated dataset specifically for the classification of loaded versus unloaded drones. To evaluate the performance of the proposed solution, both simulated and experimental tests were conducted. The results showcased promising accuracy, with a root mean square error (RMSE) of less than 10 meters for distance estimation and an RMSE of less than 3 degrees for angle estimation. Furthermore, the payload detection problem was effectively addressed, achieving a classification accuracy of over 85\% for distinguishing between loaded and unloaded drones using the trained network based on the simulated dataset. The numerical highlights demonstrate the effectiveness of using camera sensors for 3D localization, with accurate distance and angle estimations. The high accuracy achieved in payload classification showcases the potential of the proposed solution for detecting drone payloads at distances up to 100 meters. These results pave the way for enhanced safety and security in drone environments.

This paper will focus solely on the effectiveness of AV (antivirus) in detecting Metasploit payloads which have been encapsulated with different encapsulation modules. There seems to be a significant knowledge gap in the evaluation of commercial antivirus's software and their ability to detect malicious code and stop such code from being executed on IT systems. Therefore we would like to evaluate the capabilities of modern AV software with the use of penetration testing tools such as Metasploit. The research process is heavily reliant on a case study methodology as it can be argued that each payload generated reflects a case in itself. Firstly the payloads are generated and encapsulated through the self developed software, secondly they are uploaded to VirusTotal to be scanned with the use of their publicly available API, third the results are obtained from VirusTotal and stored locally. Lastly the results are filtered through with the software which in turn generates graphs of the results. These results will provide sufficient data in comparing encapsulation methods, payload detection rates, draw conclusions regarding which operating system may be most vulnerable as well as the overall state of modern AV software's capabilities in detecting malicious payloads. There are plenty of noteworthy conclusions to be drawn from the results, one of them being the most efficient encapsulation method powershell_base64 which had amongst the lowest detection rates in regards to the amounts of payloads it encoded, meaning that its encapsulation hid the malicious code from the AV at a higher degree than most the other encapsulation modules. The most noteworthy conclusion from the results gathered however is the encapsulation methods which obtained the absolute lowest detection rates, these were x86_nonalpha, x86_shikata_ga_nai, x86_xor_dynamic as well as payloads without any encoding at all, which had a few payloads reach among the lowest detection rates across the board (<20%).