In addition to the prevalent and vital applications of artificial intelligence (AI) in public safety and security, it is important to acknowledge the potential hazards associated with its usage. Nonetheless, these potential dangers can also serve as a defensive strategy in concealment of vulnerable assets from autonomous systems such as unmanned aerial vehicles (UAVs) and drones. The research community is widely cognizant of the perils associated with adversarial attacks on deep neural network (DNN) models through which they are deceived and forced to make incorrect predictions. The objective of these attacks is to induce the DNN model to generate imprecise predictions or assessments by resolving the optimization problems anchored on target network architecture. The present study showcases novel research that employs adversarial attacks to mislead self-governing airborne detection systems integrated into aerial reconnaissance, detection, and surveillance systems. Our study presents a framework for the purpose of deceiving the YOLOv4 object detection model via a patch-based attack strategy. Experiments were conducted by concealing a stationary aircraft from aerial imagery through the use of AI-based disguise utilizing an optimized adversarial patch. The application of patches on a limited surface area of an aircraft resulted in a significantly improved ability to deceive autonomous detection systems while reducing detection accuracy up to 86% and 70% in digital and physical domains, respectively. Furthermore, this study highlights the application of the aforementioned methodologies in security-focused scenarios and underscores the necessity for more resilient detection frameworks.