The rapid expansion of devices on the Internet of Things (IoT), particularly in surveillance through IoT cameras, has provided innovative cybersecurity solutions. This paper introduces SecureCamPot, an augmented reality (AR)-enhanced honey-pot designed specifically for IoT cameras. SecureCamPot utilizes AR to emulate realistic IoT camera behaviors and respond to control commands dynamically, significantly enhancing deception capabilities. Incorporating automated object prediction with reinforcement learning, SecureCamPot excels in creating deceptive interactions that mislead potential attackers effectively. Our evaluation of several metrics, including the duration of the engagement and the complexity of attacker behaviors, indicates that SecureCamPot deceives approximately 72.5% of attackers, significantly outperforming traditional honeypots.