<p>Over the last decades, unmanned aerial vehicles (UAVs), known as drones, have impacted many military applications and public services. However, UAVs have also been used for numerous illegal activities such as espionage, privacy invasion and terrorist attacks. This work consists in developing an integrated AI powered airspace monitoring system that defends restricted areas from malicious UAVs. The released system is an integrated suite of electro-optical selected sensors. It performs real-time monitoring and analyzes air images using a custom convolutional neural network based model. We set-up the proposed sky monitoring system in a no-fly zone and tested its capacity to detect mini-UAVs at different distances under different weather conditions. Test results prove the performance achievable by the proposed monitoring system and data analysis technique compared to other object detectors.</p>