Abstract
With the increase in the prevalence of armed robberies and attacks in public places, an effective and reliable surveillance system has become indispensable to fulfill various security aspects and improve the quality of human life. The most common effective video surveillance system used is called closed-circuit television (CCTV), but it can be expensive, and it requires a large amount of memory. In addition to these problems, the need for manpower to detect unauthorized activities can lead to several security problems. An intelligent surveillance system (ISS) that may overcome most of the mentioned problems is proposed in this paper. This system is built with more affordable hardware components that provide automated security services more than what regular surveillance systems can offer to the user with better quality using the assistance of Artificial Intelligence (AI), Deep learning (DL), and Image Processing (IP). The obtained overall detection accuracy of 83.1% at 22.2ms helps to inform the user in a short time during the incident. Also, the classification Mean Average Precision (mAP) of 0.98, 0.95, and 0.94 for detecting Knives, Guns, and Pistols, respectively, can reduce the rate of disturbing the user with false alerts.