<![CDATA[The presence of surveillance cameras plays an important role in helping the process of monitoring and evaluating human activities in the monitored area. This ability can prevent or trace undesirable events such as criminal acts or some accidents that related to human activities. However, most of the surveillance camera that used nowadays only held a passive role in security that can lead to an increased potential risk of negligence by the guards (users) in the process of monitoring the activities that are happening. This study aims to design a system that is able to improve the performance of surveillance cameras in detecting and calculating numbers of human based on Movidius NCS on a Raspberry Pi device so that the camera can be active and be able to provide optimal results and reduce the use of excess space on the storage. The human object detection system that is used in this research applies Deep Learning technique with Mobilenet-SSD as its network architecture model. The research trials were carried out under various conditions of light intensity starting from 50-550 lux and distance to objects in range of 1-10 meters. The results showed that the accuracy obtained by the system was able to reach 91.67% with 49.24% of storage efficiency.]]>
Copyrights © 2022
