Analysis and Implementation Face Recognition on CCTV for Smart Home Security Based on Internet of Things (IOT) with BOT Communication Media

Abstract

Nowadays, crime in the form of house robbery often occur when the owners are not in the present. Because of the mentioned reason, home security must be improved. There are various kinds of technology regarding home security; one of them is integration between CCTV and Internet of Things (IoT). With this technology, it’s expected to reduce the number of crimes.
For this reason, further, development is needed to overcome these problems, namely by creating a face recognition system on CCTV that can detect unknown objects and sent via Line Bot. After that, testing and analysis are carried out related to the system performance made.
From the results of testing features using the black-box testing method, the functions contained in the bot and the image sending feature of the Raspberry Pi can work well. Analysis of Raspberry Pi - Server - Server network analysis divides into three conditions, namely idle condition, normal condition, and full load conditions. In the idle condition has a delay of 35.99 Ms with a throughput of 188.7 Kbps, the normal condition has a delay of 37.08 Ms with a throughput of 184.2 Kbps, the full load condition has a delay of 38.93 Ms with a throughput of 176.44 Kbps. Reliability & Availability value on the Raspberry Pi - Server - Line network is 100%. For Line - Server - Line network analysis testing divides into three conditions, namely empty hour, normal hour and rush hour. In the empty hours has a delay of 109.64 Ms with a throughput of 9.9 Kbps, in the normal hours has a delay of 112.73 Ms with a throughput of 9.86 Kbps, and in the rush hour has a delay of 121.54 Ms with a throughput of 9, 56 Kbps. The Reliability and Availability value of the Line-Server-Line network is 100%.