Topic ──

Embedded CCTV recording and Face ID authentication management system using real-time object recognition using C and openCV

Summary ──

Using Raspberry Pi, C/C++, and openCV, CCTV that detects and records movement in places that need to be monitored

If the administrator needs to be away, you can turn on the monitor mode by pressing the button. When the monitoring mode is turned on, the LED flashes three times and detects movement on the screen in real time from 10 seconds later. (Use MSE.) If movement is detected, it immediately starts recording with a buzzer. The image is stored by performing recording until the movement is not detected for more than 10 seconds.

At this time, pressing the button ends recording and enters the monitoring mode. Even if it is not recording, when the manager returns to the seat, the monitoring mode may also be turned off by pressing the button. In this case, face recognition is required to turn off the monitoring mode. (Use a lower cascade.) It is not recognized if you wear a mask or if you can't see your face properly. This is to leave a picture of the face of a person who has ended the surveillance mode. If face recognition is successful, it saves a picture of the recognized face and exits surveillance mode. Facial authentication is performed up to five times once called.

Background ──

There are places where CCTVs should always be turned, but there are moments when CCTVs are unnecessary in places where people are guarding. CCTV takes up more capacity than necessary to record every moment. In addition, CCTV data that has not been used for more than three months should be erased to cover the capacity. It's okay if the video isn't used, but there's no way to turn back the deleted video if the moment comes. If only the necessary parts were stored and managed, the project was planned as it was judged that storage management could be facilitated. Thinking that "Necessary Parts" is when a person is away and cannot detect intruders, we created a program that sounds the buzzer and starts recording images when we detect movement (using MSE) In addition, a code that can terminate the Face ID when it matches was written so that the monitor could return and terminate the CCTV, and the face of the terminator was recorded in case someone else tried.

Expectation ──

In the case of this program, it is designed to monitor intruders in certain spaces that must be monitored if an administrator is away. It can be used in private garages, in front of the store's safe, inside the management office, and at the entrance of facilities with little traffic.

Anaysis ──

Details

System Flow

Baseline

1) Flow chart

flowchart.jpg

2) Process

  1. program execution : Load the device driver.