-
This project is about an online intruder alert system with face recognition on raspberry pi 3
-
Inside the system, there is a motion sensor detecting heat from humans or animals. When the sensor detects a human or animal within a range of 1-5 meters (set up), the system automatically records incoming faces.
-
system will submission system image with The MQTT protocol,image will processed by the Computer Server to determine who is coming in the room.
-
When the process is complete, the person who comes to be identified. The system will send a notification message and the processed image to the user's line.
- Raspberry Pi 3
- Webcam Camera
- Motion Sensor HC-SR501 PIR
How to connect PIR HC-SR501 with Raspberry Pi 3
-
Run file publish.py in raspberry pi
-
When the program is running and connect with MQTT Broker it will print "CONNACK received with code 0" , leave the program open
-
Train the datasets in the Train.py file prepared on the server.
-
Run the Subscribe.py file on your server
-
When the program is connected to MQTT Broker and the Subscribe topic will print "Connected with result code 0" and ''Subscribe with QOS (0,)". leave the program open.
-
When Motion Sensor captures the movement that people have come to. The camera will work. And send image to MQTT Broker
-
When the server receives an image from Rasp Pi, the program will display the topic of the received information.
-
The program will take the image from the rasp pi to process it to detect the person's page and identify it as the data we have trained.
-
When the process is finished,program with create image file is Processed,detected and identified.And send those alerts and pictures to the User via the line.
- Take all the image with have trainer's face. Put them in the input folder within the same project.
- Run the detect_face_datasets.py file to get only face-specific images in the output folder (create a prepared folder).
- Extract all the output files from the output folder into the sort_output folder, divided equally for the Train and Test, and split the files into folders named without the need to rename them.
- Open https://notify-bot.line.me/en/ login with our user and password line
- Get token for sending data to the Line by putting Token into the Python program we want to used for send.
- Give name the sender that will show who sent the message and choose whether to send the message to anyone, only us or as a group.
- Copy the token to keep.
Nuttakrid Uppatumwipanon
Bachelor's degree of electronics and telecommunication engineering
King mongkut's university of technology thonburi
Student id 58070502476
Project in ENE490 Machine Learning