These versions are all outdated. I strongly recommend you to use the new vesion instead!

Raspberry Pi Zero W security camera system.

Script that sends a video feed from the Raspberry Pi over the network and starts recording when there is motion. The recordings will be saved and removed when the storage gets full. The camera and the recordings can be looked at using a Django website (https://github.com/Ruud14/Django-Camera-View-And-Playback). This feature only works with version 3. Version 3 also has a nice feature where the recording always contains the a couple of frames from before the action happened that triggered the recording.

Schematic of version 3.

Every version has its own advantages and disadvantages but I strongly recommend you use version 2 or 3 since they are faster and are way easier to use. Version 3 isn't always better than version 2 though. It just depends on your needs.

Version	Advantages	Disadvantages
V1	- Requires less dependencies.	- Relatively slow (low fps).
	- Over raw tcp instead of http.	- Stream is more difficult to view.
V2	- Requires less dependencies.	- Requires more processing power on the server.
	- Requires less processing power on the Rb Pi.	- Recording fps depends on internet speed.
V3	- Higher frame rate recordings.	- Requires FFMPEG to be installed.
	- Recording fps isn't affected by internet speed.
	- Also records frames that happened before there the motion got detected.
	Stores the recordings in h264 format for easier playback and less storage occupation.

VERSION 1

In this version the Raspberry pi sends the camera feed to the server frame by frame using raw tcp sockets. The server receives these frames and detects motion in them. When it does detect motion it starts the recording. The main disadvantage here is that it is slow and because sending frame after frame requires a lot of processing power.

How to use:

• Get yourself a Raspberry pi and a (Night Vision IR) camera module.
• Install an os and python3 + openCV.
• Put the Raspberry Pi.py script on your Raspberry pi.
• enable the camera in the sudo raspi-config menu.
• You might want the script start on boot. I achieved this using crontab.
• Make a file called data.json and put it in the same folder as your Raspberry Pi.py.
• In your data.json add {"server_ip":"your_server_ip", "server_port": your_port_number, "secret_key": "your_made_up_key"} and change the values.
• In Raspberry Pi.py in the setup function change /home/pi/Scripts/Camera/data.json to the file location of your data.json.
• Put the CameraReceiver.py and Camera.py on your server/host machine.
• Also add that same data.json file from before to the folder your CameraReceiver.py is in.
• In Camera.py in the Camera class change video_output_folder from .\\Recordings\\ to whatever path you want your video's to be stored in.
• In CameraReceiver.py in the __init__ method change file = open('data.json') to the file location of your data.json.
• You might want to change the camera_count to the amount of camera's you're using. (The value of camera_count can be more but can't be less.)
• If your host machine's os has a GUI you might want to enable cv2.imshow("Frame",frame_data) in the recv_stream method.
• If you want to use this with a django webpage you'd want to get the content of django_views.py and add it to your django views.py. You should also change from .scripts.Cameras_Receiver import CameraReceiver in django_views.py to from PATH_TO_YOUR_CAMERASRECEIVERSCRIPT import CameraReceiver.
• Just to make sure that you don't have to change the ip's in the code every time your router resets, Reserve the ip addresses of both the host and the Raspberry pi in the settings of your router.
• Enjoy

VERSION 2

In this version the Raspberry pi hosts a little http site with the stream of the camera on it. The server then scrapes that site and grabs the video stream. Whenever motion is detected in the video stream the server will start the recording. The main disadvantage here is that the fps of the recording is slow when the internet connection is slow.

How to use:

• Get yourself a Raspberry pi and a (Night Vision IR) camera module.
• Install an os and python3 + openCV.
• Put the Raspberry Pi.py script on your Raspberry pi.
• enable the camera in the sudo raspi-config menu.
• You might want the script start on boot. I achieved this using crontab.
• Make a file called data.json and put it in the same folder as your Raspberry Pi.py.
• In your data.json add {"username":"pi", "password":"picamera"}. You might want to change the values.
• In Raspberry Pi.py change /home/pi/camera/Camera/data.json to the file location of your data.json.
• Put the CameraReceiver.py and Camera.py scripts on your server/host machine.
• You might also want this script to start on boot, again I used crontab.
• Also add that same data.json file from before to the folder your CameraReceiver.py is in.
• In Camera.py in the Camera class change video_output_folder from .\\Recordings\\ to whatever path you want your video's to be stored in.
• In Camera.py in the Camera class you can change motion_sensitivity to change the sensitivity for detecting motion.
• In CameraReceiver.py change data.json in file = open('data.json') to the file location of your data.json.
• In use.py change the ip address to the ip of your camera.
• Just to make sure that you don't have to change the ip's in the code every time your router resets, Reserve the ip addresses of both the host and the Raspberry pi in the settings of your router.
• Enjoy

VERSION 3

In this version the Raspberry pi still hosts a little http site with the stream of the camera on it. It is different from version 2 because here the recording (after movement is detected) takes place on the Raspberry pi itself instead of on the server. The movement is detected on the server side and the server then sends a message to the camera to start the recording. After the recording is done the recording is sent to the server. The main advantages here are that the fps of the recording isn't affected by the internet speed and that it records frames that happen before there the motion was detected. So the action causing the camera to record will always be visible in the recording.

How to use:

• Get yourself a Raspberry pi and a camera module.

On the Raspberry Pi:

• Install an os and python3 + OpenCV + FFMPEG.
• Put everything from inside the camera folder on your Raspberry Pi.
• enable the camera in the sudo raspi-config menu.
• You might want the HTTPServer.py and use.py scripts to start on boot. I achieved this using crontab.
• Make a file called data.json and put it in the same folder as your Camera.py, CameraReceiver.py, HTTPServer.py and use.py (if there isn't one already).
• In your data.json add {"username":"pi", "password":"picamera"}. You might want to change the values.
• In HTTPServer.py change /home/pi/scripts/data.json to the file location of your data.json.
• In CameraReceiver.py change data.json in file = open('data.json') to the same location as in HTTPServer.py.
• In Camera.py in the Camera class change video_output_folder from /home/pi/recordings/ to whatever path you want your video's to be stored in. (The recordings will get removed after sending it to the server.)
• In use.py change the ip address to the ip of your camera/Raspberry pi.
• In Camera.py change the server_ip to the ip address of your server/host.

On the Server:

• Put everything from inside the server folder on your server/host.
• In FileReceiver.py change the host in the main function to the same ip address as server_ip in Camera.py on your Raspberry pi.
• In use.py change the ip-address to the ip of your camera/Raspberry pi.
• Make a file called data.json and put it in the same folder as your detector.py, CameraReceiver.py, FileReceiver.py and use.py (if there isn't one already).
• In your data.json add the same content as in the data.json file of your Raspberry pi.
• In detector.py in the Camera class you can change motion_sensitivity to change the sensitivity for detecting motion.
• Run use.py and FileReceiver.py.
• You might also want the scripts to start on boot, again I used crontab.
• Just to make sure that you don't have to change the ip's in the code every time your router resets, Reserve the ip addresses of both the host and the Raspberry pi in the settings of your router.
• Enjoy

Additional info.

What I learned:

• My first time using a Raspberry pi.
• OpenCV in python.
• Use of struct (un)packing in python.
• Use of crontab in linux.
• Use of base64
• HTTP/MJPG streaming
• FFMPEG

Part of the script running on the raspberry pi in version 2 & 3 is based on this project.

This project assumes that the Raspberry Pi and the server/host are on the same network.

The os used on the raspberry pi is 'Raspbian Buster Lite'. The OpenCV version used on the raspberry pi is '3.4.6' The ffmpeg version used on the raspberry pi is 'git-2020-01-17-c95dfe5' The python version used on the raspberry pi is 'Python 3.7.3'