notkshitij/PiDetectify
notkshitij/PiDetectify
1
0
Fork 0
Code Issues Pull Requests Projects Releases Activity

Completely changed ABOUT.md file. Improved document structure, provided description for each file (main ones at least).

Browse Source
This commit is contained in:
K 2024-11-19 23:45:22 +05:30
parent dba7e67ae1
commit b2b44eb337
Signed by: notkshitij
GPG Key ID: C5B8BC7530F8F43F
1 changed files with 74 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

82
ABOUT.md
View File

@ -1,13 +1,75 @@
# About
- Project uses Python for motion sensing using the camera
- Uses picamera, opencv, numpy libraries for processing the camera feed and motion detection
- Uses requests library for sending notification
- Docker for running 2 containers:
- ntfy for notifications and dashboard (supports PUB/SUB too!)
- wireguard for encrypted tunnel (didn't test since my router has CG-NAT)
# Features
- Scanning the camera feed and processing all done locally, therefore no 3rd party, hence privacy and security preserved
- Docker for isolation of notifications and encrypted tunnel
- Entire setup secured using ufw firewall
This file contains description about this project.
---
## What?
This project utilizes Raspberry Pi with Pi Camera to detect motion in real-time. When motion is detected, the system sends a notification to a specified topic on an ntfy.sh server. The setup involves several components, including a camera interface, Docker setup for ntfy.sh and WireGuard VPN, and a Python script for motion detection.
---
## Features
- **Local Camera Feed Processing**: The entire camera feed is scanned and processed locally on the device. No third-party services are involved, ensuring privacy and security are preserved.
- **Docker-Based Isolation**: The notification system and encrypted tunnel (via WireGuard) are containerized using Docker, providing better isolation, security, and ease of deployment.
- **Enhanced Security with UFW Firewall**: The entire setup is secured using UFW (Uncomplicated Firewall), ensuring that only necessary services are accessible and protecting the system from unauthorized access.
## Structure
1. `setup.sh`: A shell script to install dependencies and set up the environment.
2. `docker-setup.sh`: A shell script that installs Docker and its dependencies.
3. `docker-compose.yaml`: Docker Compose configuration to run ntfy.sh and WireGuard VPN in containers.
4. `main.py`: The Python script that captures camera frames, detects motion, and sends notifications.
---
### setup.sh
- Install `python3`, `pip`, and necessary libraries.
- Install camera libraries such as `picamera2`, `opencv`, and image processing libraries.
- Install `requests` for HTTP notifications.
- Set up Docker and start the `ntfy.sh` container using the [docker-setup.sh](https://git.kska.io/notkshitij/PiDetectify/src/branch/main/docker-setup.sh) script.
## docker-setup.sh
- Remove any conflicting Docker packages.
- Add Docker’s official apt repository and install Docker and Docker Compose.
- Set up the `ntfy.sh` and `wg-easy` containers using the provided [docker-compose.yaml](https://git.kska.io/notkshitij/PiDetectify/src/branch/main/docker-compose.yaml).
- Ensure Docker runs without sudo (add the current user to the Docker group).
- Test the Docker installation by running a "hello-world" container.
> [!NOTE]
> I have referred [official Docker documentation](https://docs.docker.com/engine/install/debian/) for this.
## docker-compose.yaml
- `ntfy`: The container running ntfy.sh for sending notifications.
- Exposes port 80 and uses the `binwiederhier/ntfy` Docker image.
- Configured with timezone IST and persistent storage for cache and config.
`wg-easy`: A WireGuard VPN container that allows secure access to the Raspberry Pi.
- Exposes ports `51820/udp` and `51821/tcp`.
- Uses environment variables for configuration and requires a password hash for login.
> [!TIP]
> Refer official documentation for [ntfy.sh](https://docs.ntfy.sh/install/#docker) and [Wireguard](https://github.com/wg-easy/wg-easy).
## main.py
- Initialize the Raspberry Pi camera using the `picamera2` library.
- Capture and compare two consecutive frames to detect motion using OpenCV’s frame difference method.
- If motion is detected, the script will send a notification to the configured `ntfy.sh` topic (`motion-sensing` in this case by default).
> [!TIP] You can customize the topic name in the `main.py` file by modifying the `ntfy_topic` variable. If you are using the local Docker instance for `ntfy.sh`, the script will use `localhost:80` by default. You can change that too by modifying the `rpi_ip` variable.
## Customization
- **Motion Detection Sensitivity**: The motion detection sensitivity can be adjusted by modifying the area threshold in the `cv2.contourArea(contour)` condition in `main.py`. Decrease the threshold to detect smaller movements or increase it to filter out smaller motion.
- **ntfy.sh Topic**: The topic for notifications is defined in `main.py` as `ntfy_topic`. You can change this to any unique topic name you want to receive notifications under.
---