Catflap Prey Detector

Table of Contents

• Overview
• Gallery
  • System in Action
  • Hardware Setup
• How to get started
• Hardware Requirements
• Software Stack
• Installation & Setup
  • Prerequisites
  • Installation
  • Configuration
  • Run
• Usage
• Documentation
  • Module Documentation
  • Prey Detection API
• Contributing
• License
• Acknowledgments

AI-powered system that detects when your cat is carrying prey and automatically blocks the catflap to prevent unwanted "gifts" from entering your home.

Overview

Problem: Cats bringing prey (mice, birds, etc.) into your home through the catflap.

Solution: Camera captures frames → YOLO tracks cats and other objects (persons, etc.) → Prey Detection API analyzes images → RFID jammer blocks catflap when prey is detected. The system monitors and sends notifications for all detected object classes.

Gallery

System in Action

Day detection with Telegram notification	Night detection with infrared camera
Cat detection notification example	Person detection at night with notification

Hardware Setup

Assembled catflap with camera and antenna

Required hardware components

Key Features:

• 🎥 24/7 monitoring with night vision camera
• 🤖 AI-powered prey detection using YOLO11n and custom Prey Detection API
• 🚫 Automatic catflap blocking via RFID jamming
• 📱 Remote control and notifications via Telegram
• 👁️ Monitors and notifies for multiple object classes (cats, persons, etc.)
• ☁️ Optional long-term image persistence to Google Cloud Storage

How to get started

New to this project? Follow this step-by-step path:

1️⃣ Understand the System

• Read this README
• Check out the Architecture Overview to understand how components work together. Check the Module Documentation for details.

2️⃣ Gather Hardware

• Review Hardware Requirements below
• Order components from Hardware Setup Guide

3️⃣ Assemble Hardware

• Follow the Hardware Setup Guide for:
  • Raspberry Pi initial setup
  • Camera installation and mounting
  • RFID jammer wiring and assembly
  • Testing procedures

4️⃣ Get API Access

• Read the Prey Detection API documentation

5️⃣ Configure Software

• Set up Telegram bot (get token from @BotFather) - see Notifications Guide
• Configure environment variables following Configuration Guide
• Optional: Set up Google Cloud Storage for image backup

6️⃣ Install & Run

• Follow Installation & Setup below
• Run the system with make run
• Test with Telegram commands: /photo, /status, /ping

7️⃣ Deploy & Monitor

• Set up remote access with Deployment Tools
• Configure as a systemd service for auto-start
• Monitor logs and performance

Hardware Requirements

• Raspberry Pi 5 - Main processing unit (4GB or 8GB RAM)*
• Cooling system - Active cooling fan or heatsink for Raspberry Pi
• Picamera 3 IR Wide - Night vision camera
• Infrared illumination - For 24/7 monitoring
• 134.2 kHz FDX-B RFID reader module - For jamming signal
• 5V single-channel relay - Controls RFID jammer via GPIO
• Jumper wires - For connections
• Compatible catflap with RFID reader - Target device to block

*Note: The system only requires ~1GB RAM, so lower-spec Pi models may work. CPU affects YOLO FPS but shouldn't be an issue.

See Hardware Setup Guide for detailed assembly instructions.

Software Stack

• Python 3.11+
• YOLO11n (NCNN) - Fast object detection for tracking cats and other objects
• Prey Detection API - AI service for prey detection
• Telegram Bot API - Remote notifications and control
• Google Cloud Storage (optional) - Long-term image persistence

Installation & Setup

Prerequisites

• Python 3.11 or higher
• uv package manager

Installation

git clone https://github.com/fl2o/catflap-prey-detector.git
cd catflap-prey-detector
make install

Important for Raspberry Pi: The installation uses uv venv --system-site-packages to ensure compatibility with the Raspberry Pi camera libraries (picamera2) that are pre-installed on the system.

Configuration

Set the following environment variables (see Configuration Guide for details):

Required:

• BOT_TOKEN - Telegram bot token (get from @BotFather)
• GROUP_ID - Telegram group/chat ID (use /where command)
• PREY_DETECTOR_API_KEY - API authentication key

Optional (for long-term image persistence):

• GCS_BUCKET_NAME - Google Cloud Storage bucket name
• GOOGLE_APPLICATION_CREDENTIALS - Path to GCS service account JSON

Run

make run

(corresponds to uv run catflap-detector)

Usage

The system automatically:

1. Detects cats and other objects (persons, etc.) approaching the catflap
2. Analyzes cat images for prey detection
3. Locks the catflap when prey is detected
4. Sends Telegram notifications with images for all detected events
5. Auto-unlocks after configured duration (default: 5 minutes)

Manual Controls (via Telegram bot):

• /lock - Manually lock the catflap
• /unlock - Manually unlock the catflap
• /status - Check current lock status and remaining time
• /photo - Capture and receive current camera view
• /ping - Health check (responds with "pong")

Documentation

• Hardware Setup - Assembly and wiring instructions
• Configuration - Configuration options and tuning
• Prey Detection API - API documentation and usage
• Architecture - System design overview
• Deployment Tools - SSH, Tailscale, Syncthing, WiFi management

Module Documentation

Each module has detailed documentation:

• Detection Module - YOLO detection and tracking
• Classification Module - Prey detection API integration
• Hardware Module - RFID jammer and catflap control
• Notifications Module - Telegram bot
• Cloud Module - GCS sync

Prey Detection API

The Prey Detection API is a custom-built AI service that powers the prey detection capabilities of this system. I built this model.

Get API Access:

• Purchase an API key at: https://buy.stripe.com/dRmaEPasU1pm3Ua6o27kc00
• See the Prey Detection API documentation for setup and usage details

Custom Models: If you have specific use cases or monitoring needs beyond prey detection, I can build custom models tailored to your requirements. Feel free to reach out to discuss your use case.

Self-Hosted Alternative: You can also modify the code to use your own prey detection model instead of the API. The system is designed to be flexible - simply replace the detect_prey() function in src/catflap_prey_detector/classification/prey_detector_api/detector.py with your own model inference.

Future: A guide for training your own custom model with data collected from the Prey Detection API will be added in the coming months.

Contributing

Contributions are welcome! Please feel free to submit issues and pull requests.

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Acknowledgments

• YOLO11 by Ultralytics
• NCNN framework by Tencent