bird-feeder

This is a super simple bird feeder project that lets you use either a weight sensor, motion detection, or both to automatically snap pictures of birds on your bird feeder and optionally upload them to the cloud.

Basic Setup

1. Clone repo

git clone https://github.com/hevansDev/bird-feeder.git
cd bird-feeder/bird-feeder

2. Create virtual environment

python3 -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Configure

cp .env.example .env
nano .env  # Edit with your settings
MOTION_ENABLED=true
MOTION_THRESHOLD=1000  # Adjust based on your environment

# Optional: Upload to community gallery
ENABLE_CLOUD_UPLOAD=true
USER_ID=your_username
FEEDER_LOCATION=Your City, Country

5. Run

python main.py

Setting up the scale

For a full guide on setting up and calibrating the directly attached scale see my blog on the topic, blog on setting up Pico scale coming soon.

Settings

Motion Detection

Variable	Default	Description
MOTION_ENABLED	true	Enable motion-based detection
MOTION_THRESHOLD	1000	Pixel change threshold (higher = less sensitive)
FRAMES_BEFORE_DEPARTURE	10	Consecutive no-motion frames before "bird left"
DEBUG_MOTION	false	Print motion values for threshold tuning

Scale Settings

Variable	Default	Description
SCALE_ENABLED	false	Enable weight-based detection
SCALE_TYPE	serial	direct (Pi GPIO) or serial (Pico USB)
WEIGHT_THRESHOLD	5	Minimum weight in grams (Goldcrest = 5g)
SCALE_WAIT_TIME	1.0	Seconds to wait for scale reading after motion

Direct GPIO Scale:

Variable	Default	Description
SCALE_REFERENCE_UNIT	-388.929792	Calibration value from calibration.py

Pico Serial Scale:

Variable	Default	Description
PICO_SERIAL_PORT	/dev/ttyACM0	USB serial port for Pico
PICO_SERIAL_BAUD	115200	Serial baud rate
PICO_TIMEOUT	2.0	Serial connection timeout

Cloud Upload

Variable	Default	Description
ENABLE_CLOUD_UPLOAD	false	Upload photos to community gallery
UPLOAD_SERVICE_URL	https://...	Cloudflare Worker endpoint (pre-configured)
USER_ID	anonymous	Your unique username for the gallery
FEEDER_LOCATION	(empty)	City, Country
KAFKA_BROKER_URL	kafka-2a015ed7-bird-feeder-free-tier.d.aivencloud.com:19448	Kafka broker URL for streaming bird data

Privacy: Only photos, timestamps, weights, and location (if provided) are shared. No personal information is collected.

Camera & Performance

Variable	Default	Description
CAMERA_WARMUP_FRAMES	5	Frames to skip for auto-adjustment
PHOTO_COOLDOWN	5.0	Minimum seconds between photos
IMAGES_DIR	./images	Local photo storage directory
SENSOR_CHECK_INTERVAL	0.2	Seconds between sensor readings

Resources

• The motion detection code in this project is based on this blog by sokacoding.
• HX711 library: tatobari/hx711py
• Community uploads powered by Cloudflare Images

Bird Feeder Upload API

You don't need to deploy this yourself! The community worker is already running at https://bird-upload-api.hugh-evans-dev.workers.dev but if you want your own private gallery you can deploy your own by following the instructions in the cloudflare-worker dir.