Last updated: 2026-11-11

Current highlight: see how Raspberry Pi Zero 2 W Door Camera Display combines focused engineering with dependable day-to-day operation.

Current working deployment

Raspberry Pi Zero 2 W Door Camera Display

This project turns a Raspberry Pi Zero 2 W into a dedicated indoor door camera display. Frigate remains responsible for camera recording and person detection. The Pi acts only as a responsive viewer, showing the same detection image on a compact SPI LCD and a ViewSonic HDMI monitor when somebody approaches the door.

Frigate person detection displayed by the Raspberry Pi Zero 2 W door camera project

Example Frigate person-detection frame used by the LCD and HDMI viewers.

Detection Gallery

These event frames show Frigate detections captured during normal operation. They demonstrate different subject positions, movement, and framing handled by the Raspberry Pi display services.

Two people detected near the monitored doorway — Two-person doorway detection with both subjects tracked in the event frame.

Person walking away through the monitored doorway — Full-body tracking as a person moves away through the monitored area.

Rear view of a person detected near the doorway — Rear-view detection demonstrating tracking across changing subject orientation.

System Flow

Door CameraVideo source

Frigate NVRRecording and detection

MQTT EventPerson detected

Pi Zero 2 WLatest-frame viewers

LCD + HDMISynchronized display

SPI LCD Viewer

320×240 framebuffer display
Dedicated LCD service
Limited to 2 FPS to preserve HDMI performance
Clears automatically after the event timeout
Existing LCD behavior remains independent from HDMI

ViewSonic HDMI Viewer

1920×1080 HDMI framebuffer
960×540 image rendered in the top-left quarter
Persistent Frigate MJPEG connection requesting 12 FPS
Measured rendering of approximately 10.7 to 11.26 FPS
Latest-frame-only processing drops stale frames

Event-Driven Operation

Both viewer services subscribe to Frigate person-detection events through MQTT. A matching event wakes the display path and starts rendering current camera frames. The HDMI viewer uses a persistent MJPEG connection rather than repeated snapshot requests, improving smoothness while avoiding a queue of outdated frames.

After 60 seconds without a matching event, the displays are cleared. The ViewSonic monitor is then placed into DPMS sleep and wakes again for the next detection.

Software Design

Two independent Python services, one for LCD and one for HDMI
Frigate supplies detection events and scaled camera frames
NumPy-based pixel conversion for efficient framebuffer output
Automatic MJPEG reconnection after stream errors
Deadline-based pacing without queued frames
Systemd services start automatically and recover independently

Design Goals

Keep Frigate as the single source of truth for recording and detection
Keep the display device simple, local, and purpose-built
Show the door image only when it is useful
Keep the LCD and HDMI outputs isolated so one display cannot break the other
Minimize CPU, network, and power use while idle

Current Status

The LCD and ViewSonic HDMI services are active and have completed controlled wake, render, clear, and sleep testing. The current deployment includes persistent MJPEG streaming and independent display services for stable operation.