Camera Data Extractor for Home Assistant

Extract numeric values from IP camera video streams using OCR (Optical Character Recognition). Perfect for monitoring displays like boiler temperatures, pressure gauges, energy meters, or any device with a digital/analog display.

Features

• Video Stream Support: Connect to RTSP and HTTP video streams from IP cameras
• OCR Value Extraction: Automatically extract numeric values from camera images
• Region of Interest (ROI): Define specific areas of the frame to analyze
• Template Matching: Handle camera movement/rotation by using reference templates
• Auto-Discovery: Discover ONVIF-compatible cameras on your network
• Web UI Panel: Easy-to-use interface for configuration and ROI selection
• Home Assistant Integration: Exposes values as sensor entities for automations and monitoring
• Two Installation Methods: Available as both a Custom Integration and an Add-on

Screenshots

Main Dashboard

View all configured cameras and their extracted values at a glance.

ROI Selection

Easily select the region of interest by clicking and dragging on the camera frame.

Camera Configuration

Configure camera settings, preprocessing options, and scan intervals.

Installation

Choose one of the following installation methods based on your needs:

Method	Best For	Requires
Add-on	Home Assistant OS / Supervised	Nothing extra
Custom Integration	Any HA installation	Tesseract OCR on host

---

Method 1: Add-on Repository (Recommended for HA OS)

This is the easiest method for Home Assistant OS and Supervised installations. The add-on runs in its own container with all dependencies included.

Quick Install

Click the button above, or follow the manual steps below:

Manual Install

1. Go to Settings → Add-ons → Add-on Store
2. Click the three dots menu (⋮) in the top right → Repositories
3. Add this repository URL:

   https://github.com/tursoft/ha-camera-data-extractor
   

4. Click Add → Close
5. Find "Camera Data Extractor" in the add-on store and click it
6. Click Install
7. Configure your cameras in the Configuration tab
8. Start the add-on
9. Click Open Web UI to access the interface

Add-on Configuration Example

cameras:
  - name: Boiler Temperature
    stream_url: rtsp://192.168.1.100:554/stream1
    username: admin
    password: your_password
    value_name: Temperature
    unit: °C
    roi_x: 100
    roi_y: 50
    roi_width: 200
    roi_height: 80
    preprocessing: auto
  - name: Pressure Gauge
    stream_url: http://192.168.1.101/video
    value_name: Pressure
    unit: bar
scan_interval: 30
log_level: info

---

Method 2: HACS Custom Integration

For advanced users or non-OS installations who want sensor entities.

Quick Install

Click the button above, or follow the manual steps below:

Manual Install

1. Open HACS in Home Assistant
2. Click the three dots menu (⋮) → Custom repositories
3. Add https://github.com/tursoft/ha-camera-data-extractor as an Integration
4. Search for "Camera Data Extractor" and install
5. Restart Home Assistant
6. Go to Settings → Devices & Services → Add Integration
7. Search for "Camera Data Extractor" and configure

Method 3: Manual Installation

For manual installation without HACS.

1. Download the latest release from GitHub
2. Extract and copy the custom_components/camera_data_extractor folder to your Home Assistant's custom_components directory
3. Restart Home Assistant

---

Prerequisites (Custom Integration Only)

Note: If you installed the Add-on, skip this section - all dependencies are included.

The custom integration requires Tesseract OCR to be installed on your Home Assistant system.

Home Assistant OS / Supervised

Tesseract should be available automatically. If not, you may need to install it via the terminal add-on:

apk add tesseract-ocr

Home Assistant Container

Add Tesseract to your Docker image or mount it from the host:

apt-get install tesseract-ocr

Home Assistant Core

Install Tesseract on your system:

# Ubuntu/Debian
sudo apt-get install tesseract-ocr

# Fedora
sudo dnf install tesseract

# macOS
brew install tesseract

Configuration

Adding a Camera

1. Go to Settings → Devices & Services → Add Integration

2. Search for "Camera Data Extractor"

3. Enter your camera details:

   • Camera Name: A friendly name for your camera
   • Stream URL: The RTSP or HTTP URL of your camera stream
     • RTSP example: rtsp://192.168.1.100:554/stream1
     • HTTP example: http://192.168.1.100/video
   • Username/Password: Optional credentials for camera authentication

4. Configure extraction settings:

   • Value Name: Name for the extracted value (e.g., "Temperature")
   • Unit of Measurement: Unit for display (e.g., "°C", "bar", "%")
   • Scan Interval: How often to capture and analyze frames (seconds)
   • Preprocessing: Image preprocessing method for better OCR accuracy

5. Set up the Region of Interest (ROI):

   • Set coordinates to define the area containing the value
   • Use the Web UI panel for visual selection (recommended)

Using the Web UI Panel

After installation, a new Camera Data Extractor panel appears in your Home Assistant sidebar.

1. Cameras Tab: View configured cameras and capture live frames
2. ROI Selection Tab:
   • Click and drag on the frame to select the region containing the value
   • Fine-tune using the coordinate inputs
   • Save as a template for cameras that may move
3. Templates Tab: Manage saved templates and test template matching
4. Discovery Tab: Scan your network for ONVIF cameras

Template Matching (For Moving Cameras)

If your camera can pan/tilt/zoom, use template matching to find the correct region even when the camera moves:

1. Position your camera to show the display
2. Go to the ROI Selection tab
3. Select the region containing the value
4. Enter a template name and click "Save as Template"
5. Use the camera_data_extractor.use_template service to update the ROI based on the current camera position

Common Camera Stream URLs

Brand	Typical RTSP URL
TP-Link Tapo	rtsp://user:pass@IP:554/stream1 (HD) or stream2 (SD)
Hikvision	rtsp://user:pass@IP:554/Streaming/Channels/101
Dahua	rtsp://user:pass@IP:554/cam/realmonitor?channel=1&subtype=0
Reolink	rtsp://user:pass@IP:554/h264Preview_01_main
Amcrest	rtsp://user:pass@IP:554/cam/realmonitor?channel=1&subtype=0
Ubiquiti	rtsp://user:pass@IP:554/live
Generic ONVIF	rtsp://user:pass@IP:554/stream1

Note for Tapo cameras: You must create a "Camera Account" in the Tapo app (Settings → Advanced Settings → Camera Account). Use these credentials for RTSP access, not your TP-Link account credentials.

Services

camera_data_extractor.capture_frame

Immediately capture a frame and update the sensor value.

service: camera_data_extractor.capture_frame
data:
  entity_id: sensor.boiler_temperature

camera_data_extractor.update_roi

Update the Region of Interest coordinates.

service: camera_data_extractor.update_roi
data:
  entity_id: sensor.boiler_temperature
  roi_x: 100
  roi_y: 50
  roi_width: 200
  roi_height: 80

camera_data_extractor.use_template

Find and apply ROI from a saved template (for moving cameras).

service: camera_data_extractor.use_template
data:
  entity_id: sensor.boiler_temperature
  template_name: boiler_display

Sensor Attributes

Each sensor provides these attributes:

Attribute	Description
raw_text	Raw text extracted by OCR
confidence	OCR confidence percentage
last_frame_time	Timestamp of last analyzed frame
camera_name	Name of the camera
roi_x, roi_y, roi_width, roi_height	Current ROI coordinates
preprocessing	Preprocessing method used
stream_url	Camera stream URL

Automation Examples

Alert on High Temperature

automation:
  - alias: "Boiler Temperature Alert"
    trigger:
      - platform: numeric_state
        entity_id: sensor.boiler_temperature
        above: 80
    action:
      - service: notify.mobile_app
        data:
          title: "Boiler Alert"
          message: "Temperature is {{ states('sensor.boiler_temperature') }}°C!"

Log Temperature Changes

automation:
  - alias: "Log Boiler Temperature"
    trigger:
      - platform: state
        entity_id: sensor.boiler_temperature
    action:
      - service: logbook.log
        data:
          name: Boiler
          message: "Temperature changed to {{ states('sensor.boiler_temperature') }}°C"

Update ROI for PTZ Camera

automation:
  - alias: "Update Boiler Camera ROI"
    trigger:
      - platform: time_pattern
        minutes: "/5"
    action:
      - service: camera_data_extractor.use_template
        data:
          entity_id: sensor.boiler_temperature
          template_name: boiler_display

Preprocessing Options

Option	Description	Best For
auto	Automatically determines best method	General use
none	No preprocessing	Clear, high-contrast displays
threshold	Binary threshold (Otsu's method)	LCD/LED displays
adaptive	Adaptive threshold	Varying lighting conditions
invert	Inverted colors	Light text on dark background

Troubleshooting

OCR Not Reading Values Correctly

1. Adjust ROI: Make sure the ROI tightly frames only the numeric display
2. Try Different Preprocessing: Switch between auto, threshold, and adaptive
3. Improve Lighting: Ensure adequate lighting on the display
4. Increase Resolution: Use a higher resolution stream if available
5. Check Decimal Places: Adjust the decimal places setting

Camera Connection Issues

1. Verify Stream URL: Test the URL in VLC or another media player
2. Check Credentials: Ensure username/password are correct
3. Firewall Rules: Make sure Home Assistant can reach the camera
4. RTSP Port: Default is 554, but some cameras use different ports

Template Matching Not Working

1. Retake Template: Capture a new template in good lighting
2. Check Confidence: If confidence is low, the camera view may have changed too much
3. Reduce Movement: Template matching works best with small position changes

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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

Support

• Report an Issue
• Home Assistant Community Forum

Author

Muhammet Turşak Email: tursoft@gmail.com Website: https://tursoft.net