-
-
Notifications
You must be signed in to change notification settings - Fork 0
Setup
- Python (python.org)
- Esphome (pip install esphome)
Download full repo or git clone
- [optional] change the name in the substitutions section
- [optional] change the timezone
- meter_reader_tflite : Start with a minimal config with no camera_window (zoom) Copy the model you choose from ./model to the config.yaml directory set it's name in model: "dig-class100-0180-s2-q.tflite"
- packages : Select the board by commenting or not the board include
change your wifi passwords and SSID
The number of wifi must be adapt in wifi.yaml
comment out or clear "initial_value" initial_value: ""[]"" This are the crop zone for digit identification
esphome compile config.yaml
If success - connect your board with usb (only needed the first time, the next upload can be done with OTA)
esphome run config.yaml
- Add the new detected board
- Pause the AI processing
- Enabled Flash LED
- You can view and download the image by clicking the camera.

If you have OV2640, OV3660, OV5640, SC101IOT, SC030IOT camera
Go to and test zoom factors http://[hassio]/developer-tools/action > esphome.[s3cam_tflite]_set_camera_window
When your parameters are good, you can set them in config.yaml
meter_reader_tflite:
camera_window: # can be set in http://[hassio]/developer-tools/action > esphome.[s3cam_tflite]_set_camera_window
offset_x: 928 # multiple of 4
offset_y: 480 # multiple of 4
width: 448 # multiple of 4 and 4:3 proportions with height for best quality
height: 88 # multiple of 4 and 4:3 proportions with width for best quality
- Download the image
- Copy it into ./tools folder
run
python draw_regions.py myimage.jpg
Draw you crop zones, press s key to save and q to quit regions.json is generated with your crop zones
You can set them in globals_AI_component.yaml or test them in http://[hassio]/developer-tools/action >esphome.[s3cam_tflite]_set_crop_zones
esphome run config.yaml