doorcam

Raspberry Pi Peephole Camera

Requirements

This application requires OpenCV and the optional python package(s) it installs. You'll need to build and/or install that before you proceed with installing doorcam.

Installation

1. Install the requirements listed above
2. Clone the repo

   git clone https://github.com/retrontology/doorcam
   

3. Install the python requirements

   cd doorcam
   python3 -m pip install -r requirements.txt
   

Config

• analyzer:
  • contour_minimum_area: Minimum contour area of difference between frames of the analyzer to trigger a detection event.
  • delta_threshold: Threshold setting passed to threshold command for detecting difference between frames of tha analyzer
  • max_fps: Maximum desired fps. Minium fps relies on speed of single thread
  • undistort: Whether you want to undistort the camera image before analayzing it or not.
  • undistort_balance: The balance used for the undistort function if enabled
• camera:
  • D: Array of distortion coeffecients for applying fisheye undistortion. Obtained via the calibrate.py program.
  • K: Camera intrinsic matrix. Obtained via the calibrate.py program.
  • format: A four letter string used for setting the format of the capture device.
  • index: Index of the video device to be used for capture. i.e. if you want to use /dev/video2, your index would be 2
  • max_fps: Desired capture fps for the video device
  • resolution: Desired capture resolution for the video device
  • rotation: Rotation desired for frames retrieved from the video device. Is very intensive and can reduce fps if not None/null
• capture:
  • enable: Whether or not to enable saving events to disk
  • keep_images Whether or not to keep saved images
  • path: Where the images will be saved
  • postroll: Amount of time in seconds to capture after the last frame where motion is detected
  • preroll: Amount of time in seconds to capture before the first frame where motion is detected
  • rotation: The desired rotation to apply to the frames during post-processing.
  • timestamp: Whether or not to add timestamps to the saved images
  • trim_old: Whether or not to trim/delete old events.
  • trim_limit: Amount of days of events you want to keep. All videos older than this window are trimmed/deleted if trim_old is true.
  • video_encode: Whether or not to encode the saved images to a video file
• screen:
  • activation_period: How long in seconds you want the screen to activate for when either motion is detected or you touch the screen.
  • backlight_device: Path to the backlight device
  • color_conv: Color conversion to use for rendering to the framebuffer. Refer to https://docs.opencv.org/4.5.3/d8/d01/group__imgproc__color__conversions.html
  • dtype: The dtype to use for determining the width of each framebuffer pixel. Refer to https://numpy.org/doc/stable/reference/arrays.scalars.html#sized-aliases
  • framebuffer_device: Path to the framebuffer device to use for display.
  • resolution: The resolution of the framebuffer for resizing the frame for display.
  • rotation: The desired rotation to apply to the frame retrieved from the camera. Is significantly faster in this application as the image used is 1/4 the size of the original
  • touch_device: Path to the touchscreen device
  • undistort: Whether to undistort the frame on the screen
  • undistort_balance: The balance to pass to the undistortion function
• stream:
  • ip: The IP address of the desired network device to use for the MJPG server
  • port: The port to listen on for the MJPG server

Usage

usage: run.py [-h] [-c config.yaml] [-d] [-f]

optional arguments:
  -h, --help            show this help message and exit
  -c config.yaml, --config config.yaml
  -d, --debug
  -f, --fps