#camera #camera-image #interface #api

cameraunit

An efficient and ergonomic interface to capture images using cameras

9 stable releases (4 major)

5.1.0 May 14, 2024
5.0.0 Apr 4, 2024
4.0.0 Apr 2, 2024
3.1.1 Oct 3, 2023
1.1.0 Sep 30, 2023

#195 in Images


Used in 2 crates

MIT/Apache

21KB
190 lines

cameraunit

cameraunit provides a well-defined and ergonomic API to write interfaces to capture frames from CCD/CMOS based detectors through Rust traits cameraunit::CameraUnit and cameraunit::CameraInfo. The library additionally provides the cameraunit::ImageData struct to obtain images with extensive metadata.

You can use cameraunit to:

  • Write user-friendly interfaces to C APIs to access different kinds of cameras in a uniform fashion,
  • Acquire images from these cameras in different pixel formats (using the image crate as a backend),
  • Save these images to FITS files (requires the cfitsio C library, and uses the fitsio crate) with extensive metadata,
  • Alternatively, use the internal serialimage::DynamicSerialImage object to obtain JPEG, PNG, BMP etc.

Usage

Add this to your Cargo.toml:

[dependencies]
cameraunit = "5.1"

and this to your source code:

use cameraunit::{CameraUnit, CameraInfo, DynamicSerialImage, OptimumExposureBuilder, SerialImageBuffer};

Example

Since this library is mostly trait-only, refer to projects (such as cameraunit_asi) to see it in action.

Notes

The interface provides two traits:

  1. CameraUnit: This trait supports extensive access to the camera, and provides the API for mutating the camera state, such as changing the exposure, region of interest on the detector, etc. The object implementing this trait should not derive from the Clone trait, since ideally image capture should happen in a single thread.
  2. CameraInfo: This trait supports limited access to the camera, and provides the API for obtaining housekeeping data such as temperatures, gain etc., while allowing limited mutation of the camera state, such as changing the detector temperature set point, turning cooler on and off, etc.

Ideally, the crate implementing the camera interface should

  1. Implement the CameraUnit and CameraInfo for a struct that does not allow cloning, and implement a second, smaller structure that allows clone and implement only CameraInfo for that struct.
  2. Provide functions to get the number of available cameras, a form of unique identification for the cameras, and to open a camera using the unique identification. Additionally, a function to open the first available camera may be provided.
  3. Upon opening a camera successfully, a tuple of two objects - one implementing the CameraUnit trait and another implementing the CameraInfo trait, should be returned. The second object should be clonable to be handed off to some threads if required to handle housekeeping functions.

Dependencies

~17MB
~243K SLoC