Currently the dark calibration is done on a remote computer connected to the AXIOM Beta over Ethernet.
In the simplest case, a predefined number of darkframes are captured and fed into raw2dng for averaging. The resulting combined darkframe is uploaded to the camera and loaded there. Frames with and without darkframe compensation applied are captured and fed into raw2dng for analysis if the compensation worked.
The process has to be repeated for different gains and exposure times, to cover all use cases.
For details, please see Step 2 and 3 from the following page:
This entire process should be done in one script to run inside the camera. The script should:
- guide the user where needed (e.g. attach lens cap)
- print progress information
- summarize the results (a HTML page with figures and graphs would be very nice)
- clean up temporary files.
Optionally, the image processing operations can be accelerated in FPGA, making the calibration process very quick.
Prerequisites: remote camera to test will be provided, development can be done on any linux system
Required programming skills:
- shell scripting (preferred: bash, python)
- C and VHDL for the FPGA task
- easy for a CPU-based solution (but the calibration will be slow)
- medium if FPGA acceleration is used
Mentor: Sebastian, Alex