Investigate compression techniques

[wtbarnes]

(this text is paraphrased from some conversations on Teams)

Investigate whether JPEG-LS lossless compression is a viable option as a compression technique for quickly reading off images from the camera. This is partly because we may not be able to readout at 1 HZ (or greater) without compression.

In order to test this, we'll need a good model for the visible-light leakage, both the steady and shot noise components. The noiseless simulations will likely compress too well and not give a realistic estimate of the compression. The images we test on should be at the 1 s cadence

Some helpful links

• https://en.wikipedia.org/wiki/Lossless_JPEG#JPEG-LS
• http://www.stat.columbia.edu/~jakulin/jpeg-ls/mirror.htm

[wtbarnes]

This will levy a processing requirement on the IDPU

Do we have to uncompress onboard as well?

Potential workflow: compress on the camera, uncompress on the IDPU in order to co-add.

[wtbarnes]

Nominal performance of the spacewire is 24 Mbit/s but we have a 16 bit detector so in order to meet that, we need compression that gives us at least a factor of 2.