Image Processing

One of my hobbies is working with archival spacecraft imagery to produce pretty pictures. Very few people realize it, but a huge amount of spacecraft data is publically accessible and relatively straightforward to work with with the right tools. Most of this data sits in data archives, referenced now and then for research purposes or as a source for secondary data products like maps. With a little bit of knowledge, these archives can be a way to explore the Solar System from the comfort of your own home.

The problem is the learning curve. When I picked up this hobby in 2013, I was intimidated by the sheer volume of things I needed to learn. How does the command line work? What do these obscure file formats mean? What are the best practices for making pretty images that are also realistic? I spent much of my first year feeling constantly overwhelmed by the stuff I needed to know, and was frustrated that the information was available, but scattered all over the web. This page is intended to help you minimize the hours searching for answers by providing tutorials and helpful links in one place.

Getting Started

Your first stop should be to Emily Lakdawalla's free Basics of Digital Imaging course at the Planetary Society. It's a great intro that covers most of the basics, and it's a resource that I wish I had been aware of when I first started out.

Another good resource is the UnmannedSpaceflight Forums, started by Doug Ellison in 2004 (then a hobbyist, now a navigation engineer on Curiosity). It has plenty of up-to-date discussion from enthusiasts following along with current missions. It also has a section for questions about image processing and sharing techniques that the forum users have developed.

I keep a list of my tools in rotation here.. Most of it is freeware, with the exception of Photoshop CC (which I am slowly trying to move away from). Most of the tutorials on this site are written using Photoshop, but use features that are standard in most photo editing suites. Don't feel like you have to use these tools to get good results - if you know your way around software that performs the same function, in most cases time is better spent adapting the process to that program instead of learning something entirely new.

Techniques, Recipes, and Projects

Image processing generally relies on a handful of core techniques. These techniques are usually very simple modifications or manipulations of the data, but can be used in sequence to pull out additional details hidden in the data. Some of these techniques blur the line between amateur and professional - some of these techinques are adopted from the scientific literature; others are skills I've learned in the hobby and have transferred to my professional work.

Usually, I don't have a plan or an end-goal in mind when I choose to work on a specific imaging data set. I see an imaging data set and decide it will make for a pretty picture. However, from time to time I do take deeper dives with a specific purpose in mind, like testing newly reported discoveries with archival data. I will add write-ups of these projects here.

Words of Advice

Don't be intimidated by your current skill level. Many of the pictures I started on were ones that I had never seen before despite knowing a lot of the processing community. I worried if I shared the results of a bad processing job, it would become the "definitive" version of the picture. What helped was the realization that the original data would always be there, and that if someone else (including Future Me) felt like they could do the job better, they were always free to do so.

And some advice: there is always room to grow and get better. I took up this hobby in 2014, but I'm still learning or figuring out new and better ways to do things. My work is often full of flaws that only I can see - but looking for these flaws and figuring out how to fix them next time is part of the process. Always be on the lookout for things you could have done better. And finally unlike Apollo, failure is an option. Sometimes working with a data set doesn't work out. I have dozens of failed projects, many of which still sit in the back of my mind. As your working knowledge grows and matures, keep these in your back pocket - you might develop skills that make these projects possible in the future.