It's all about (Image) Quality!
One of the top things on our mind right now is image quality and figuring out how to actually measure and determine what is a "good image quality"?
The reason for this is of course our mission to the moon. For this we are building a self developed camera system, which mades up the camera head that goes on-top of our lunar rover Asimov. The address the most obvious question right aware, why the hell are we developing a whole new camera system? Aren't there many good ones already?
The answer is quite simple: Yes, of course there are existing solutions out there! The simple reason we went out to design one ourselves is because we thought we could build a camera that really fits our mission needs and also provides a value to future explorers. When it comes to cameras and the Google Lunar XPRIZE it's very important to read the fine print of the competition guidelines. In the so called "Master Team Agreement", basically the rules of the competition, there is a very specific definition for how good images from the moon should look like. We all know about the basic goal of capturing nice HD footage from the moon and send it back to earth. But did you know that for the panoramic image alone you would need to capture more than 21.000pixels in the horizontal direction (360°) and achieve a contrast of at least 40% between neighboring pixels? The reason for these nit picking tiny details in the rules is that there are really many way to capture "HD" footage. You have probably also noticed how poorly some so called HD camera perform, especially in low light conditions. Now think about such a camera on the moon. The lunar regolith acts partially as a reflector, the sun is extremely bright and the sky is almost pitch black. Long sharp shadows are crouching all over the scene. As it turns out, operating in such light conditions (not taking into account the thermal and radiation requirements!) is really hard for a digital camera. Your normal smartphone would probably throw in the towel and leave you with images that you would not look forward to share with the world.
Really? You sent a rover to the moon and all you got are these lousy images?!
The three key requirements we identified for our camera system for the moon where, being radiation tolerant, have a small form factor with low power requirements and of course satisfy the image requirements. Ok, we cheated a bit these where actually four important requirements for our camera. However, to these vital requirements we determined that it would be nice to also be capable of learning more about the lunar surface. Thus we added a special set of 9 exchangeable filters (e.g., infrared, colors,..) to the camera. Just like on the Mars Curiosity rover. This enables us on the ground to see what would otherwise be hidden from us.
Ok, so let's suppose you acquired or build a camera that fit's your personal and the Google Lunar XPRIZE MTA image requirements, how do you proof it works according to the specs? The issue with image quality is that its subject to your opinion if its good or bad, thus the first step is to figure out some quantifiable metrics and then define a test setup which you can replicate over and over again. A very good basis is to have a selection of test charts available. These are charts with pre defined lines and drawings in varying sizes that help you fine tune your precious optics, the focus unit and even evaluate the quality of the overall imaging unit. However, since early on we noticed a problem with many of the test charts we had tested. A number of them where simply purchased online, e.g., at Amazon and eBay, and other where printed on laser office printers. After many hours of testing we always ended up with the same frustrating result that we could proof that our image quality perfectly matches up with the print quality of the test charts. Normally this is good news, as it means "test passed" at least for this chart, however it also got us to wonder if the image quality of our camera was actually being limited by the printout quality of the test charts. After conducting several tests with a number of good and bad printers we could confirm our suspicion, thus we needed to find a new way to create our test charts.
Luckily we've found a pretty old school technique to save the day. Back in the 90s cheap repro film where the best way to make high quality PCBs at home. Today of course making PCBs became a solely digital process, however the cheap repro films and shops who offer to make them still do exist! While in the past few years the need for such films subsided we where really excited to see that there are still some companies out there who make these. By the way, these films are not really printed but rather "copied" or "exposed" onto the film itself, like with an analog photo. This provided us with de facto image quality for our test charts of up to 2540dpi with real fine lines and a black that is truly black. This repro film technique, intended for a totally different purpose provided us with extremely good high quality test charts, that finally for the first time, made us hit the real limits of our camera solutions! :-)
Another awesome side effect of self printed test charts is of course the custom branding: Hell Yeah! ;-)
Long story short, to comply with the fine print regulations of the Google Lunar XPRIZE competition we have developed a new lightweight, low-power and high quality camera system that also future space mission can benefit from. Even if it would have not been for this Google Lunar XPRIZE fine print rules about image quality we would have likely also ended up developing a new solution in-house as we like to fine tune our mission to provide as much value to all of you as possible. To proof and test the image quality of our camera we needed to find and use an old school printing technique and self designed testcharts.
Robert from the Part-Time Scientists