Many of our projects are muti-faceted. The MAPLE project, being led by PDF Christina Smith, is no exception. Primarily tasked to couple a CW laser with an all-sky camera, this work also includes side investigations, such as the exploration of HDR imaging led by PVL MSc Trainee Alex Innanen.
by Alex Innanen
For the past… perhaps longer than I would like to admit, I have been working on building a setup for an HDR imaging system using a DMD. That’s a lot of acronyms right off the bat so let’s break that down.
HDR stands for High Dynamic Range – that is, in imaging, achieving a greater range of luminosity values. Have you ever tried to take a picture with a mix of bright light and shade? The light tends to overpower the shaded regions, and you can’t see details within them. Take this picture of my office I took with my phone: you can’t see anything inside the office because of the sheer brightness of the window. Conversely, if I try to resolve the inside of the office, the scene through the window gets washed out. HDR imaging (HDRI) techniques can help us see both!
As for a DMD, or Digital Micromirror Device, this is an array of very very tiny mirrors – around 13 microns across – which can be switched “on” or “off” individually. This allows us to direct light, reflected off the micromirrors, either towards or away from some observer. This is also fun because you can switch the arrangement of ‘on’ and ‘off’ mirrors to make patterns. It’s almost Valentine’s day, so here’s a heart!
So how does a DMD help us increase dynamic range? Imagine I’m taking a picture of some very lovely clouds, but the sun is right in the frame completely washing out my clouds. If you’ve ever been outside on a sunny day, you might have shaded your eyes against the sun, or held out a hand to block it out. Physically blocking the sun – either with a mechanical apparatus or with a kind of ‘mask’ right on the lens – has been used pretty extensively for both earth- and space-based applications. The Hubble’s Advance Camera for Surveys (ACS), for example, includes such a sun block – called a coronagraph – which allowed it to directly image the exoplanet Fomalhaut b.
Turning ‘off’ certain bright pixels in an image can achieve the same result, but without the necessity of sun trackers or mechanical structures obscuring other parts of the image. That’s what we’re hoping the DMD will do for us.
But first I have to put it together.
Simple enough, right? It turns out nothing is that simple. Whether it’s technological problems (having to replace my hard drive twice in the space of about 2.5 months) or geometrical considerations (more on this later) or simply not having all the right parts (a couple of orders to ThorLabs cleared this up) the process of designing an optical setup can really stretch out. It can take a lot of creative problem solving and completely scrapping previous ideas that it turns out just aren’t going to work. Here, presented for your enjoyment and edification, are two tales of design woe…or possibly triumph.
I inherited the concept of the DMD setup from my nominal counterpart Alex Seguin – he was working with a different DMD which never ended up working. I got the advantage of a brand new and much more user-friendly device. Excellent, so I could just put the new DMD into Alex’s setup, right? Wrong. The way that the mirrors in the old DMD were laid out meant that they turned the incident light either towards or away from the observer in a flat plane. The mirrors in the new DMD directed the light in both the vertical and horizontal directions. Here’s a very quick diagram to illustrate this.
(It may be obvious why I did not go into fine arts)
So while Alex’s setup could be in two dimensions – flat – mine had to add in a third dimension. Cue the first ThorLabs order. The setup is now gloriously three-dimensional, and adjustable, which lets us find that perfect configuration of angles and vertical and horizontal distances to resolve the image from the DMD.
Another problem I encountered was mounting the DMD system. The DMD attaches to a large PCB which in turn connects it to a computer. The PCB is very electrostatically sensitive – basically in a setup full of metal parts it has to be mounted away from all of that. There’s also the issue of mounting it near enough to the DMD that the cable connecting them – the big flat orange thing in the image of the DMD above – can reach.
As I was thinking about how to mount the PCB I remembered my 4th year engineering capstone class. There was a day when the professors brought in a stack of what appeared to be recycling and rolls of tape. The task was to make a “prototype” of our project using cardboard.
And so was the PCB Shelf-inator 3000 born.
There’s still work to be done before the setup is ready to go forth and boldly block out the sun, but I’m confident that any future obstacles can be met with some careful thought and the application of a bit of tape.
No comments:
Post a Comment