Pics o’ the moment   Leave a comment

One of the threads running through my lab over the years is “scientific imaging”. This is slightly distinct from photography in that an image used for scientific purposes is considered *data*, just like a graph. The tools used for imaging can be very similar (or identical to) the tools used in photography, but imaging is not limited to cameras and lenses operating in the visible waveband.

Here’s a photo:

But the way this image appears is different than how the objects appeared. Specifically, the digital data file of the image was deliberately altered by “raising the black level”- any numbers below a certain value (for this image, the values was 30) were changed to be ‘0’, while numbers higher than 30 were unchanged. This made the trees black (and thus appear in silhouette), and made the sky a deeper hue of blue.

As a photograph, altering the image to make it appear as I choose would pass without mention: photographers have used a wide variety of techniques (burning and dodging, for example) to alter photos since photography existed.

But in science, altering the image means I have altered the *data*. Consequently, there are rules about which particular manipulations are acceptable, and which constitute (scientific) fraud. It’s not clear what that means in the context of the above photo, so let’s look at some images that contain scientific data. Here are two images, one is the ‘raw’ image, and one is processed: it’s not immediately obvious, but one image looks ‘sharper’ than the other.

This image was taken of our cells after fixing and staining the actin cytoskeleton with Oregon Green/phalloidin. Phalloidin is a chemical that binds to a specific type of actin (F-actin), and Oregon green is a fluorescent molecule that glows green.

Here’s the important point- the image contains data about where the actin is. More than that, there is also information about how the cell is attached to the filter. This is information, because the actin cytoskeleton is created and remodeled by the cell in response to it’s environment- by changing the environment and observing the actin cytoskeleton, we can learn about how the cell response to it’s environment.

The altered image contains the same information- that’s critically important in image processing. The specific alteration was deconvolution. In essence, the signal (information we care about) was extracted from background noise (what we don’t care about). Using this image as scientific data, I simply disclose that the image was deconvolved (and provide details about the specific algorithm) and all is well. With scientific imaging, the rules about what is acceptable are based on the same rules regarding data manipulation- removing outliers, or emphasizing specific features, for example.

But understanding how images contain data also means we can do some fun things: measure the spectra of stars, for example: here’s a photo I took using equipment in our lab:

This image was taken with our 35mm format camera and a 15mm f/3.5 lens. I slid a small piece of a transmission diffraction grating into the filter holder of the lens, and thus was able to image the diffracted spectrum from the sun (visible at the top of the frame).

The two ‘rainbows’ (a full one and a second partial one at the bottom of the frame) are a measurement of the different colors of light given off by the sun. The sun is a (very) hot ball of gas, and so the spectrum is continuous- black body radiation. The secondary spectrum is identical to the first, but is stretched out more- here’s a shot trying to capture most of the secondary spectrum:

Now, here’s an image of streetlights:

The ‘rainbows’ look very different! The secondary spectrum to the closest light is a little visible, but the spectrum does not look smooth- there are blobs here and there. Each blob in a spectrum is an image of the light, so what actually matters is the (discrete) positions of the blobs (as opposed to the existence of blobby things…)

Here’s the spectrum of a typical Mercury streetlight:

There are two emission peaks in the violet (centered at about 405nm and 440nm), two in the green (about 528nm and 555nm), and then weaker features in the red (650nm), blue (460, 480nm) and ultraviolet (there are actually two, closely spaced near 360nm). The ultraviolet lines make Mercury bulbs hazardous, so don’t play with one unless you know how to block those lines.

Looking at each streetlight spectrum, you can see two blue blobs, and then three more, maybe green yellow and red. Furthermore, the red blob looks brighter than it should (based on the size of the red emission peak), possibly providing information about how the camera amplifies the different colors. To dig further would require a discussion about Bayer filters (how single chip color cameras function), resolution, and interpolation. So we’ll stop at ‘blobs’, for now.

Here’s an image of Antares and it’s spectrum:

Again, it’s a continuum (stars are hot gas). This image was taken with our 400mm f/2.8 lens (and a small piece of diffraction grating) and if you look carefully, you can see the spectrum of a different star.

This type of imaging, of obtaining spectral information about the objects, is called “hyperspectral imaging”, “multispectral imaging”, or “imaging spectroscopy”.

The 400mm lens is effectively a 6″ diameter refracting telescope, as we can also do some astrophotography- putting a 2x teleconverter changes the lens to a 800mm f/5.6- which gives great images of the moon: These images were taken at slightly different times, the redder image was taken closer to moonset (is that word?).

We look forward to presenting other images- we have lots of optical tools that can be adapted to photography, and we hope to show some interesting results.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: