Io erupts, in color

The last one of New Horizons imaging instruments has finally checked in with a lovely image from the Jupiter flyby. I've already shown you plenty of images from the high-resolution LORRI camera (see, for example, this one, and these). I talked about data from the UV spectrometer, ALICE, and from an infrared imager, Ralph LEISA. At last there is an image from the other component of Ralph, called MVIC. The data from MVIC has been hotly anticipated, because MVIC is the New Horizons instrumentthat can provide images in four different visible and near-infrared wavelengths, which can be composited into color views approximating what the human eye would see. The first MVIC view to be released is a spectacular one of Io erupting:

As it turns out, this image is not, in fact, a straight red-green-blue composite. The New Horizons team explains on their website that even with this thin crescent view, the sunlit part of Io was still overexposed as seen through two of the broadband filters, the red and infrared. The only two useable images for this composite were the one taken through the blue filter and one taken through a narrow-band "methane" filter at 890 nanometers. That's also the near infrared, not very much longer-wavelength than the red light visible to human eyes, so for a target that doesn't have a lot of methane on its surface, it's not a terrible choice to use as the red channel in a color image. They didn't specify what they did to create the green channel, but one thing they could have done would be to average the blue image and methane image together to synthesize a green image.

MVIC has a pixel resolution of about 20 microradians. LORRI's resolution is four times better, at about 5 microradians. So it would be very nice to be able to combine the color information from the MVIC image with the sharpness of a LORRI image to make a beautifully detailed color version. In fact, New Horizons did take three LORRI images of Io at just about the same time as this MVIC observation, so the data exists to make a very pretty, very detailed color version. Here's mine:

For those of you who would like to play along at home, here's what I did to make that image. I use Photoshop, but the same operations can be performed in other image processing software.

First of all, I checked out the LORRI images on the New Horizons website. There were three taken at 00:35 UT, just 10 minutes later than the MVIC observation. The three had different exposure settings of 4, 20, and 75 milliseconds. The 4- and 20-millisecond images look pretty similar and mostly show the sunlit crescent. The 75-millisecond image overexposes the crescent but contains better imaging of the nightside plume. So the first thing I did was to combine the left side of the 20-millisecond image with the right side of the 75-millisecond image to bring out details in the plume. This also brought out some details along the terminator that weren't visible in the 20-millisecond image. I had to line up the two images first, which I usually do in Photoshop by pasting one as a layer over the other, then setting its blending properties to "difference" and shifting it around until the two images cancel each other out almost perfectly.

The next step is to resize the MVIC image so that it matches the resolution of the LORRI image. By measuring from horn to horn along Io's crescent, I found that the released MVIC image appeared to have roughly half the resolution of the LORRI image. (This means they must have resized the MVIC image, enlarging it by a factor of two, before releasing it.) So I enlarged the MVIC image by a further factor of two, and aligned the enlarged MVIC image with the composite LORRI image. The match wasn't perfect, because the two images weren't taken at quite the same time. New Horizons moved about 3,000 kilometers further away from Io in the 10 minutes separating the two pictures, and the terminator also moved as Io rotated by about 1.5 degrees. But features seem to line up to an accuracy better than one MVIC pixel, so that's not bad.

There are several ways to do the next step; here's how I did it in Photoshop. What you want to do is to retain the color information from MVIC, but the contrast information from the LORRI image. Color images are most typically shown as three-channel, red-green-blue color composites. But that's not the only way to represent an image. Another way is the L*a*b* color space, which also has three channels, but instead of each channel representing brightness or darkness in red, green, or blue, L*a*b* uses two channels to encode color information and one channel to encode "lightness." I won't go into it more here; you can check out the Wikipedia article on the L*a*b* color space if you want more detail. In Photoshop, I converted the MVIC image from RGB color to Lab color (Image > Mode > Lab color) and then threw out the "Lightness" channel. I pasted in the LORRI image as the new Lightness channel. Voila! An image with the beautiful detail of LORRI, and the color information available from MVIC. The final image was a bit blue for my taste, so I shifted the color a bit toward orange. Finally, I converted the picture back to the RGB color space so that it could be displayed here.