New Horizons sees surface features on Pluto, begins raw image release

Today the New Horizons team released a new animation of images taken on approach to Pluto. The animation clearly shows how Pluto wobbles around the Pluto-Charon barycenter. It also shows something more exciting to the scientists: variations in brightness across the surface of Pluto.

We knew those variations existed, but they didn't expect to see them so early in New Horizons' approach, according to project scientist Hal Weaver. He said at a press briefing today that they didn't expect to see the surface variations for another month. They were made visible thanks to a process of stacking multiple photos taken one after another, and deconvolving the result. Hal said that the astronomy community's deconvolution algorithms were developed after Hubble was launched with its initially blurry vision -- so the Hubble problem has had a silver lining for the New Horizons mission!

The most common question I've been asked about this image today is: Is Pluto lumpy? The answer is no; Pluto is round. It appears lumpy because of those very brightness variations that the scientists are so excited about. Wherever a bright spot is near the edge of the disk, it looks like an upward lump; wherever a dark spot is near the edge, it looks like a downward lump. The variations in surface brightness are probably going to make it very hard to figure out what kind of topography Pluto actually has until we're quite close to it. In the meantime, they make animations of Pluto very interesting!

The other big announcement today is that the New Horizons team has begun sharing their raw images on the Web! Unlike with the Jupiter flyby, where they shared lossless PNG-formatted data, they are putting the Pluto images out in lossy JPEG format, which reduces their quality. While of course I'd love to have PNGs, lossy JPEG format is a tradeoff I'm certainly willing to accept in exchange for rapid access to images that help me to follow the mission! They say that images will generally appear on the raw images website within 48 hours of their receipt.

I've spent some time on the raw images website this afternoon, and thought I'd give you a tour of the kinds of images you'll find there.

First of all, I can decode the filenames for you. The filename convention is:

lor_NNNNNNNNNN_MMMMM_sci_P.jpg

Where

• "lor" tells you that it's a LORRI image
• NNNNNNNNNN is the mission elapsed time in seconds
• MMMMM is an "ApID", a hexadecimal code that contains information about the observation (see table below)
• "sci" tells you that the data has been lightly processed from the raw engineering data (it is "Level 2" rather than "Level 1" data)
• P is the version number (there may be multiple versions of images; it looks like the raw page links only to the latest-version one for each observation)

Here is a lookup table for the ApIDs:

So far, all of the available data appears to have ApIDs of 0x630 or 0x633. That tells you that only the Side 1 computer has been used (presumably Side 2 will only get used if something bad happens). It also tells you that all these images have been transmitted to Earth losslessly. That will continue to be true for all of these early images because they are being used for optical navigation and the New Horizons navigators need every scrap of precision they can eke out of the data. We'll begin seeing lossy data closer to encounter. The main difference between 0x630 data and 0x633 data is that the 0x633 images have been binned 4x4 before transmission to Earth. That means that the images are only 256 pixels square. As I explained in an earlier blog post, this binning is necessary to allow LORRI to be able to take the long exposures needed pick out the faint light from Pluto's smaller moons Nix and Hydra from very far away without smearing the images.

Here's an example 0x630 image. These images are usually taken with relatively short exposures, so you can pretty much only see Pluto and Charon. Few or no background stars are visible. This image had an exposure time of 0.1 second.

Here's an example 0x633 image. These are taken with much, much longer exposures in order to reveal the faint moons. Lots of background stars are visible; astrometry.net readily identifies the star field. (Thanks to some friends on Twitter for reminding me about astrometry.net!) The long exposures cause pixel bleeding from the relatively bright Pluto and Charon, which appear as jagged spears of white pixels trailing to the right. This one is a 10-second exposure.

I downloaded a bunch of these pictures and tried stacking them to reduce artifacts and aligning them to reduce the visual confusion of dancing stars in the hopes that I could spot Nix and Hydra, but I don't know that it was effective:

There are a few weird-looking photos on the raw images page. Some of them, like this one, are very light-colored with a vertical stripe down the center. If you look within the vertical stripe, you'll see Pluto and Charon. The website helpfully explains what is going on in these images:

Sometimes only partial images are downlinked in order to reduce downlink data volume using a process called "windowing". The region of interest, also called the "window", is captured from the LORRI detector and the rest of the image is not even downlinked. However, the subsequent ground processing populates the missing regions with zeroes and then subtracts the bias level, which produces negative numbers in the final calibrated image. This combination of negative numbers in the "Blank Region" and "Real Data" in the windowed region, and the automatic intensity scaling used to produce the JPEG image (which ranges from the minimum to maximum values in the image), produces a banded structure in the JPEGs. A similar effect also applies to images for which data have been "lost" (e.g., data dropouts at the Deep Space Network).

A few images, like this one, are very light-colored with scattered darker stripes running left-to-right. The website explains:

Sometimes data are lost during the transfer from the New Horizons spacecraft to the Mission Operations Center at APL (e.g., bad weather at one of the DSN sites can sometimes cause data dropouts). When these dropouts occur, the subsequent processing of the images is affected and some of the missing pixels are assigned incorrect intensity values. Furthermore, these incorrect intensity values can adversely affect the automatic scaling used to produce these figures, with most of the field appearing to be white and the missing pixels set to black. If an image "corrupted" in this way is considered "critical", the Mission Operations Center may request a re-transmission of the data from the spacecraft. But if the loss of the image doesn't significantly impact ongoing operations or science analysis, we will not try to re-transmit the image from the spacecraft.

The following table shows when you can expect New Horizons to be taking optical navigation images like the ones available on the raw images website so far. I posted this once before; thanks very much to Kim Ennico for getting this information to me! There are two periods during Pluto approach when New Horizons is not taking new images, but is instead focusing on transmitting data to Earth. We're in the middle of the first one of those (April 19 to May 8), and there will be another gap May 15-27. But those gaps are when the lion's share of the data arrives on Earth. So stay tuned to the raw images website for more and more data!