A map of Saturn

This map of Saturn was created from 56 images obtained by the Cassini spacecraft of Saturn's southern hemisphere in September 2004. At this time Cassini was observing Saturn almost continuously through several color filters over a period of several weeks to track atmospheric motions. Longitudinal coverage was complete so this data set was suitable as source data for a high resolution map. For the northern hemisphere I used the northern half of a map I created in 2002 from Voyager data - for the northern hemisphere of that map I utilized 31 Voyager 2 images. The final map has a size of 9000x4500 pixels to match the resolution of the Cassini data but the version here is 2880x1440 pixels.

For the Cassini images, the first step was to calibrate the images and then to clean them up a bit using Photoshop. The next step was to reproject the images to simple cylindrical projection, correcting for the different illumination and atmospheric scattering across the field of view on the fly. I used CB1 (effectively red) and BL1 (blue) images and made synthetic green images from these since Cassini wasn't observing Saturn using the green filter. This worked very well. I then mosaicked everything together into a big full color map, using a combination of Photoshop and specialized software I wrote to remove seams and the effects of different illumination that weren't fully corrected for during the previous processing steps.

As previously indicated, for the northern hemisphere I used the northern half of a map based on Voyager images I created back in 2002. For that map, the first step back then was to radiometrically correct the Voyager images, i.e. remove the effects of differing camera sensitivity across the field of view etc. (the image corners were noticeable brighter than areas near the center in the raw images). I then processed the resulting images in Photoshop to remove noise, using both built-in filters and manual 'painting'. The next step was to reproject the images to simple cylindrical projection, correcting for the different illumination and atmospheric scattering across the field of view on the fly. For this I needed to partially reverse engineer the spacecraft's viewing geometry, I had information on the spacecraft's location relative to Saturn but the camera angle information (i.e. in which direction the camera was looking) was very inaccurate, typically off by 0.1 to 0.5 which is huge considering the fact that the camera's field of view is less than 0.5. Usually I used three sets of images obtained through orange, green and blue filters and used these as red, green and blue, respectively. In a few cases orange images were not available so I had to determine which combination of the green and blue images yielded the best orange image in a least squares sense and then synthesize the orange images. This worked very well and in the resulting map it is impossible to see where I used real orange images and where I used synthetic ones.

I then combined the southern hemisphere Cassini map with the northern hemisphere of the Voyager 2 map to get a global map. I then adjusted the color a bit. In particular, I wasn't completely satisfied with the color of some of the cloud belts in the southern hemisphere. The reason was that I had to use synthetic green. As a consequence of this I modified the color of several individual cloud belts with the result that the map is probably indistinguishable from the results I wold have gotten had green data been available.

Normally this would have been the final processing step but due to Saturn's relatively low contrast some faint seams were visible in the map, manifesting themselves as very faint, low contrast and broad north/south trending "belts" that looked very slightly more grayish than other areas. Normally artifacts like this one are completely "hidden" by high contrast details (sort of "can't see the forest because of the trees"). Although this was not noticeable in renderings it could be seen in animations of a rotating Saturn. The only way to get rid of this was to create a "mask" from the map in Photoshop and run this together with the map through a special utility I wrote for processing the map. This made these artifacts considerably less conspicuous but they are not gone. One reason is that the mask I mentioned was difficult to create due to the low contrast of these artifacts - it was very difficult to visually estimate their exact size and shape and I didn't find a way to bring them out in any useful form by processing the map. The final result is great for rendering stills but I'm not completely satisfied with it when doing animations where Saturn is rotating fast although overall I'm happy with it - it'a big improvement over the old 2002 map.

What this all means is that an improved version of the map will hopefully show up in the future.

Click the map below to view the full resolution (1013 KB 2880x1440 pixel JPG) version.

A map of Saturn (small version)

You may notice that the northernmost cloud belt is not straight in the east-west direction but has a kind of wavy structure, unlike the other belts. When it is rendered it shows up as a dark hexagon around the north pole. This is not an error in the map, Saturn really has a hexagon around the north pole, one of this planet's most enigmatic features.

I will probably replace the Voyager 2 data with Cassini data once the north pole gets illumimated following equinox in 2009.

The map is in simple cylindrical projection with latitudes running with a uniform interval from -90 (bottom) to 90 (top) and longitude 0 at the left edge of the map (although longitudes are of no importance for planets like Saturn which have no fixed features, just clouds). Latitudes are planetographic. It should be rendered by projecting it onto an ellipsoid with an equatorial radius of 60268 km and a polar radius of 54362 km or some equivalent units. It is important not to use a sphere, this oblateness is one of Saturn's more distinguishing features and is for example easily visible in small amateur astronomical telescopes from Earth. See also my planetery rendering tips page for further info.

Back to main page