Martian Vistas

Here is my resent analysis of PSP_001414_1780 overlaid onto the southern portion of the image at 1/4 of the full resolution (and some smaller sizes).

Quarter resolution (7449 x 5562) (10.8Mb)
One Eighth resolution (3725 x 2781) (2.9Mb)
One Sixteenth resolution (1863 x 1391) (884kb)

Here is the Green->Blue->Red Fourier Transform analysis overlayed onto the HiRISE image for the section just south of Victoria (including the flat area around Victoria top centre - in green). This shows how well it does at picking out the differing ripple sizes. As you can see, it does not pick up some of the more isolated big ripples, however I would argue that this is not really an issue as they sit in a field of much smaller ripples (that the algorithm is detecting) and would be easy to navigate around. They are also blindingly obvious in the raw HiRISE image so I don't think we'll run into them! Apart from that, this method does a very good job of finding areas of differing ripple sizes.

Full resolution (29794 x 5152) (20.2Mb)
Half resolution (14897 x 2576) (6.8Mb)
Quarter resolution (7449 x 1288) (2.2Mb)
One Eighth resolution (3725 x 644) (660kb)

Here are the raw analysis files as uncompressed PNG files as requested by Paolo in the UMSF thread.

Greyscale, Black -> White, 256 levels (4.8Mb)
Colour, Green -> Blue -> Red, 256 levels (4.9Mb)
Colour, Green -> Blue -> Red, 3 levels (4.8Mb)

Now that I've convinced myself that I can reliably map areas with larger and smaller ripples, it's time to concentrate on the road ahead. Here is the HiRISE image overlaid with my analysis for the terrain south of Victoria crater. At the bottom right is the start of of what looks to be almost ripple free terrain for as far south as the full HiRISE image shows. So crossing this area from Victoria crater at the top left to the green pastures at the bottom right will be the tricky bit.

Full resolution (8614 x 10240) (31.8Mb)
Half resolution (4307 x 5120) (9.4Mb)
Quarter resolution (2154 x 2560) (2.7Mb)
One Eighth resolution (1077 x 1280) (772kb)

Here is my latest Fourier transform analysis colour coded onto the HiRISE image PSP_009141_1780 at one quarter of it's full resolution. I've left it split into the sections I made for running the analysis to make things more manageable, working with HiRISE data you can soon end up with vast images! One quarter is enough to see the problematic ripples that this is trying to detect and confirms that it is working really well. In fact, much better than I ever imagined possible, almost every ripple is marked correctly. The latest attempt to add resolution has really helped as it is now even able to resolve some of the more isolated big ripples.

Quarter resolution (7644 x 1280) (968kb)


Quarter resolution (7644 x 1280) (1.2Mb)


Quarter resolution (7644 x 1280) (1.6Mb)


Quarter resolution (7644 x 1280) (2.2Mb)


Quarter resolution (7644 x 1280) (2.7Mb)


Quarter resolution (7644 x 1280) (2.9Mb)


Quarter resolution (7644 x 1280) (2.7Mb)


Quarter resolution (7644 x 1280) (2.7Mb)


Quarter resolution (7644 x 1280) (2.9Mb)


Quarter resolution (7644 x 1280) (2.7Mb)


Quarter resolution (7644 x 1280) (2.6Mb)


Quarter resolution (7644 x 1280) (2.2Mb)


Quarter resolution (7644 x 1280) (2Mb)


Quarter resolution (7644 x 1280) (2Mb)


Quarter resolution (7644 x 1280) (1.5Mb)

Another attempt at my Fourier analysis of the ripples at Meridiani. This time I have computed values over 64x64 squares instead of the 128x128 squares and also only shifted 32 pixels for each derived pixel. The result is a much higher resolution colour overlay. I've also changed the mapping of values to colour to stop so much of the area north of Victoria being saturated red.

Full resolution (2048 x 5380) (4.2Mb)
Half resolution (1024 x 2690) (900kb)

Here are a couple of sections at 1/4 of the full HiRISE resolution. This is just enough to see the individual ripples and be able to gauge how well the method of generating the map is.

Quarter resolution (5656 x 1280) (1.9Mb)

Quarter resolution (5624 x 1280) (1.9Mb)

Opportunity is about to set off on a long trek south towards a big crater named Endeavor. In an attempt to help determine the best route to take avoiding the worst of the ripples that have caused problems in the past, I have started to work on using a Fourier analysis of the HiRISE images to map them out.

This is a work in progress but here is my first attempt at analyzing the whole of PSP_009141_1780. Red indicates areas with large ripples whereas green shows nice smooth areas.

The crater towards the top is Endurance near where Opportunity landed, in the middle is the large crater Victoria, where Opportunity is now and where Opportunity is headed is off the image in the direction of the bottom right.

Full resolution (2048 x 5380) (2.8Mb)
Half resolution (1024 x 2690) (828kb)

The bright ripple on the edge of Victoria Crater named "Bagnold"

Full resolution (1172 x 2060) (892kb)
Half resolution (586 x 1030) (216kb)

Anaglyph (3D glasses required). There is a large parallax between the left and right images so this isn't the easiest anaglyph to view I'm afraid.

Full resolution (1532 x 2072) (1Mb)
Half resolution (766 x 1036) (244kb)

Tracks - New, 1 year old, and 2 years old (Earth years)

Full resolution (2204 x 1104) (1.0Mb)
Half resolution (1102 x 552) (292kb)

Anaglyph (3D glasses required)

Full resolution (2188 x 1090) (936kb)
Half resolution (1094 x 545) (268kb)