PDA

View Full Version : OT: The most bizarre landscape I have seen



Evan
09-26-2012, 05:18 AM
Seen in a photo, anyway. This is a small area on Mars not related to the current Rover mission. When I saw this image I had to model it to try to make some sense of it. The area is about 1.5 square kilometres in size meaning that if a person were in the picture you could see them. The entire scene is beyond fantastic. It is apparently a result of a wide variety of minerals that were dissolved in water and deposited where they dried to create this weird Mars-scape. They have also discovered that there is still liquid water on Mars and on a hot summer day it sometimes seeps out of faults in some crater walls. The ground temperature in some areas can actually reach up to 80f sometimes.

This is the satellite image looking straight down. The HiRise camera has very high resolution of about 16 inches so it can pick up a lot of detail. This reminds me of the fractals I was playing with recently.

http://ixian.ca/pics9/riot2d.jpg

This is the model made in SketchUp. I wish I had the ability to download the really accurate terrain models but the files range from 300MB to a gigabyte or so. Even so, the shape of the landscape is just as strange as the appearance. It was very difficult to get the image correctly aligned with the model. Nothing is familiar. The small ridges are little sand dunes maybe 20 feet tall.

http://ixian.ca/pics9/riot3d.jpg

The low gravity allows sand to form tall sharp edged shapes that are impossible on Earth.

Arcane
09-26-2012, 05:23 AM
Alien landscape is alien.

goose
09-26-2012, 08:01 AM
Intriguing. Looks almost like one of your fractal images.

boslab
09-26-2012, 08:14 AM
reminds me of fractured flint!
mark

vpt
09-26-2012, 09:31 AM
Probably just piles of gold, diamond, and other useless rocks.

gwilson
09-26-2012, 10:01 AM
Those golden spots are enormous chunks of GOLD!!!! Actually,they could be!!

Evan
09-26-2012, 01:13 PM
Or they could just be sulphur. Mars had a very busy volcanic period that lasted a billion years. I sure would like taking a trip to Mars just to look around. The tallest volcano in our solar system is on Mars. It is named Olympus Mons and is about 14 miles tall thanks to 1/3 gravity.

hitnmiss
09-26-2012, 01:19 PM
Sort of reminds me of one of the big hot springs in Yellowstone...

interrupted_cut
09-26-2012, 06:27 PM
Where did you get the elevation data for the model?

Evan
09-26-2012, 08:02 PM
I convert the colour shaded relief image back to a grey height map by splitting out the hue map. As long as the colour map is correctly arranged it works fine. The colour map used by NASA is ok up to a certain relative altitude and then they go to "illegal" colours that don't convert correctly. In this case the problem was that the entire area was much larger with much larger altitude differences. That reduced the contrast greatly on this area which makes it very difficult to pull out the differences accurately. It is fairly close but not entirely correct.

I downloaded the entire 350 meg DTM file last night and will have a look to see if I can improve the height map quality.

The colour DTM looks like this with the area in question highlighted:

http://ixian.ca/pics9/colordtm.jpg

bob_s
09-26-2012, 08:06 PM
Reminds me of the ceiling of a sub-glacial ice cave.

aostling
09-26-2012, 09:23 PM
The entire scene is beyond fantastic.

It is indeed. It bears some resemblance to tafoni http://www.tafoni.com/Welcome.html, a rock-weathering pattern observable on Earth. But it is beyond that too.

You have discovered this. There is no other way to put it!

Evan
09-26-2012, 11:01 PM
I didn't discover it Allan. NASA is well aware of this formation which is why there is a satellite image of it. They have also created the digital terrain model from a stereo pair to they consider it important. I recently found out that making a single DTM from a stereo pair requires approximate one full week of graphics analyst time. That explains why they make so few.

Smokedaddy
09-26-2012, 11:56 PM
Looks like a close up view of a Campo del Cielo meteorite.

What would be another way of saying/explaining to someone ... the color map used by NASA uses "illegal" colors?

http://qt-project.org/doc/qt-4.8/QColor.html

-SD:

aostling
09-27-2012, 01:17 AM
I didn't discover it Allan. NASA is well aware of this formation which is why there is a satellite image of it.

Your model is the first I've seen, and it is stunningly rendered. You imply that you have not exaggerated the height relative to the lateral dimensions. With the low Martian gravity the angle of repose is so much greater, it really is like nothing seen before.

dp
09-27-2012, 02:44 AM
Your model is the first I've seen, and it is stunningly rendered. You imply that you have not exaggerated the height relative to the lateral dimensions. With the low Martian gravity the angle of repose is so much greater, it really is like nothing seen before.

The Z axis appears exaggerated to me and I've worked with surface/side-scan radar and sonar and recognize the characteristics. Doesn't mean it is exaggerated, but I can't tell the difference based on my experience. It is normally done to amplify surface features of which color mapping is one such use. Another contributor is the low-res color data Evan had to work with. The steps in the 3D image are characteristic of low-res stair casing. But I wasn't there and don't know - I just know that what Evan has produced is stunning.

Evan
09-27-2012, 04:33 AM
I don't know if it is exaggerated since the models I am using aren't calibrated. However, the models I have seen from NASA make a point of noting "Vertical scale not exaggerated" since they do look so. The low gravity allows formations to exist with an angle of repose even for loose material that looks almost vertical. It's definitely an alien planet and Earth based experience is of no use in judging what you see on Mars. I can forsee many hazards that would exist for visitors not accustomed to the differences. An almost vertical sand dune face could fall on top of you instead of sliding down a slope.

I have just determined that I can import the official DTMs into Blender as a calibrated mesh and export them to Collada format that SketchUp can read. It will be interesting to see how they look.


The steps in the 3D image are characteristic of low-res stair casing.

It isn't. The model has about 500,000 vertexes (verti?) vertices. However, the height data is only 8 bit so that does produce some artifacts.

Evan
09-27-2012, 04:52 AM
What would be another way of saying/explaining to someone ... the color map used by NASA uses "illegal" colors?

That is difficult to explain fully without first doing a course on how colours are related to hues and hues are related to luminance. I was trained in all this in my former job at Xerox since I was a colour machine tech along with every other product Xerox made. There is quite a bit of math involved, mainly trigonometry. The "correct" colours must be restricted to the phase circle of primary additive and subtractive colours arranged at 60 degree positions around a polar vector colour map. Any colour that isn't in that correct sequence will produce a sudden jump in luminance, either brighter or darker and ruins the linear height relationship to the polar hue sequence. The colour map I posted shades from red to brown and then white at the highest altitudes. The brown and white do not exist in the colour vector circle and so produce incorrect luminance values.

Sorry about the increased rate of typos lately, especially missing last letters like "s". My proof reading seems to have deteriorated slightly since the stroke.

Smokedaddy
09-27-2012, 02:24 PM
......

Evan
10-06-2012, 02:06 PM
I was poking around on NASA's HiRise website and found this terrain model of the subject of this post. I am rather pleased because it confirms the accuracy of my model, especially the vertical scale which I nailed pretty closely. Mine is at a lower angle but in all it is very close to the NASA model and they had the benefit of using the full resolution terrain model which is over 600 megabytes.

NASA's model in B/W:

http://ixian.ca/pics10/nasariot.jpg

My model:

http://ixian.ca/pics10/ewriot3d.jpg

I have developed a protocol for using the full scale terrain models using Blender and Meshlab. Blender sure has a steep learning curve and Meshlab is severely handicapped by not having any undo function. It is also very time consuming since even with a reasonably fast machine some of the files are using up all of my 16 gigs of ram. To take full advantage of these huge files I need a machine with at least 64 gigs ram. (!!!)

dp
10-06-2012, 02:53 PM
Just WOW!

While on vacation in the Okanogan area of Washington (Oroville) I took a drive along the Similkameen river to see what will happen when they put in the high dam. It will flood Cawston, BC, for example, and generally make a mess of things all over the area. Then as I went through the canyon where the new damn is proposed I recalled your flooding animation of Gale Crater. I'd like to try that for the flooding of the Similkameen river valley and show towns, homes, ranches, the rail tunnel, and other landmarks slipping under the flood.

I'm going to have to study your methods.

Evan
10-06-2012, 04:46 PM
I will have to look up the GIS maps for the area and see what I can do. If I can find the time, that is. I am right now helping out a guy in Poland alpha testing a neat software project for good quality ray tracing inside SketchUp. It's already well along and really to the beta test stage. He has a web site on it but I'm not posting it yet because I already found a bug in his latest version that will lock up Win7 solid.

aostling
10-06-2012, 08:00 PM
I am right now helping out a guy in Poland alpha testing a neat software project for good quality ray tracing inside SketchUp.

So much for that odd notion, that you were "retired."

I found the website of your partner Tomasz. With ray tracing, and perhaps Monte Carlo techniques, this could adapt Sketchup for use in spacecraft thermal analysis.

Evan
10-07-2012, 12:18 AM
Hmmm. His name isn't Tomasz.

aostling
10-07-2012, 03:40 AM
Hmmm. His name isn't Tomasz.

My detective skills evidently need sharpening. Tomasz is presumably Polish, and has this website asking for testers: http://tomaszjaniak.wordpress.com/2011/04/01/shadow-analysis/.

Evan
10-07-2012, 04:55 AM
That looks interesting but it isn't the same project. I will post about it as soon as the next version is available. I don't want anybody using the current version because it has a bad bug that will bite and may cause data loss.

2ManyHobbies
10-07-2012, 08:26 PM
Evidence of strip mining and slag heaps? :p

Doozer
10-08-2012, 08:34 AM
Anyone see Waldo in there??

--Doozer

A.K. Boomer
10-08-2012, 10:04 AM
Tell you the truth it looks like something that came out of my sink drain a few weeks back...

Evan
10-08-2012, 02:35 PM
This will give you an idea of the scale of the landscape. It is even more amazing. I have worked out how to deal with huge file sizes and can now make use of the extremely high resolution images. This is the yellow spot that I suggested might be sulphur. It very much does look like a pile of sulphur in this image. The pile is about 50 feet across and about 20 to 30 feet tall. Just try to imagine standing in that place and what it would be like to walk around there. It beggars the imagination.

http://ixian.ca/pics10/sulphur.jpg

aostling
10-08-2012, 07:44 PM
The pile is about 50 feet across and about 20 to 30 feet tall. Just try to imagine standing in that place and what it would be like to walk around there. It beggars the imagination.

If NASA knows about this pile and has not released the image to the media they need a lesson in public relations.

It appears to be in a crater. Or is it on a shelf? I'd really like to see your 3D model of this.

topct
10-08-2012, 08:14 PM
So if we could smell Mars, it would really stink?

Evan
10-08-2012, 08:20 PM
Keep in mind that there is one very important aspect to this type of terrain map. It is entirely missing any data on overhangs or undercuts. There is no feasible way to gain that information from overhead, not even with widely separated stereo satellites. I don't know if I can model this particular small area with meaningful elevation resolution. Unlike the images the elevation data must loaded in its entirety and this area comes from a sequence that has a very large terrain map. I will try it and see if it is possible since I have downloaded the full resolution terrain map. The elevation posts are situated on square metre corners and the vertical accuracy is as good as 10 cm. I find that amazing as well since it is much better than any declassified data available for this planet. Of course, that statement is self explanatory as to the reason why. It also shows what is possible even with declassified equipment and methods.