I recently heard that part of the mission is to actually climb the central mountain of Gale crater. Why is that? Wouldn't the most evidence for the actions of water be found at the bottom?

They're not looking for actions of water in this particular crater; they're looking for evidence of past water as recorded in the rock record. For that, they'll be studying Martian history as they drive up through the layered sediments from which the mountain is built, oldest at the bottom, youngest at the top.

As you climb the mountain the layers show the younger and younger history of the changes, that way they can figure out a more complete history of the crater. They do indeed expect to find the answers to the main questions at the bottom, the climb to the top will take many more years than the initial two year mission.

I did get the orthos and my maps on GM but the major issue is that GM DTM is MOLA, so very coarse. I investigated to see if it is possible to override the elevation map but was told that it can't be done in the current implementation of GM. You need to install USGS ISIS, then using the command image2qtree you can generate an image tree that can be imported in GM. You are right that the trick is to find the lat/long. The HiRISE images and DTMs have .LBL files that do include the lat/long of the four corners but they are not very accurate. As a result if you generate a kml file for each strip you will have inaccurate mosaics. What I suggest is to do a mosaic as jmknapp did, determine the four corners for the mosaic and generate the kml file for the entire mosaic. For example for one of the HiRISE orthos I use:



This would take PSP_009294_1750.pgm and generate a directory PSP_009294_1750_out with all the tiles and kml file. Once you have that, you launch GM, switch to Mars, and upload the kml file (File->open, go to PSP_009294_1750_out/PSP_009294_out.kml)

Hope this helps.

Paolo

Yup, I saw that while working on the opportunity map background.


That was the idea and right now I was attempting to use jmknapp's mosaic as starting point but it looks too heavy to my computer.

You have the obvious two options: subsample or crop. Well, you have a third one: get a beefier computer ;-)

Paolo

I used a 64-bit Ubuntu system with i7 processor and 8GB RAM to process the DTMs and a similar Windows 7 system with Photoshop for doing the mosaic. At that, working with these images is a bear--many programs will just roll over and die at random, and everything runs slowly. I was amazed when someone upthread said that this was a "small image" and that they load gigabytes or whatever of Venus image at a clip. Have to wonder what kind of $$ that would entail in infrastructure.

If it could help, I could tile the image for you & send a link to the individual tiles, if you know the dimensions of the tiles you will need. I'm not familiar with how Google Earth (or Mars) works but would love to see that kind of visualization applied to the HiRISE/DTM data.

Ok, here's a first and crude attempt on this thing.
"Crude" because I couldn't edit the mosaic to isolate just the picture and apply transparency to the rest of it. It was simply too big. Anyway, here are a few snapshots.



I used 8192x8192 tiles the last time I worked on this subject. This will probably be my next attempt, tomorrow.
I think my machine will be capable to tile the picture, otherwise I will contact you Joe. Thanks!


Frankly, MapTiler (http://www.maptiler.org/) is a very simple tool. You only need a source image and know its boundaries in lat / long and the tool creates a directory tree with the picture as small tiles in several LODs and a KML file to see the picture in GE. Similar to image2qtree which Paolo explained a few posts above. Do you wanna have a try? If yes, I can give you the exact parameters I used with your mosaic.

If you have the netpbm toolset you can use pnmtile. Joe, the way that these monster images are dealt with is tiling. There are file formats, including TIFF, that support tiling and multiresolution natively. It is not really the computer that is the issue, it is the file format. If you take your maps for example and convert them back to JP2 and use kdu_show, for example, you should be able to move around your bib maps very well.

Paolo

Good to know. I did not particularly liked the idea of having to install ISIS to do the tiling. It is like having to buy Home Deport to use a hammer.

Paolo

Very true. Those big HiRISE JP2 are more lightweight to handle on OpenEV/FWTools than a 30k x 30k JPEG.

Phew! Here's the second attempt.
First of all I split the mosaic in a 3x3 tileset because the original picture was too big for the second step in my PC which was to apply a transparent background instead of the original white color. Then all nine images were processed with MapTiler and here are a few snapshots of the result.


The whole thing is accessible from a single KML file but is split in a directory tree taking 210MB. Now, let's see how can I share it.

Wow, I have to get that. With Google Earth you can overlay tracks, pushpins and the like, right? It'd be fascinating to see the planned path up the mound.

The image underscores that the HiRISE/DTM coverage doesn't go very far up Mt. Sharp. I wonder how soon the HiRISE people plan on doing more DTMs covering all the way to the peak?

I have temporarily hosted the map at Dropbox. Here's the KML file to access it.

Works very slick--GE started up, switched to Mars, homed in on the image, no muss, no fuss. Really fascinating looking at the perspective views, to see the undulating slopes.