MRO Hirise data processed with ISIS3

[cartrite]

I down loaded edr data from PDS taken by the hirise camera on the MRO which shows the "Face on Mars" (orbit 03234-2210). This smaller version of the output shows a fairly decent image. The full resolution image is a half meter per pixel. The original data had an average resolution of .299 meters per pixel on the red channels.
The blue, green, and Infrared had a resolution of about .6 meters per pixel but only had data for part of this image. The red channel has 10 ccd's and the others only have 2. 2 are for bg and 2 are for ir. My next challenge will be to align the blue and green channels on red ccd's 4 and 5. This will produce a color image centered on the "face". The image looks a little out of shape because it projected to a simple cylindrical map. When I'm finished I'm planning to create a level12 vt. With all the other levels down to 6.




cartrite

[cartrite]

Here is a link to a page at my site that contains full res images of the above image and the second mosaic I created last night of the Proposed MSL Site in Melas Chasma. I'm planning to create a model with 128 points per degree of this area to use with the color portion of this mosaic.
EDIT: I ran into a problem which is going to delay my plans. The CCD IR10_1 is not working right on the HiRise camera. This means that every image sent down so far has ony one area out of 10 that has data for color images. There were supposed to be two but the near infra red filtered CCD is returning data that is either corrupted or too noisy to create aa quality image. I'm working on a way to fool the software into using the data from the BG12_1 CCD as the IR10_1. Hopefully I'll come up with something.

There are some color images at my site that are unprojected camera and a few projected to simple cylindrical.

Bottom line is it may take years before enough images were sent to earth to accumilate enough data to create mosaics, ( color or greyscale ). That is if the spaceship survives that long in it's hostile environment.

From orbit 4052-2045 it looks like the IR10_1 CCD is running a little better. Here are a couple of thumbs of CCD's RED4 and 5, BG12 and 13, and IR10 and 11. Other tiles are at my site.




The grey image with all RED CCD's can be seen here. http://hirise.lpl.arizona.edu/PSP_004052_2045

EDIT:The link above is a link to a different page which has been recently updated. It dose not contain the "Face on Mars" or files from orbits 002551_1700 and 004052_2045.
cartrite

[cartrite]

A few images from orbit 4060-1440 that I processed with ISIS3. The info can be seen at http://hirise.lpl.arizona.edu/PSP_004060_1440
These images are 2048x2048 tiles from a .5 meters per pixel image I processed today.





I think the max resolution is 9.75 meters per pixel for level12 vt's. Too bad that would make the full image that these images were cropped from only about 276 pixels wide. I'm still trying to find a way of using the full .5 mpp textures on a model which would be a workaround for the level12 limitation.

cartrite

[selden]

9.75 meters/pixel is for tiles that are 512 pixels on a side. You can at least halve that since modern systems support up to 1K/side. If you use 4K tiles (although many systems can't) you'd get about 1.2 meters/pixel.

[cartrite]

Hi Selden,
9.75 meters per pixel is plenty big enough but each image the hirise camera takes would only be about 276 pixels wide by about 1638 at 9.75 meters per pixel resolution. Those are the dimensions of the images I mapped at the 9.75 resolution. If the hirise camera was taking images that were overlapping enough to create mosaics of larger areas of the surface, a level12 vt could be produced of a couple of 512 or 1024 tiles but that is probably not going to happen. Even if they get around to overlapping the red ccd's in some areas which has about 20 channels across, those images are gray scale. The color portion is only 4 channels across. So color images are probably never going to overlap. To date no images have overlapped. I'm not sure if they ever plan to overlap any. I guess it depends how long the camera lasts.

They are also on a tilt. So the upper left is a couple of fractions of a degree further west than the lower left and the upper left is also a couple of fractions of a degree further south than the upper right . The actual size would have to be smaller if the left and right edges had the same longitude and the upper and lower edges had the same latitude.

Anyhow any vt done would be a series of stripes no larger than 276 pixels across at level12. The lower levels would be even smaller. etc. Unless some of the hirise images overlapped. Given the size of the planet and the resolution of the camera images I don't think that will ever happen in color or gray scale. Unless of course that is part of the mission plan for some areas.

Any image taken so far could be an ultra close up. They can be about 5000 pixels across and 60000 pixels high. That would be the equivalent of a level20 or so vt!!! Celestia doesn't have any plans to support level 20 vt's does it?

Anyhow, thats why I'm going to see if the model approach will work. Had limited success so far with that idea.
cartrite

[cartrite]

Here are a couple of images I've done recently I thought looked cool that are not on my site.






cartrite