Curiosity Image Retrieval Tools, scripts and software Options

[atomoid]

FYI seems to be some duplication or parsing issue going on with my favorite Mars image browser Midnight Planets.
a slew of 'new' images from sol540 came in today. At the time it said "Posted Sol 774 00:52 (2 hours ago)"
This appears to have resulted in apparent duplicates in the 540 directory under a different filename.
http://www.midnightplanets.com/web/MSL/ima...94E01_DXXX.html
http://www.midnightplanets.com/web/MSL/ima...000E1_DXXX.html
only one of the images exists on the NASA site.

[mhoward]

Atomoid: Yeah, that's an ongoing legacy of the filename change for Mastcam, MARDI and MAHLI images on the web. Every so often, I'm not exactly sure why, a bunch of images get updated to the longer filename format used by the PDS. (Which is a good thing, I hasten to add; the problem was they all should have been using the PDS format in the first place.) But since I've already downloaded the originals, my program sees both. For the moment, when I notice this, I'm running a batch job to clean up the old ones. Which I've just done, so this batch should be cleaned up now. Thanks for mentioning it.

[fredk]

Hi Joe - I noticed a little quirk on your image browser. These two images:
http://mars.jpl.nasa.gov/msl-raw-images/ms...076D01_DXXX.jpg
http://mars.jpl.nasa.gov/msl-raw-images/ms...077D01_DXXX.jpg
clearly have exactly the same pointing, but are shown differing by about 5 degrees in bearing:

Any idea why the discrepancy?

[jmknapp]

Sometimes the SPICE-reported pointing is off and in flux for the latest images as they presumably rectify the data over several days, but for these images over a month old it looks like we're into residual error territory.

BTW, not sure what's going on with SPICE and the latest reported drive, but it's erroneous. Hopefully that will get fixed shortly!

[elakdawalla]

Joe, I've noticed that for some sites, not all of the Navcam images are getting displayed. For example, at this site:
http://curiosityrover.com/synth/?camera=NL&station=31
There should be a lower tier of Navcam images around the Chemcam targetable region to the right side of the rover:
http://www.midnightplanets.com/web/MSL/sol/00125.html

Should that lower tier be displaying?

(Doing book-related work now, switching constantly among your Synth site, Midnight Planets, and the Analyst's Notebook)

[yogi]

Atomoid: Yeah, that's an ongoing legacy of the filename change for Mastcam, MARDI and MAHLI images on the web. Every so often, I'm not exactly sure why, a bunch of images get updated to the longer filename format used by the PDS. (Which is a good thing, I hasten to add; the problem was they all should have been using the PDS format in the first place.) But since I've already downloaded the originals, my program sees both. For the moment, when I notice this, I'm running a batch job to clean up the old ones. Which I've just done, so this batch should be cleaned up now. Thanks for mentioning it.

Have you had any luck figuring out how those 23 char (and 24 chars, as you mentioned earlier in this thread) imageIds are structured? There are also 30 char imageIds such as 1166MR0052750040602156E01_DXXX. The only format I've found documented in msl_camera_sis.pdf is the 36 char imageIds (40 with the file extension), and those ones seem to be the only ones recognized by this filename decoder as well. I haven't found documentation on how the 23, 24 or 30 char imageIds are structured.

Update ah, mcaplinger wrote a bit about that in post #311. I guess I'll have to read some more through the SIS to see if I also find more into on the 30 char imageIds.

[yogi]

However, if you're going to use it yourself, you might want to be aware of a few current limitations:

FYI I've also been looking at the JSON metadata; I'm looking at maybe working on something [screenshot attached] similar to Midnight Planets, but for Android, over the Christmas holidays, just because MP is awesome and because I need some holiday project. Like you say, the coordinates are relative to the site frame origin; It took me a while to figure that out, wish I'd read this thread first In the ones I looked at, the orientation of the frame doesn't seem to match Mars local level exactly so I'm not sure if those are in SITE frame or in RNAV/rover frame - note how the horizon doesn't line up in the screenshot; never mind all the other errors due to my buggy/sloppy prototype.

A peculiarity that you didn't mention is that some of the JSON files have some broken links with "####" in them (likely an error in some processing stage?). For example sol1094.json points to the image 2N223487758EFF####P0635R0M_.JPG which doesn't exist (this particular id is 27 chars long, they really seem to come in all sizes...).
Update, actually it does exist, it's just that the imageId part needs to be url encoded. Seems like an oversight in the code produces the JSON.

Like you say the incomplete metadata is annoying, e.g. the location and camera parameters (including center and optical axis, ie. C and A) are often completely missing. I haven't checked yet if they are in SPICE (I'm still struggling with SPICE). It's also weird that the rover orientation is given as a quaternion (probably rotation of RNAV relative to SITE?) while the other directions (e.g. camera_vector) are then given as a unit vector - I manually convert this to elevation & azimuth like you do. The only thing I've used SPICE for so far is to obtain the locations of the SITE frame origins relative to each other, since that info isn't in the JSON metadata.

How closely have you implemented the CAHV/CAHVOR model? I've read the paper but haven't implemented it yet. I noticed that in Midnight Planets, the image boundaries are not always straight lines, so I'm guessing you're doing some transformation using the CAHV(OR(E)) parameters? If yes, are you doing that transformation inside the shader?

Attached thumbnail(s)

 

[yogi]

FYI I've also been looking at the JSON metadata; I'm looking at maybe working on something [screenshot attached] similar to Midnight Planets, but for Android, over the Christmas holidays, just because MP is awesome and because I need some holiday project.

In case anyone is interested, I did work on my Midnight Planets clone over the holidays and uploaded it to Play Store a couple of days ago.
It's alpha quality and I'm sure I got some transformations wrong; on the plus side, there is support for Google Cardboard VR headsets (currently available for $5 from GearBest).
I'd love any feedback or suggestions!

Attached thumbnail(s)

 

[James Sorenson]

Very nice Tobias! I have been waiting for something like this for android for quite awhile now. MMB and Midnight planets are awesome, however I only have android devices and it's unfortunate that it has been only restricted to iOS. I have a Galaxy Tab 4 7 inch tablet. I installed it, I can zoom and pan around in the app as viewed from the navcams just after landing, however I can't go to any other locations because the bar isn't showing on my tablet. Not sure if it's a bug in the app or my tablet. But any who, lots of definite potential and I can see this app going far in my opinion. Keep up the great work!

[Sean]

In case anyone is interested, I did work on my Midnight Planets clone over the holidays and uploaded it to Play Store a couple of days ago.
It's alpha quality and I'm sure I got some transformations wrong; on the plus side, there is support for Google Cardboard VR headsets (currently available for $5 from GearBest).
I'd love any feedback or suggestions!

SO glad to see something like this for Android! I'm having the same problem as James, no taskbar. Its running on a Galaxy Note 4/Android 5.1.1. The scene it is stuck on loads fine and is very responsive however. Eager to check out the rest of the app, especially the Cardboard VR support.

Thanks for making the app and good luck with the rest of development.

[yogi]

Very nice Tobias! I have been waiting for something like this for android for quite awhile now. MMB and Midnight planets are awesome, however I only have android devices and it's unfortunate that it has been only restricted to iOS. I have a Galaxy Tab 4 7 inch tablet. I installed it, I can zoom and pan around in the app as viewed from the navcams just after landing, however I can't go to any other locations because the bar isn't showing on my tablet. Not sure if it's a bug in the app or my tablet. But any who, lots of definite potential and I can see this app going far in my opinion. Keep up the great work!

Oh, shoot! Sorry about that, I just saw your message and it's 3:45am. I don't have any immediate idea what could cause it, but I'll try to get the emulator working tomorrow night so I can test it across a wider range of configurations. It works correctly on all four of my devices; my devices all run Android 5.0.2 or 6.0.1, but Sean is running 5.1.1 so it can't just be an issue with older versions of Android.

[yogi]

Very nice Tobias! I have been waiting for something like this for android for quite awhile now. MMB and Midnight planets are awesome, however I only have android devices and it's unfortunate that it has been only restricted to iOS. I have a Galaxy Tab 4 7 inch tablet. I installed it, I can zoom and pan around in the app as viewed from the navcams just after landing, however I can't go to any other locations because the bar isn't showing on my tablet. Not sure if it's a bug in the app or my tablet. But any who, lots of definite potential and I can see this app going far in my opinion. Keep up the great work!

I just pushed version 0.1.3 to Play Store, can you see if that fixes it? I'm taking a stab in the dark - I found one issue that *might* cause it, but if that was the cause then it should only affect devices running Android versions < 5.0, and I haven't been able to reproduce the bug even on an Android emulator running 4.1.2.

If you would like to continue to give feedback, please email me at the gmail address listed on Play Store ( marsviewandroid@gmail.com ) and I can add you to the alpha testers list on the Play Store developer page. That way, I can push fixes to you for you to verify before I bother any other users with a new update to download. (The invitation to email me there and get added to the alpha testers list applies to anyone who reads this!).

[James Sorenson]

I just installed version 0.1.3 and now the bar shows. I'm now able to change locations without issues. Excellent!

The only minor issue I see is Mastcam images are hidden behind the Navcam images.

Sure I can do that.

EDIT: I have Android 5.1.1 as well.

[yogi]

Those who want to join beta channel for my app can add themselves here (let me know if there is a problem):

https://play.google.com/apps/testing/org.codepilot.mars

I just pushed version 0.1.4 to beta (should show up in a few hours) with the following minor improvements:

1. Cardboard stereo VR mode: launcher moved into the main app (a few people weren't able to find the launcher before). This mode is otherwise unchanged and still in early alpha.
2. Made app more responsive while moving to site/drive locations with lots of images or when rotating the device between landscape and portrait mode.
3. On Android 4.2+, skip images that are hidden behind others to make better use of the network.

Update from a few hours later: version 0.1.5 which should appear in the next few hours now also has a blending option that I'm pretty happy with; it helps reduce the stark contrast between adjacent images.