Winter Quarters, at Low Ridge Haven |
Winter Quarters, at Low Ridge Haven |
Jun 4 2006, 02:13 AM
Post
#346
|
|
Senior Member Group: Members Posts: 1229 Joined: 24-December 05 From: The blue one in between the yellow and red ones. Member No.: 618 |
Works for me. Sunset and Lonesome Hill. Be nice to see realistic color.
Astro0 could do a "Spirit rides into the Sunset" composite. So could you, I'll wager. -------------------- My Grandpa goes to Mars every day and all I get are these lousy T-shirts!
|
|
|
Jun 4 2006, 02:25 AM
Post
#347
|
|
Member Group: Members Posts: 408 Joined: 3-August 05 Member No.: 453 |
The JPL update confirmed James' third alternative. The precision of the knowledge of the orientation of the rover was improved because they had many more sun pointing shots over time with which to refine the calculation. This issue is perhaps not as straighforward as it might seem - at least to me. Those Sun images are roughly 20 pixels across, and let's assume the centroid can be measured to within 1 pixel. Since the Sun's angular diameter as seen from Mars is about .35 degrees, the position of the Sun can in theory be measured to within say .015 degrees. This is of course assuming the rover's position is know (it is), and the rovers's clock and Pancam pointing accuracy are at least as good as the measurements require. Then, a series of measurement with the Sun near the zenith (to avoid refractive effects, which can't be that big a deal on Mars anyway) can be used to obtain the rover's pitch and roll, assuming a reasonable knowledge of its yaw. Yaw can not easily be measured this way - think of it this way, the Sun at zenith would look the same for all yaw values (assuming a level rover). Towards sunset a similar set of measurements can be used to refine yaw as even refractive effects would not affect the "sideways" position of the Sun is the sky. These new measurements would then be fed into the next cycle. Given the in theory very accurate measurements of the Sun's centroid, I presume then that it is a greater inaccuracy of the rover's clock and/or Pancam pointing accuracy that limits the rover's attitude accuracy and hence benefits from averaging over multiple solls? Well, a fascinating topic I think, and I'd be interested to learn more if anybody has any more information (as opposed to my speculation ) Airbag. |
|
|
Jun 4 2006, 11:05 AM
Post
#348
|
|
Dublin Correspondent Group: Admin Posts: 1799 Joined: 28-March 05 From: Celbridge, Ireland Member No.: 220 |
Then, a series of measurement with the Sun near the zenith (to avoid refractive effects, which can't be that big a deal on Mars anyway) can be used to obtain the rover's pitch and roll, assuming a reasonable knowledge of its yaw. Yaw can not easily be measured this way - think of it this way, the Sun at zenith would look the same for all yaw values (assuming a level rover). Airbag. True but: A: The rover is not level. B: The sun shots are taken at a variety of times of day not just one. C: In this case we have sun shots taken over an extended period of time and the suns overall track has changed significantly (at least relative to the accuracy of the Pancam). Its a very interesting topic all the same. |
|
|
Jun 4 2006, 02:04 PM
Post
#349
|
|
Member Group: Members Posts: 408 Joined: 3-August 05 Member No.: 453 |
True but: A: The rover is not level. B: The sun shots are taken at a variety of times of day not just one. C: In this case we have sun shots taken over an extended period of time and the suns overall track has changed significantly (at least relative to the accuracy of the Pancam). Its a very interesting topic all the same. Some thoughts on your comments: A: For the purposes of these hand waving arguments, I'd say 11 degrees is effectively level (as compared to the 90 degrees zenith angle). B: I suggested two sets of measurements should be sufficient. Looking at the positions of the Sun shots (using MMB's pano mode for sols 807-855), it appears the measurements fall into two clusters - one near the zenith, and the other about 20-30 degrees above the horizon, with some outliers in between. C: This I believe is at the heart of the issue, assuming the Pancam pointing knowledge is indeed less than the Sun's measured accuracy (in pixels). This link states the pointing knowledge is 0.1 degrees, which although about a magnitude larger than my centroid accuracy estimates above, is still much less than the almost 2 degrees attitude adjustment. So I'm still puzzled why the adjustment was so large. Airbag |
|
|
Jun 4 2006, 02:49 PM
Post
#350
|
|
Dublin Correspondent Group: Admin Posts: 1799 Joined: 28-March 05 From: Celbridge, Ireland Member No.: 220 |
C: This I believe is at the heart of the issue, assuming the Pancam pointing knowledge is indeed less than the Sun's measured accuracy (in pixels). This link states the pointing knowledge is 0.1 degrees, which although about a magnitude larger than my centroid accuracy estimates above, is still much less than the almost 2 degrees attitude adjustment. So I'm still puzzled why the adjustment was so large. Here's a useful link If I'm reading this correctly then in general 2 degree overall pointing was deemed sufficient as that is the accuracy required for ensuring that they HGA can be pointed at Earth so that the procedures used to refine the accuracy of the attitude could terminate as soon as the error was within that limit even though they have the capacity to be more accurate than that. One point that you made already that is probably worth investigating further is whether the clock timing accuracy (and whatever overall positional uncertainty there is for the location of the rover) could contribute to the error. I'm still trying to find a paper that I read at some point last year that explained the attitude calculation and sun shot processes in even more detail than the one above, I'll post a link if I track it down. |
|
|
Jun 4 2006, 05:23 PM
Post
#351
|
|
Member Group: Members Posts: 408 Joined: 3-August 05 Member No.: 453 |
Ah, thanks for that link - it made for fascinating reading. I guess it is the 'Sungaze" mode were were talking about then. I agree, it seems that HGA pointing needs drove the 2 degrees "is good enough" model. One of the sections mentioned that the attitude calculations go on until the "covariances converge to the 1.5 degree level"; presumably that means when subsequent estimates agree to within some function of pitch, yaw and roll that results in a delta of less than 1.5 degrees. BTW, did you see that mention of in theory being able to use the gyros as a gyrocompass to detect heading, if only the software supported it? (as a fallback if the Pancam was no longer working). It must have been very satisfying to have worked on this attitude/postion software and see such great results. Airbag |
|
|
Jun 4 2006, 09:57 PM
Post
#352
|
|
Member Group: Members Posts: 133 Joined: 3-June 06 From: the jungle of Nool Member No.: 799 |
1/2 size Preliminary Left eye McMurdo panorama:
everything from L2/L7 to L2...L7 images with lots of synthesized missing data rectangles. No attempt to match contrast / brightness / color between images except what Autostitch provides. Color by Horticolor: Do not look upon me; Lest with this piteous action you convert my stern effects: then what I have to do will want true colour; tears perchance for blood. |
|
|
Jun 4 2006, 10:13 PM
Post
#353
|
|
Dublin Correspondent Group: Admin Posts: 1799 Joined: 28-March 05 From: Celbridge, Ireland Member No.: 220 |
BTW, did you see that mention of in theory being able to use the gyros as a gyrocompass to detect heading, if only the software supported it? (as a fallback if the Pancam was no longer working). Yeah - it's a very neat set up - it's pretty impressive that you can have a solidstate closed unit that can give you a decent compas (and general attitude) bearing solely from the motion of the planetary surface you happen to be perched on. It must have been very satisfying to have worked on this attitude/postion software and see such great results. Personally I'd have said that about almost any part of these magnificent machines but this is one of the very impressive parts. |
|
|
Jun 5 2006, 05:58 AM
Post
#354
|
||
Senior Member Group: Admin Posts: 3108 Joined: 21-December 05 From: Canberra, Australia Member No.: 615 |
Works for me. Sunset and Lonesome Hill. Be nice to see realistic color. Astro0 could do a "Spirit rides into the Sunset" composite. So could you, I'll wager. A very quick version from your suggestion Shaka. Might be nice to do this in much higher resolution as better calibrated versions become available later. Cheers Astro0 |
|
|
||
Jun 5 2006, 07:37 AM
Post
#355
|
|
Senior Member Group: Moderator Posts: 4279 Joined: 19-April 05 From: .br at .es Member No.: 253 |
1/2 size Preliminary Left eye McMurdo panorama: Beautiful, as usual! PS: I wasn't aware you were using autostitch. Just in case, there is a thread with some tips on this nice tool. |
|
|
Jun 5 2006, 09:10 AM
Post
#356
|
|
Senior Member Group: Members Posts: 1619 Joined: 12-February 06 From: Bergerac - FR Member No.: 678 |
Hello,
Very nice AstroO, but, I think you should darken the pancam. Also, here is my own version of the McMurdo pan, only the images taken toward horizon. I reserved an other treatment to the rest of the pan. -------------------- |
|
|
Jun 5 2006, 03:52 PM
Post
#357
|
|
Senior Member Group: Moderator Posts: 3431 Joined: 11-August 04 From: USA Member No.: 98 |
|
|
|
Jun 5 2006, 06:44 PM
Post
#358
|
|
Member Group: Members Posts: 133 Joined: 3-June 06 From: the jungle of Nool Member No.: 799 |
Welcome, Horton! Here is a combined Autostitch/MMBview version of the Spirit Sol 855 Navcam pan. It looks a bit odd, but the advantage is that the horizon is properly levelled. I couldn't for the life of me get Autostitch by itself to level it properly. Er, thanks mhoward. I think autostitch must weight each join match equally -- so the near point errors introduced by non-panoramic images cause uncorrectable distortions in the panorama. A better scheme would be to weight points near the top of the panorama more heavily. Perhaps a user option parameter? 100 to -100? ( top to bottom weighting ?) I think that I will end up using PT Assembler for the "final" product. For those not aware, the rover camera images are NOT proper panorama images: The cameras do not pivot through the optical axis of the camera. For people interested in real panoramas with a lot of depth this is an important factor. Also also a uniform exposure and white balance is critical to a totally successful panorama. Again lacking in the rover images. But we do what we gott'a do and work with what we have. We go to the computer with the images we have and not the images we would like. |
|
|
Jun 5 2006, 07:35 PM
Post
#359
|
|
Senior Member Group: Members Posts: 4245 Joined: 17-January 05 Member No.: 152 |
Can anyone suggest what Spirit's up to with this curious navcam series?
|
|
|
Jun 5 2006, 10:40 PM
Post
#360
|
|
Senior Member Group: Moderator Posts: 2262 Joined: 9-February 04 From: Melbourne - Oz Member No.: 16 |
Can anyone suggest what Spirit's up to with this curious navcam series? CODE 860 p0060.01 40 40 0 0 0 80 photon_transfer_left_navcam
860 p0061.01 40 40 0 0 0 80 photon_transfer_right_navcam -------------------- |
|
|
Lo-Fi Version | Time is now: 19th March 2024 - 07:22 AM |
RULES AND GUIDELINES Please read the Forum Rules and Guidelines before posting. IMAGE COPYRIGHT |
OPINIONS AND MODERATION Opinions expressed on UnmannedSpaceflight.com are those of the individual posters and do not necessarily reflect the opinions of UnmannedSpaceflight.com or The Planetary Society. The all-volunteer UnmannedSpaceflight.com moderation team is wholly independent of The Planetary Society. The Planetary Society has no influence over decisions made by the UnmannedSpaceflight.com moderators. |
SUPPORT THE FORUM Unmannedspaceflight.com is funded by the Planetary Society. Please consider supporting our work and many other projects by donating to the Society or becoming a member. |