I have been working on trying to assemble Sojourner's color images and to clean them up to the point of being presentable. I have posted an image or two in the past, but I now have completed the whole set for Sols 64-81 (Sol 56 would be the next date with images, working backwords). during this time, Sojourner was in or near the rock garden. These images really show how Sojourner could have gotten stuck if it did indeed circle the lander after the lander died.

Very nice, Ted

Phil

Here, here Ted!

I've been thinking that much more could be done with the Pathfinder/Sojourner images...especially with all the talent we have here.

Thank you, these are fantastic.

I assume you separated the images into separate bands and geometrically corrected them to eliminate the horrific levels of radial chromatic abberation in the camera. Did you do any special work to improve the mosaic-sampling of the in-camera color pixel mosaic?

Yes, I separated them into bands and realigned them. I constructed the color image from binned images, which were used as a color overlay, to cut down on the problems caused by the mosaic sampling. Otherwise you had funny slivers of color that no amount of realignment can fix.

Here is the next set - Sols 37-56. This set really shows what an obstacle rocks were for Sojourner.

Ahhhh!

Ted, when are you going to get all of your wonderful images published in book form? Could you perhaps produce something via Lulu.com?

Ian.

Do you think that MRO would be able to spot Sojourner from orbit? It would interesting to see how far Sojourner continued after contact was lost....

Probably. The advantage it has is that its color capability might make Sojourner stand out.


As for book form, while I have thought about it, I haven't even found time to build a new website lately. Maybe someday.

Here is another set.

Groovy!

Phil

It really amazes me that nothing was ever done with these images in terms of cleaning them up and assembling the mosaics for public release. Especially images like those from Sol 21. This is part of a bigger project of mine using the Pathfinder dataset. With the 10 year aniversary coming up, I am hoping to have some finished products soon. In terms of the distance covered and being so low to the ground, I am tempted to call the Sojourner set "A Crawl On Mars" - No offense, Nix.

I must admit, I thought the color camera on Sojourner was useless. Nobody would ever get anything out of it. Ted has really done a great job with this.

Phil

I started working with the set in hopes of coming up with an approximate color overlay from the Sol 37 data for some of the black and white cameras. During this, I figured out how to break the images into color channels and align them. Then, I came across some black and white rear camera mosaics, and finally understood how the images fit together. After toying around, I realized that these images might not be hopeless.

Here is the next installment - The images after Sojourner left Yogi, but before Sol 18.

Excellent results. I've tried to do something with these images but never got anything close to your results.
Perhaps I'll just try to reproject some of them and correct for vigneting.

It is more of a resizing issue and realigning issue than a reprojecting issue. One thing I make sure to do is clean up salt and pepper noise and bright, monochrome bars on the left side of some of the frames that wreck havoc on cleanup efforts. Also, to be frank, some of the images are a lot better than others.

In case you (the reader) might think "so what is so special about these fuzzy images?", compare Ted's efforts with one of the very few rear camera color images on NASA's Planetary PhotoJournal and you'll see just how much of an improvement they are, especially in terms of color registration and balance.

Airbag

I couldn't help but revisit these having seen your great results. The Red and Blue channels seem to line up reasonably - just the green channel that's so warped....but looking at the green channel in photoshop, it has a trace of one of the other channels which I can't seem to cull out.

DOug

The green channel is godawful. Another problem is the stretching of the images. You will have wonderful mosaic, but in one of the frames, there will be a bright glint of light off part of the rover, which makes the rest of the image appear nearly black. I am still working on the next set, but here is a preview, a shot of Yogi from Sol 8.

Thanks for the images, Ted. Those initial releases of the Sojourner images were gosh-awful. And it shows how far rover-eyes have evolved since then...

--Bill

Yup - there's no Pathfinder IMP or Sojourner 'RAD' equiv PDS release ( which is real pity ). Are you going straight from the images on the PDS sojourner page, or are you using the IMGs?

Go on....what's your green-channel secret...I think you're still using it because I tried a synthetic green and it doesn't quite match with your results

Doug

I am using the IMG files. The whole Patfhinder set is sloppy. Sometimes the labeling is plain wrong. I think it had mostly to do with a lack of funding for archiving the data.

As for the green channel, I manually resize to fit. Then I warp it to correct for any areas that are still out of alignment. I align the mosaic and do the best correction for brightness variations before I realign the colors. For every red pixel, there are six green pixels. Same goes blue.

So the green channel has far better sampling than the other two channels, and so it is extremely damaging to try to work without it. I use a process I first used on Voyager images with vibration problems to correct somewhat for the ghost images. In some shots, it works really well, sometimes it doesn't.

Here are the remaining images. There is also an image taken on Sol 2, but it is essentially blank, so I did not include it.

Here is the full Sojourner color set.

Excelent! If I'm not mistaken, there are some larger images, but overexposed, specially on the red channel. Did you try to do something with those? Perhaps you did and I'm just mistaken :-)

These are all the images taken by the rear camera. Perhaps you are thinking of images from the front cameras?

Click to view attachment Click to view attachmentClick to view attachmentClick to view attachment

I've checked. These are the images I'm talking about, in my sort of "color raw" version. This is as far as my processing as gone with this images. I can't seem to find them on your set. Perhaps you've cropped the image and this is what's been confusing me!!!

OK, I found them in my set - I had somehow overlooed them. I am not sure what I can pull out of these...

I just downloaded the files from the net!
http://pdsimg.jpl.nasa.gov/data/mpfr-m-rvr..._rear/index.htm

The preview files online are gifs and I worked from those!

I've tried to reproject some of your rover images and it looks interesting. I'll post the images when I have time

I had used the gifs from the color section, and hadn't realized there were a few images not in both sections. The gifs are fine to work with. Since you are dealing with 8-bit data, you don't have problems that you would have with other missions.

While you did some great work with these images, Ted, they're still really poor images -- always have been. How anyone involved in MPF thought they were going to get any significant scientific or engineering information out of those cameras is beyond me.

-the other Doug

They were more for navigation and to get the APXS targeted than for producing pretty pictures. The real problem for Pathfinder all around was a low transmission rate, due to the lack of an orbiter, and also a low bandwith link between Sojourner and the lander. Hence, the images from the mission are small, compressed, and, in the case of Sojourner, often fragmentary. Another problem, especially for the rear camera, was that they ran out of time, and never got to finish calibrating them. Still, it is humanity's first crawl on Mars, and it is a goal of mine to get what I can out of the images.

Very good point, Ted. It's vitally important to know what the APXS is sitting on when it makes a measurement. For that you don't need high resolution. So Sojourner could look at a target with its front cameras, twirl round and drop the APXS on the target, and fire off a little rear camera image to check it was really on the target.

Phil

And it's the first Color Matrix CCD used on Mars!!! Speaking of that, any idea if a IR cut filter was used? The colors see so strange on some images... and the level of chromatic aberration... seems consistent with a color matrix ccd without a IR cut filter in place.

It's really puzzling why such a camera was chosen... why not just put a BW lowres camera in the back of the rover, like the Huygens probe? I'd trade resolution for bit depth ;-)

I think it had a lot to do with the clock running down. They ran out of time to develop and calibrate the camera. The idea of a color camera on the rover seemed good, but budget/schedule/data rate issues got the best of it.

I processed the four remaining images. The overexposure is a big problem. I finally understand why they chopped up all but these four images. Apparently, the dynamic range wasn't wide enough to send back a full frame with all three colors without major over or under exposure.... What a horrible camera. I reduced the size of three of the frames since they did not contain sharp enough data to make it worth keeping them full size.

(evil evil grin...)

Now you can go back and restore the Explorer 6 and Vanguard 2 spin-scan images of Earth

I wish....I have never been able to find any of that data, save a single image from Explorer 6.

(grins)
Well.... Ya.
Probably the same cruddy <makes Sojouner color pics look like Imax...> pic I've seen in the Explorer 6 NASA SP series science results volume.

I recall vanguard 2 was bumped by the last stage during separation and kiked into a nutating wobbly spin, making reconstruction of the crude spin-scan images impossible. May have been the origin of the recognized need for nutation-dampers in spin stabliized spacecrzaft.

Yes, it was. I wonder, though, if modern computers could reconstruct the data.

Here is the set with the "emergency mode" images that have serious exposure issues.