Reading thru lots of little PDF's I've found, i get a figure that mentions up to 12 Gbits for a single full HiRISE image.....BUT - look at this...

Click to view attachment

To me, it would seem that they might get as few as one full HiRISE image per day

BUT - at the best data rates of say, up to 6Mbits/sec - 12 Gbits is only 34 minutes of downlink.

Can anyone sync those two facoids? It seems that the suggested performance, and the expected return, don't match by an order of magnitude.

Doug

For starters, the maximum data rate that is being assumed is more like 4 Mbps. For an 8-hour pass, the s/c is only in Earth view about half the time, so two passes a day at 4 Mbps would return about 115 Gbits assuming no time to lock up, no retransmits, etc. So the graph seems roughly correct.

But the max data rate is only doable when the Earth-Mars distance supports it, which is why the graph goes up and down. And yes, at the low data rate periods you shouldn't expect to see many full-size max res images from HiRISE.

I suspect that most HiRISE images will be summed down from the maximum resolution, both to save volume and because the HiRISE MTF may not be all that great at full-res.

The HiRise website says a feature of the mission will be "Hundreds of full-resolution 20,000 x 40,000 pixel images". Over the 700 day primary science mission, that should mean roughly one a day. The site also says "Thousands of high-resolution sub-scenes with 1 to 2 meters/pixel content", so we should expect roughly 10 or more images with 2x2, 4x4 or 8X8 binning a day. Plus some 4000X40,000 pixel swaths in 3 colour.

I think scientists will have to make a good case for why they need a full 6x12 km chunk of Mars at 30 cm resolution. Because such images won't be coming down by the dozen every day.

I guess it helps to think of MRO as a spy satellite. You have to pick your targets carefully.

May not be that great?.....huh Do you know someting we dont lol

Perhaps somebody should consider the benefits of a dedicated Mars telecommunication orbiter...

(ducks and runs)

Bob Shaw

No, just look at the diffraction limit. HiRISE has a 50 cm aperture and is sampling 30 cm from 300 km. MOC has a 35 cm aperture and is sampling 140 cm from 400 km. All other things being equal, you can't get 3.5x better resolution at the same image quality by increasing the aperture by less than 2x. (HiRISE has 12 micron pixels and MOC 13 micron, but that only helps HiRISE a little.) HiRISE will have better image quality than MOC, certainly, but not 3.5x better; it's not physically possible.

Perhaps they'll combine the two, downsample and subsample

i.e. take the full thing, but send the full width at 1/4res, and the central 4k pixels at full res.

I've seen fairly confident mention of 5mbps, and one of the reports I mention elsewhere in this subforum talks about 6mbps. Obviously, I appreciate this will only be at certain ranges, and only with larger DSN assets, and only some of the time. I guess once you factor all that in, and the requirements of the other instruments and so on and so forth - a couple of images a day makes sense.

It does seem a bit mad, however, that only something like 70,000 sqkm or so will be imaged at that res - for a planet with as much land as Earth, it's an area just the size of Lithuania.

I guess a lot of it will be monochrome, probably 2x downsampled ( good for SNR ), and thus approx 50-60cm/pixel. It's going to be good stuff, but what's MOC up to now, 200k images +? It'll take a long time for MRO to match that

Doug

You're assuming the MOC camera is diffraction limited; the material from the HiRISE team implies the limit on MOC is the low signal to noise ratio resulting from scanning single pixels for the extremely short exposures needed for high resolution. They plan to improve the signal to noise ratio by clocking multiple pixels in the scanner to integrate the signal as the image drifts across the scanner. "Each CCD has 2048 12 x 12 um pixels in the crossscan direction and 128 TDI elements (stages) in the along-track direction." The 128 lines of time delay and integration (TDI) are used to get a very high (100:1) signal to noise ratio.

For more details see the HIRISE Instrument Development Report.

PS
Just saw the following in the last page of the report: "The nominal high resolution image is 20,000 pixels by 40,000 lines and can take from 4 to 48 hours of transmission time depending on range to earth and compression factors."

Steve

Because it is, pretty much; at least as closely as an R-C Cassegrain system can be. As for your point about HiRISE having better SNR; that's certainly true, although the TDI will degrade the MTF at Nyquist, especially if the spacecraft attitude control has any jitter in it at all. I stand by my original assertion: HiRISE will have better image quality than MOC, but not by as much as a simple ratio of their ground sample distance would suggest.

I think you're all comparing apples and pears - size of data product vs optical resolution!

Bob Shaw

Ahh - BUT - if the actual resolving power of HiRISE turns out to be only, say, 60cm/pixel - then onboard downsampling can produce a data product 1/4 the size, so we can have 4x as many of them

Doug

I think we spoke at cross purposes when I said that MOC was not diffraction limited. I was not questioning the optical system, I meant to say that the resolution of the system was not limited by optical diffraction but by something else (pixel size). A bit of digging finds that the optical diffraction limit leads to a resolution on the ground of about 70 cm/pixel. (See http://www.msss.com/mer_mission/finding_mer/. ) In standard-mode scans, the pixel size is 1.5 m x 1.5 m, much less than the diffraction limited resolution. cPROTO mode came close to reaching the diffraction limit, oversampling along track at 50 cm/sample while the across track resolution was still limited by the pixel size at 1.5 m.

We'll see whether HiRISE lives up to its promises when the images start to come down. In that regard, has anyone reported on the resolution of the lunar and stellar test images?

Steve

The diffraction limit is often treated as some kind of an absolute limit to the resolution of an optical system, but this is not really true. What it defines is the minimum distance where two objects can be definitely separated. To get an idea about what an object is, it needs to be several times larger than the diffraction limit. On the other hand it is possible to determine the position of small isolated objects (e. g. stars) with a precision that is much higher than the diffraction limit, since this is limited by the precision with which you can measure the position of the centre of the diffraction pattern.

tty

Question: in the normal mode of operations, will there be any need for MRO to point significantly off the nadir point (like at the horizon)? I ask because I was fooling around with a SPICE kernel that evidently is a dry run of the science phase, covering Dec. 2007. At one point that I saw, The HiRISE/CTX boresight swings up and above the horizon, pointing into space. Could just be a what-if, who knows.

MRO will routinely point off the nadir to image selected targets, but not as far as the limb. From the MRO website:

"The orbiter will typically keep its science instruments pointed to nadir (looking straight down at the surface). A few times per day, and for about fifteen minutes each time, the orbiter will point side-to-side in order to capture high-priority science targets that don't fall directly beneath the spacecraft. The spacecraft can point off-nadir up to 30 degrees."

There were some problems with the attitude simulation in some of the test files; perhaps this is what you are seeing.

That might explain why the pointing went way above the horizon. I was trying to make an animation of a mapping orbit & wasn't expecting to see that much of a swing. Makes for a dramatic effect though. Here's the (flickery) animation I got: PSP sim animation (8MB MPG) (using your maps as texture BTW).

I think a good strategy could be send the entire swats at binned/lower resolution, hold data in memory, select interesting regions and then ask to MRO to send back full res data only for these cropped areas.
However, i do not know if timing and onboard memory will allow this, do someone knows?

I would not be surprised at all if they didn't take near-limb images on occassion. Hubble takes near-limb-shots of Earth. A narrow angle image taken at extreme slant-range on Earth and Mars are pretty decent "flat-field" calibration images.

If they got the horizon in the picture and could remove the pesky martian atmosphere (perhaps a giant Hoover is in order?), that would be a really cool picture!