MRO image release 25th April 2007

regards
Robert

I particularly like the craters on south polar layers one - especially the pattern of winding furrows.

I still have not got myself set up to deal with these gigantic images. Apart from a bit of zoomifying, all I ever do is look at the browse versions.

Phil

Notice the trio of small new craters on smooth plains in the NW part of "Craters on South Polar Layered Deposits" - two large ones barely touching, and a tiny "Proxima Centauri" crater at about 2 o'clock...

It would be cool to survey all these fresh polar deposits for recent asteroid hits and determine the ratio of "solos" to binaries and multiples (as this one presumably is).

You gain so much more from looking at bits of the full-resolution though... Still, I do know what you mean, it can be difficult to look at these large images...

After trying and discarding a number of software tools for viewing these giant images, I finally stumbled upon "IAS Viewer" which works great, though it is a bit of a pain to download (you have to create an account and answer questions); still, it is free.

http://www.ittvis.com/download/download.as...lproductcode=93

The thing I like about IAS Viewer is that it doesn't attempt to work with the entire image all-at-once in memory, but does a multi-resolution thing and only reads as much detail as can be seen at the current zoom level and screen resolution - which means it is using Jpeg2000 the way it was meant to be used.

FWTools does the same sort of thing I think...seems ok with 2 gig of ram.

Doug

I had mixed luck with FWTools. It worked great on my computer at work, but would only work with small images on my home computer (which has double the memory of the one at work).

Hirise folks should concentrate getting the zoomified images released at the same time as each weekly image release. To be honest, I see little point in providing lossless images of these sort of resolutions. Most of us just want to look at a few interesting spots in each image at full resolution. Lossy compression would be sufficient for that and the zoomify interface is the perfect interface.

Though suggestions (like that marvellous 3D of Itokawa) for small asteroids being barely-bound collections of larger lumps might confuse things at this sort of crater scale. Was the body a true binary, or did it simply fragment in the atmosphere? Which leads me to wonder just how much can a body's components separate in the few seconds before hitting the ground?

Andy

Whereas personally, I hate zoomify. It's slow, painfull, useless if you want to look at a large area at full res. I would rather they focused on getting the full size images out so I can look at as much or little of it as I want to. The JP2's are a product of the workflow - the Zoomify stuff involves further work over and above that.

You have to remember that HiView - the online app for viewing the images at full res in a slightly zoomify way - is a work in progress and will essentially provide what you're talking about when it's finished.

The short answer: We don't have the resources to do zoomify for everything. In fact, it's not even clear how much longer we will even be able to do the zoomified images. Our colleagues at NASA Ames have been doing this, and I think they will continue to do this for the weekly captioned images, but I have big doubts about them being able to keep up once the dam breaks with our first PDS release in June. In short, these are not, nor will they ever be our official release products.

However, things are looking pretty good for us to be able roll out decent jpip/jpeg2000 support coinciding with our first data release to the PDS in June, although this isn't set in stone yet. We will be testing this very thing next week some time to see if we can really do it for the PDS release. I would say that we have a pretty good chance of making it work, but we'll see. Once you've seen jpeg2000 working via the jpip protocol, you will not *ever* want to mess with zoomify again. It really is that much better.

But for most people wanting to view images from MRO, "Zoomify" is the only practical way of doing it.

I understand where you are coming from and let me assure you, that this is will not be the case once we have our jpip servers running. We will be providing suitable client software for one thing. What's going on behind the scenes with the jpeg2000 stuff is *very* different. The image format is such that it is saved with multiple resolution levels built right into it, and as you zoom in and out, you will get only the amount of data that can be displayed in your window at the selected resolution level. Which makes it possible to essentially do everything that zoomify does and more with out the lossy compression that zoomify introduces, and without having to waste disk space on the image pyramids that zoomify needs.

That being said, I'm sure there are those people that will still like zoomify better for whatever reason, unfortunately, it just isn't practical for us to support it. I do think that most people will be pleasantly surprised when they see the jp2's working in conjuction with client software that knows how to work with these images properly, though. It really is quite impressive.

Thinking about the way I would want to use images myself, the methods described here are or will be great for browsing. But the one special function I will really need will be the ability to outline a rectangle on a low res version and save that area at full resolution for use in some other software. I hope that will be part of the new software.

Phil

FWTools does that perfectly with the JP2's if you can't wait for HiViwe Phil.

Saving of subimages is supported, although I think it is currently limited to the part of the image that is visible within the view window (but I could be wrong about that). Our initial release will probably support IASViewer as opposed to our own HiView software, athough it looks like there is some good progress being made with HiView. It's difficult to say at this time if a version of HiView will be made available at the same time as our PDS release.

As for FWTools, yes, FWTools, works essentially like the new client software, however, the difference is that FWTools doesn't support the jpip networking protocol, at least not as far as I've been able to tell. This means that in order to view a file using FWTools, you still have to download the whole file first before you can view it. The jpip protocol allows you to only download the portions of the image you are interested in at the resolution that you want, so you can avoid having to download 2.5 Gigabytes of data to view a single image.