Image overlap

Discuss your experiences with and ideas about Stardust@home here.

Moderators: Stardust@home Team, DustMods

Post Reply
Franz
DustMod
Posts: 23
Joined: Fri May 19, 2006 6:27 pm
Location: Netherlands

Image overlap

Post by Franz »

How large is the overlap between adjacent images?
If an impact trail is located near the edge, is it possible to examine the neighboring images?

Take for example the tutorial view #10. The trail is quite large and it is highly imaginable that the trail spans multiple images, something like 4 images if the the corners of the images would lie near the center of the trail. One might need all the imagary to motivate a conclusion.

Bosniac
Posts: 22
Joined: Fri May 19, 2006 4:05 pm
Location: Sara-is-an-ox; B&H

Re: Image overlap

Post by Bosniac »

Franz wrote:How large is the overlap between adjacent images?
If an impact trail is located near the edge, is it possible to examine the neighboring images?

Take for example the tutorial view #10. The trail is quite large and it is highly imaginable that the trail spans multiple images, something like 4 images if the the corners of the images would lie near the center of the trail. One might need all the imagary to motivate a conclusion.
Well u might only be asked to identify the trail of possible particle in that image (although I don't "see it at the end of its track" in view #10, but I suppose its in upper left corner)
You see things; and you say Why? But I dream things that never were; and I say Why not? (B.S.)

Franz
DustMod
Posts: 23
Joined: Fri May 19, 2006 6:27 pm
Location: Netherlands

Post by Franz »

Perhaps I miss-formulated my question.

I find the tutorial to be slightly misleading as the presented particle trails lie well within the edges of the image. What if a particle trail is smack on the edge of an image? Say that one suspects this to be the case, how can one 'navigate' the VM to that position so that the candidate trail is better centered within an image?

If the scanning microscope constructs seamless images, and these 'raw cuts' are distributed between volunteers, isn't there a greater chance that a particle trail is overlooked if it is on a seam?

I would assume that there is an image overlap to prevent this happening, so my question is: how much is this overlap?

Second question: If there is insufficient overlap in the scanned imagery, how can one stitch neighboring images to examine the seams?

After all, it is a VIRTUAL microscope, so it should (in my opinion) be able to stitch multiple neighboring physical images and present a virtual view of what I am directing at. One should be able to navigate in a X-Y-Z direction, not just depth as the tutorial suggests.

Before the tutorial I had expectations with regard to the navigating abilities of the VM. The tutorial signals that navigation is (highly) restricted. One should have a versatile tool if one is expected to wander in undiscovered territory.

Bosniac
Posts: 22
Joined: Fri May 19, 2006 4:05 pm
Location: Sara-is-an-ox; B&H

Post by Bosniac »

Yes that is a good question. Will we be looking at a "zone" (several images in square formation, so that we can scan adjacent "images") or just some random images.
You see things; and you say Why? But I dream things that never were; and I say Why not? (B.S.)

Sharqua
DustMod
Posts: 245
Joined: Thu May 18, 2006 1:02 am
Location: Bradenton, FL

Post by Sharqua »

Very good question, Franz. I am interested in the response.

Your English is exceptional, by the way.

Regards,

-Donna

Emilie LeBlanc
Posts: 4
Joined: Fri May 26, 2006 8:07 am
Location: Dieppe, New Brunswick, Canada
Contact:

Post by Emilie LeBlanc »

Very good points. I too am also interested in the response to these questions. :)

Chronocide
Posts: 9
Joined: Mon Jun 05, 2006 12:29 am

Post by Chronocide »

I remember reading about overlap and, luckily, it didn't take me long to find the article again. Here is a link and an excerpt.

Link to Planetary Society Article

Here's how it will work: as in the original plan, the automated microscope will scan the entire surface of the collector, recording digital images of each miniscule portion of the aerogel. Since each image will cover an area of 260 x 340 microns, and since each image will include a 10% overlap with its neighbor, the microscope will need to focus on 1.6 million different locations to cover the entire surface of the collector.

My take on an object like #10 cutting the corner of "my" image, which means its start and end points will be on two other images, is that Stardust at Home is just one research project. After we find 40 or so micron sized dust particles they are not going to just chuck the rest of the aerogel in the trash bin. Even if we miss it, somebody will find #10 someday. It's all good.

Post Reply