1

Topic: Glacier displacement with low SNR

I used the SRTM  DEM(resampled from 90m to 30m)as the first reference image and computed the displacement using windows from 128 to 32, step 4 pixels, time span from 5 months to 4 years. Strangely, the SNR of glacier area is very low so when discard the low correlation area the glacier is removed from the displacement result(Fig1, the ortho and SNR images geographic linked). I followed the process and method in the paper and forum, but they didn't show any solution. Can someone give me some advice? Thanks million for your help.
PunBB bbcode test

2

Re: Glacier displacement with low SNR

What are you correlating, what are your two images?

Sebastien

3

Re: Glacier displacement with low SNR

Sebastien_Leprince wrote:

What are you correlating, what are your two images?

Sebastien

Thanks for your kind reply.
1. ASTER L1A data: 20010430/20060428, 20010430/20010905
2. I generated shaded DEM using SRTM, then orthorectified 20010430 using shaded DEM as the reference and SRTM DEM as the elevation. Subsequently, orhto-image 20010430 orthorectified 20060428 and 20010905 using SRTM DEM.
3. Ortho-image 20010430 correlated ortho-image 20060428, so did 20010430 and 20010905.
4. The SNR is low(<0.5) in glacier area in both correlation results.
I think it isn't due to long time span(such as 5 years) and different acquiring season of data, because all of the correlated results have the same pb more than just the above two pairs.

Cui

4

Re: Glacier displacement with low SNR

What happens if you correlate using windows of 64-32, or simply 32-32 pixels? 128-32 may be a bit brutal.

Use 4 robustness iterations, if the SNR is below 0.9, discard the values. For glaciers, it's usually best to be at most 1-2 years apart.
Feel free to send me the 2 ortho-images you're correlating if you'd like me to have a quick look, you can zip them and use this dropbox:
https://dropbox.caltech.edu

Sebastien

5

Re: Glacier displacement with low SNR

Sebastien_Leprince wrote:

What happens if you correlate using windows of 64-32, or simply 32-32 pixels? 128-32 may be a bit brutal.

Use 4 robustness iterations, if the SNR is below 0.9, discard the values. For glaciers, it's usually best to be at most 1-2 years apart.
Feel free to send me the 2 ortho-images you're correlating if you'd like me to have a quick look, you can zip them and use this dropbox:
https://dropbox.caltech.edu

Sebastien

Hi, Sebastien, thanks for your patience. I have sent my data to you. It will help you understand my pb.
1. What's the meaning of different initial window size, such as 128*128 and 64*64(I couldn't figure it out from the forum and COSI-corr guide)?
2. Your suggestion(window size from 128-32 to  64-32) generated a better result. The  percent of pixel values SNR>0.9 come from 20% to 28%, but the mean values of the whole displacement image from 9.7 to 2.8. Which one should I believe?
3. Can I apply the displacement result in other research without comparing with in situ data and just compute the residual offset using the method in this paper(Scherler,2008 RSE)?
Thanks a lot.
Jiangpeng

6

Re: Glacier displacement with low SNR

Yes, got your data, thanks.
In theory, the maximum displacement that can be measured is 1/2 the correlation window. Therefore, in theory, if you have a 32x32 pixel window, you can measure up to 16 pixels of displacement, or 16*15=240m with ASTER. In practice however, this bound is a lot lower due to noise, multi-temporal changes, etc. Therefore, rather than hoping for a 1/2 ratio, we're usually happy when we can attain a 1/5 ratio. Meaning measuring ~5-6 pixels of displacement with a 32x32 pixel window.
It's the reason we implemented the multi-size windows. Often times, you want to be able to measure larger displacement, but still would like the advantage of smaller windows in terms of local averaging. Therefore we start with a larger window to get the large first order displacement, then we refine with a smaller window. That's what the 64-32 or 128-32 means. We do 2 passes, the first pass with a larger window, the second with a smaller one, re-adjusted according to the previous measurement.

For your case, you have to acknowledge that your images are not ideal. For the images separated by less than 1 year, you shouldn't expect a large displacement, therefore 32x32 pixel windows should suffice. However, the images from the same year aren't from the same season and you can notice a lot of differences in snow cover. That clearly constitutes noise that will damage the correlation.
On your image pair from 2001-2006, there are other problems. First, note that the main glacier is covered by clouds in some areas, so you can't expect the correlation to work there. On another area, you see the strong shadow cast from the clouds. You can't correlate there either.
Then, there are more problems: the displacement to be measured is quite large with respect to the width of the glacier. Remember, to measure large displacements, you have to use large windows. The problem, is that at the resolution of the ASTER image, a 64 pixel window will likely be larger than the width of the glacier. The correlation therefore becomes more difficult as the displacement is not coherent anymore within each correlation window.
You can see that on the faster parts, the correlation tends to give results going uphill on the glacier. This makes no sense, and it simply means that the texture on the glacier completely disappeared and cannot be traced anymore. By chance, the correlation finds other structures on the glacier that looks like what was there before, but it's obviously wrong.

My guess for these images: use a 64-32 scheme for the longer term pair, and 32-32 for the shorter term image pair, with a step of 2 pixels and robustness of 4 pixels. Then apply a directional filter, in the same fashion as Scherler et al (you'll have to do that, say in Matlab or IDL, etc.). Meaning you will manually define a direction for the glaciers you are interested in (eg central flow line), and you discard displacement vectors that strongly disagree with this a priori direction of flow.

You would get much better results if you could find images without the clouds on the glaciers, and with no more than 3-4 years between pairs.

Sebastien

7

Re: Glacier displacement with low SNR

Hello, Sebastien, thanks very much for your detailed explanation.
Because of the bad weather, the five ASTER images I have bought are the best images in NASA's inventory from 2000 to 2012 year, that is 20010430, 20010905, 20060428, 20070821, 20100525, so there is not more images to use.
Your reply is very helpful to me.
Thanks again.
Jiangpeng.

8

Re: Glacier displacement with low SNR

You're welcome. Maybe one last thing: when you create your ortho images, you can select "from raw images" for every image, and make sure you click the "multiple to image resolution" check. That way, all your images will be ortho-rectified completely, and you won't end up with large black parts in your ortho images. The correlation tool will find the correct overlap directly. The images you correlate don't have to have the same size (as long has they have the same resolution and grid multiple of the resolution).

Sebastien

9

Re: Glacier displacement with low SNR

Sebastien_Leprince wrote:

You're welcome. Maybe one last thing: when you create your ortho images, you can select "from raw images" for every image, and make sure you click the "multiple to image resolution" check. That way, all your images will be ortho-rectified completely, and you won't end up with large black parts in your ortho images. The correlation tool will find the correct overlap directly. The images you correlate don't have to have the same size (as long has they have the same resolution and grid multiple of the resolution).

Sebastien

OK, I will follow your advice in later work.
I wish you all the best for your career.
Jiangpeng