1

Topic: Glacier volcano interaction

Hello!

My name is Pablo Zenteno, Geographer and research assistant at the Centro de Estudios Científicos in Valdivia, Chile.

Fisrt of all I want to sincerely give thanks to all the COSI-Corr team for their valuable help!

We are a research group studying the interaction between glaciers and volcanoes in the southern Andes and we are also starting a research in Peninsula Antartica.

I have recently  processed two ASTER images with only one month apart (14 November and 16 December - 2007).

Surface features in the central morraines are very visible and with the co-registration it is very clear the displacement between one month.

After correlation of the images I haven't applied any NL filter yet, can you please give some advice for this.

I have displacements of 0 to 44 m for the 32 days between the two images.

Best wishes

Pablo

2

Re: Glacier volcano interaction

Dear Pablo,

Thank you for your support and interest in COSI-Corr smile

In order to make the most out of your measurements, I would suggest:

1- Try to identify if your correlation results are affected by wave artifacts, which are due to the ASTER platform oscillations. To do so, remove all data with low SNR (<0.9), and with large and unphysical displacement (use the discard value tool). You can also crop all values outside +/-15m, it will enhance the dynamic range of possible artifacts. I will refer you to this publication:
http://www.tectonics.caltech.edu/slip_h … SE2008.pdf

If you observe wave artifacts, you can apply a destriping procedure, COSI-Corr now offers an automatic way of doing it. Typically, you want to remove most of your signal due to the process to be imaged (crop the values to some low displacement to retain most of the artifacts while discarding most of the useful signal), generate the destriping model from this image, and then apply it to the raw correlation results. Let me know if my explanations are not clear.

2- Once raw correlation images are destriped, make sure low SNR and large unphysical displacement values are discarded.

3- Apply the NL-Means filter. We were planing on writing more about our specific implementation to give more guidance, but we still haven't found time to do it, sorry about that. It will require some trial and error, but try to start with the default parameters. Select the EW and NS band. If low SNR values have been discarded already, you don't really need to select the SNR band (you can try but it won't change the results much). The only parameter you really need to focus on is the "noise parameter". The larger this number, the stronger the denoising. It is realted to the noise standard deviation, and for your application, I would guess that "H" should probably be chosen between 2-4. You want H to be small enough so that filtering doesn't blur the measurements too much, but large enough so that denoising still occurs. The next parameter you can play with is the search dimension. You can probably try dimensions from 20 to 50 pixels. Other parameters are discussed in the user's guide. You can also experience with them, but if the basic parameters don't produce satisfying results, the more advanced ones won't help very much. Correlation images can be corrupted with many outliers, and unfortunately, the NL-Means filter will not help with these. Only expect the NL-means filter to smooth out the results, while nicely preserving the structure of the displacement field.


Cheers,
Sebastien

3

Re: Glacier volcano interaction

Dear Sebastien,

thanks a lot for the literature, it has been a lot of help! :-)

I'm now on the first steps you told me, .

I was doing some experiments at a calving glacier
"glaciar Jorge Montt" in the  Southern Patagonian Icefield (-48.35S/-73.47W) using ASTER images with near one year of difference and taking in account a displacement of  nearly 500 m per year.

For the correlation I used the following parameters:

Frequential Window Size: 64-Step:4 -Mask Treshold:0.900000 -Nb Robust Iteration:4
and No Resampling

After destriping and discard values the final vector field really improved and shows consistent values with the ice flow from the main tongue and also from tributary glaciers.

Still I have to try tomorrow the NL - filtering.

Some other questions on my mind:

After the correlation I don't see visible striping on the data. Is this good or am I doing something wrong? Maybe I'm don't understanding this very clear: generate the destriping model from this image, and then apply it to the raw correlation results.

What implies the "Mask threshold" value ? I was looking at Leprince et al. 2007 in IEEE but I feel really lost with that.

Thank you very much again Sebastien.

PD:I feel very enthousiastic with COSI-Corr :-) and I will encourage some colleagues doing netectonic in Chile.

4

Re: Glacier volcano interaction

Dear Pablo,

1- The mask threshold in the correlation helps to discard frequency components with low power. The phase of low power components is mostly random and it's usually beneficial to discard it. The lower the threshold, the more frequency components are discarded. A value of 0.9 should work well in almost all cases.

2- You can use the correlation tool with the multiscale approach: the correlation window should be at least twice as large as the largest displacement to be measured. However, when you enlarge the correlation window, you are going to lose some spatial resolution and the displacement field will be more blurry. To avoid this problem, you can use several window sizes, for example 64 down to 32, or 128 down to 32, etc... COSI-Corr first correlates with the larger windows, then re-center the correlation according to the first result and recomputes the re-centered correlation with a smaller window, until the smallest specified window is reached. Therefore you can measure large displacements and still maintain the spatial resolution given by smaller windows. For your case, I would assume that a correlation with windows 64 -> 32 pixels should be optimum. Do not use windows smaller than 32x32 pixels, results will become meaningless with all 8 bits quantized images (e.g., ASTER, SPOT, etc...).

3- Striping artifacts usually have an amplitude around +/- 10m. If glacier displacements are larger than that, the striping will be saturated and the images won't have enough dynamic to show the stripes. To check for stripes or waves artifacts, you have to crop all the values that are outside the +/-10m or +/-15m range (in short, you have to discard most of the signal from the glaciers). You can use the "discard value" tool if you want to save the results, or you can adjust the display "Enhance -> interactive stretching" in the ENVI display, and reduce the range, then click apply.

4- In the destriping tool, there is a field "image to define correction from". This is the file you're going to input to determine the destriping model. In your case, the destriping is a simple averaging of the residual displacements along the across-track of the sensor. Assume you have glaciers in the correlation with large displacements. You certainly don't want to average the signal of the glacier to build the destriping model. Thus, you can build a correlation file where glaciers are masked out. An easy solution is to discard all measurements that may represent some signal by cropping large values. This is what is represented in Fig. 4 of Scherler's paper (note that glaciers appear white because they have been cropped out). Then COSI-Corr determines the global bias, and this global bias will be subtracted from the original measurements. Therefore, you should input the original correlation file in the field "image to apply correction". Note that you don't always need to differentiate between the image used to derive the correction, and the image to which the correction is applied, depending on how your glaciers (or other processes) are distributed in your image. For another example of destriping, you can look at Fig 1 of this paper:
http://www.tectonics.caltech.edu/slip_h … SS2007.pdf

Hope things make more sense. It's a very easy process, but a little tedious to explain, sorry:)


Cheers,
Sebastien