1

Topic: spot 6

Hello there

I want to use Cosi-Corr Software to compute displacements accrued along a glacier using SPOT-6 imagery.

I have read software manuals and found that I should have level 1A Panchromatic products.

but unfortunately I don't have it and instead I have got two Spot-6 ortho images.

Is there a way to process these two to provide a good result with quality and precision?

I directly went to Correlation step but there is no suitable and well resolution results.

I was wondering if everyone would be so kind as to help me.
Thanks

2

Re: spot 6

Hi Anna,

if you have already orthorectified SPOT-6 or any other image for that matter, which I suspect won't be on a same grid/same resolution, you'll need to put them first on a common grid. The easiest/fastest way to do this is to use the ENVI layer stacking tool (Basic Tools -> Layer Stacking). Select the two orthos, keep everything as close as the original images (same projection, resolution...). Use bilinear interpolation.

Francois

3

Re: spot 6

Hi dear Francois

Thank you for your reply.

I have done layer stacking as you said. but there is no meaningful difference between new and old results.

I have a question, can we get same results as level 1- A by using ortho images?

or is it better to find and work with level 1- A products?
Thanks

4

Re: spot 6

Anna,

ok, sorry I misunderstood your question. You correlated the SPOT-6 orthos and your output map contained a lot of noise and/or did not have any interesting signal, and you were asking if there was a way to improve the results.
At that stage of the processing, apart from testing different correlation parameters, in particular increasing the window size if the correlation map is noisy, there is not much to do. Experiment with that and see if that improve the results.

Regarding improved result that you could get with level-1A instead of orthos, I cannot say. May be the signal you're after is too small for the SPOT-6 resolution or may be the change between the images is too strong and the correlation fails. In these cases, the 1-A format won't help. The most likely situation for which 1-A would help is if the problem comes from uncorrected topography or misregistration.

Francois