1

Topic: Vector Displacement Map

Hello:

How can I keep the length of the vector arrows as a field in the feature attribute table of the exported shapefile?

Thanks,

ElSayed Hermas

2

Re: Vector Displacement Map

Dear ElSayed,

I'm not sure I understand your question. What do you mean by "keep the length"?

In the vector field generation tool, you can change the arrow length in the menu: Processing -> Tool GUI, then use the "ratio arrow length" variable for proper display (it depends on the size of your display and the length of the arrows). The scale-bar displayed at the bottom is for informative use only as it is not very accurate. The arrow scale and density of the arrow display is preserved when you export the arrow field as a n ENVI vector layer. Let me know if I understand your pb correctly.

Sebastien

3

Re: Vector Displacement Map

Dear Sebastien:

I thought I can obtain the length of each arrow in the attribute table of the envi vector layer that has been produced from the vector field?

ElSayed

4

Re: Vector Displacement Map

Dear ElSayed,

This option is not available. You can however obtain the length of a displacement vector by computing its Euclidean norm using "Basic Tools -> band math" and input "sqrt(b1^2 + b2^2)", b1 and b2 being the EW and NS component of the displacement field. It will produce an image representing the displacement.

If you use the Euclidean norm computation to quantitatively determine some physical parameters, be aware that this method can overestimate the "true" value of the displacement because the noise on the Euclidean norm is not Gaussian, but rather has a Rayleigh distribution. In short, unless you see a clear directional trend in a given neighborhood of vectors, the length might be dominated by the noise level (this is easily seen because a length is never negative). This case will mostly happen if you are examining small displacements (less than a few meters).

Sebastien

5

Re: Vector Displacement Map

Dear Sebastien,

I am an undergraduate student new to ENVI and COSI-Corr.  I am trying to obtain a velocity field of the Rupal Glacier but I'm having a few difficulties.  I created a vector field of the glacier successfully but I am having some problems with how to determine the velocity of the displacement arrows.  I generated the displacement vector using the Band Math and the input you provided but I still can't figure out how the velocity and arrows correspond to create a table for my velocity field.  Thank you!

-Tony

6

Re: Vector Displacement Map

Hi Tony,

I'm not sure I understand your question. The velocity is simply the distance the glacier traveled divided by the time separation between the acquisition of your two images. The distance can be determined by the Euclidean norm of the EW and NS component, and the time depends on the images acquisition times. All these can be done using the ENVI "Band math" tool.

It looks like you're confusing the display of the vector field with the computation of the velocity, which you have to do yourself. Displaying the displacement field is just that: a visualization, it's not a computation.

Sebastien

7

Re: Vector Displacement Map

Hello all,
My query is: with the reference of the paper "Remotely sensed dune celerity and sand flux measurements of the world's fastest barchans (Bodele,Chad)" (Vermeesch, Drake,2008), how do we make a Euclidean distance map using both East-West and North-South displacement maps? And once this map is prepared, how are the attributes visualized?and how are the pixels <0.1m selected in ENVI in order to define interdune areas ? As I am using Cartosat-1 (2.5m) data for 2011 and 2006 for Thar Desert, India, what are the things to be kept in mind while processing the data?Also, is NL Means filter necessary?If so, at what step?

I am new to ENVI and Cosi-corr, so I apologize for the plethora of questions at one go..
Lydia

8

Re: Vector Displacement Map

Euclidean norm = sqrt(EW^2 + NS^2). The NLM filter is not necessary.

Sebastien