How do I create a mask to specify where features should be automatically detected during image co-registration?

Hello, I have a question about the capabilities of the Image Registration App. I haven't yet upgraded to the right version to get this app, but I'm considering it.

I am trying to calculate glacier flow velocities using images from cubesats. These images only have a ~10 m geolocation accuracy, so before velocities can be calculated using feature tracking, I need to coregister the images. I assume the Image Registration App is the way to go for this kind of thing. The problem is that the coregistration needs to be based on features detected only in the rocky regions around the glaciers, because the surface features within the glacier areas are expected to shift between images. So I need to create some sort of mask to tell the software where to look for features, and then transform the entire image based only on those features, interpolating to correct the glacier areas. Is masking a capability that the Image Registration App has? Or is there some other tool that I could use? I assume I could manually choose control points, but I would prefer to use some sort of automated detection.

Thanks!

[Chad Greene]

Hi Karen,

Is it possible that you can do the template matching on the unregistered images, and then use a rock mask to set known zero velocities accordingly?

[Image Analyst]

Can you attach a pair of images? And indicate what are the regions that are unmoving and need to be registered? Are you able to get a binary image of what pixels should be used in the registration? If so, attach that also.