Post

2 followers Follow
0
Avatar

Overlap and Angles, Single grid vs. Cross Hatch

Overlap.  On a cross hatch mission, does MPP take into account the overlap produced by the perpendicular passes or just the overlap in one direction and you basically end up doubling it?  I think that is the case.  If so, do you have any idea of what reduced overlap settings would result in an overall overlap to a certain setting.  In other words, do you know what would you set for overlaps on a cross hatch in order to end up with an overall 80% overlap taking into account both direction paths?

Gimbal angle.  For precision topography, not 3d modeling,  is the data showing that introducing a slight oblique angle on one of the passes of a cross hatch is beneficial.  Or is nadir on both passes still better?

Thanks!

Dave Pitman

Official comment

Avatar

The overlap settings tell how a single image will overlap with its neighbors. 80% overlap means that any image will overlap with 80% of the next image. 60% of the next. 40% of the next and 20% of the next. If you do this in both axes along track and across track you will get 25 looks at everything. 

The Grid mission settings doesn't change this. So if you do a grid mission at 80% overlap you will end up with 50 looks at everything (which is too much). The goal should be to get around 25 so it is safe to use a lesser overlap setting of roughly 65-70% overlap when doing a grid mission and you will still get a reasonable number of looks at everything. 

It is funny you ask about the gimbal stuff because we were just having this discussion yesterday with someone. Must be something out on the internet these days... Mixing one pass of oblique images has been proven to be useful in linear mapping with high quality cameras. The northern pass was taken facing back across the rest of the passes. Every 4th image was included. It turned out shockingly well. It was with a P1 though so that might have something to do with it. 

Aside from that we don't have much data or inkling to support the statement that mixing in a single oblique pass changes much. 

 

 

Zane
Comment actions Permalink

Please sign in to leave a comment.

4 comments

0
Avatar

Got it.

One other quick question.  When I look at radial distortion plots of the images captured with the P4P, the area of highest distortion is on the left and right perimeter and lesser at the top and bottom. Do you have any thoughts on the idea of capturing a high side overlap, like 90%, and then cropping the images in on the sides to greatly reduce that area of distortion from the images?

Dave Pitman 0 votes
Comment actions Permalink
0
Avatar

This is one of the reasons you have to use so much overlap. Much of the image isn't really that great. You can get away with less overlap with better quality cameras. 

You CAN crop the sides off the images as long as you are careful with the EXIF and XMP data and are sure to treat the whole dataset identically. 

Zane 0 votes
Comment actions Permalink