Clipping boundary for referenced raster image rotates on its own

[Bender_B_Rodriguez_III]

I am having a problem with the rectangular clipping boundaries for some raster images I am referencing, or rather the frames of the clipped images. Whenever I save and close the drawing file, then re-open it, the clipped image frames align themselves with the axes of the raster images, instead of staying aligned with the original polyline rectangles I selected in the drawings as clipping boundaries. It's driving me nuts, but no matter what I try I cannot keep it from doing this. See attached screenshots to see what I mean - the red arrows show the rectangular polyline I am using as a clipping boundary, with the "before" shot showing what it looks like when I close the drawings, and the "after" showing what they look like upon re-opening. 

The images I am referencing/clipping are aerial photos which are slightly skewed off of the orthogonal compass directions, whereas the drawings themselves (and therefore the rectangular polylines that create the borders for the clipped images) are aligned to the compass. So whenever I re-open the drawings, the clipped image frames have mysteriously rotated to align with the skewed axes of the aerial photos, and I have to re-establish the clipping boundaries again. I have been using IMAGECLIP to do this, which has worked fine on many other drawings of exactly the same nature, with the same set of aerial photos, and which have stayed put in all the previous drawings. Note that the images themselves do not rotate or move, only the clipped image frames do; the actual images are still where they should be, but they have frames which are skewed to the rest of the drawing after re-opening. Further, the re-aligned image frames still hold at the SE and NW corners of the rectangular polyline border, which seems even odder yet.

The only thing different about these new drawings is that they are rotated 90* around the Z-axis, so that North is to the left on the drawings instead of towards the top of the sheets (I had to lay them out this way to align the project area with the sheet size, etc.).  After I performed the Z-axis rotation, I set USCFOLLOW back to "0" and then changed back to World UCS, in order to be able to use real coordinates. So the UCS axis in the drawing is Y (north) to the left and X (east) towards the top of the monitor.

If it was an issue with the UCS Z-axis rotation I'd have expected to see the clipping boundaries rotated 90*, but not the screwy 1.7* angle it is that the axes of the aerial photos are skewed off of North. So I am stumped. Anyone else have any experience with this?

I'm running plain old ACAD 2018 on a Dell PC with plenty of RAM, etc., no issues there, and am working only in Model Space - no P-space Viewports.  

[Bender_B_Rodriguez_III]

Further weirdness....... Going back and monkeying around, trying different things, I used the "Polygonal" option to define the clipping boundary, instead of the "Select Polyline" option I'd used previously. And then saved and re-opened. And nothing shifted or rotated or moved - it looked the same on re-opening as it did when I saved and closed it, and worked just as one would expect it to. So I suppose that's a reasonable work-around, but still I am wondering why it keeps screwing the clipping boundary up when defined by "Select Polyline"? The mystery continues.......

[yangguoshe]

you  may  draw  an  orthogonal rectangular   first， while running imageclip command， new border  /select  polyline

[Bender_B_Rodriguez_III]

That’s how I drew the original boundary to clip the image with – a orthogonal rectangle. The Polygonal work-around seems to work ok, using the rectangle as a guide for the Polygonal clipping outline, but I’d like to know why it just doesn’t work as it should in the first place 😊