"Just drop your photos into PTGui and it will figure out how they overlap. "

[Velson Horie]

Most of my panoramas are imported and aligned as well defined grids with ca 25% overlap.

So I use the project>grid process.

1. in the latest 2x4 grid, I  loaded the files. For about 3 seconds on the panorama editor, the tiles were aligned as specified in the grid dialog.  Then 2 were placed at the far left and 6 at the far right, with acres of black space between them.  Obviously at this point, the program knows there are not going to be any control points on tiles which now were not overlapping.  So I had to add those manually, at which point the panorama took shape.

When I specifically called for creating CPs between tiles 1 and 2, about 30 were found.

2.  PTGui currently does not work out the control points on the adjacent tiles as specified.

Could you add a tick box or similar in the grid dialog box that automatically processes all specified contacting/overlapping tiles to create points?  

[John Houghton]

Velson, PTGui already has an option on the Control Points menu to generate control points for all overlapping images.  After running Align to Grid, you can simply use that option, assuming that the images have been reasonably well positioned.  How well they are positioned will depend on parameters entered into the Align to Grid tool and focal length setting.  If you are doing flat (mosaic) stitching using viewpoint correction, that should work.  Alternatively you can pretend that the images were shot with a long focal length (e.g. 1000mm) and stitch in the conventional way.  See FAQ at https://www.ptgui.com/support.html#6_5 .

John

[Velson Horie]

John

Thank you, but

I have a 2*5 mosaic, aligned to a grid with a 1500 mm focal length and I have clicked to align images and  " generate control points for all overlapping images"

See attached screen capture

PTGui recognises the grid, identified that images 9 and 10 are above image 1, then created the relevant CPs.  But it does not recognise that 2 is next to 1, or that 9 is above 2, or 3 next to 2 etc.  So the "panorama" proposed is a random arrangement of separated tiles on a black background, which is disconcerting.

It is not too much a of a faff to go to each pair and generate CPs for each.

However it seems as though the grid arrangement is not being fully fed back into CP tool.

[John Houghton]

Velson, The problems may be due to the nature of the artwork that you are copying.  The standard control point generator doesn't work well with line drawings or etchings.  It's not possible to see from the small images exactly what type of subject you have here.  However, I attempted to simulate what you are doing using similar crops made from a photograph using Photoshop.  Needless to say, they stitched perfectly using either a short focal length lens + viewpoint correction, or a very long focal length without viewpoint correction.  If you use the Align to Grid tool to position the images (or simply drag them by hand roughly into position in the Panorama Editor window), then running the control point generator (for all images) followed by the optimizer should work ok.  And once the images are reasonably well aligned, you can generate more points if necessary using the "Generate control points here" option.

John

[Velson Horie]

John

My problem might be that my early panoramic successes were with MS ICE. This allowed me to predefine the position of >1000 tiles  on a grid, which ICE used as the starting point of analysing, no matter what the content of the tiles were - completely empty, lined or with useful features.

I keep on assuming that "Align to Grid" would be the starting point of the arrangement of tiles. But the predefined grid appears to be completely ignored in the initial random placement of the tiles in Editor window space and in the subsequent analysis.  Is there a way that the grid could be prioritised and, only when the placement on the grid is established, that CPs and overlaps should be looked for and used to tweak placement of tiles within the grid structure?

I changed my image capture from using a macro lens (1000 tiles) to A3 scanning (15 tiles) which are manageable to rearrange, but 1000 wouldn't be.

Velson

[John Houghton]

On Tuesday, November 18, 2025 at 7:56:20 AM UTC Velson Horie wrote:

I keep on assuming that "Align to Grid" would be the starting point of the arrangement of tiles. But the predefined grid appears to be completely ignored in the initial random placement of the tiles in Editor window space and in the subsequent analysis.  Is there a way that the grid could be prioritised and, only when the placement on the grid is established, that CPs and overlaps should be looked for and used to tweak placement of tiles within the grid structure?

Velson, If you run Align to Grid with appropriate parameters, you should have the roughly aligned image showing in the Panorama Editor window.   You can then either generate control points and run the optimizer, or run Align Images.  Exactly what the Align Images operation does is defined on the Project Settings tab (available in Advanced mode).  Clearly, if you have an already roughly aligned image showing in the Panorama Editor, you don't want PTGui to attempt to "roughly arrange the images" and thereby possibly undo the good work of Align to Grid.  Either way,  creating control points should assign points on the image overlaps and the optimizer will attempt to improve the existing alignment using the control points.

John

[Velson Horie]

John

It seems I am doing something wrong. This is what I get with a 3x8 grid after aligning.  It was more or less like this before aligning. Once again, the grid specification that image 1 is next to 2 has been ignored.  But I am reassured by the comment from the Optimiser that the maximum CP distance is 9.5 pixels, which is very good.

[John Houghton]

Velson, Your Align to grid parameters should be something like this:

Your shooting configuration parameters should match the ordering of your images as loaded.  Mine were two rows starting at bottom left, L to R, with approximately 20% overlap.  Just loading in the images and setting focal length to 1000mm, and then running Align to grid gave me this in the Panorama Editor window:

I was then able to generate control points and optimize to give a nicely aligned panorama after using Centre panorama and Fit panorama from the edit menu on the Panorama Editor window.  The optimizer figures are : Average cp distance 0.57 and max distance 2.21.

John

[Velson Horie]

John

Thank you.

The scans were taken zig-zag. In this case, image 12 was correctly placed above image1, and CPs were automatically generated.

This is the panorama generated by applying in the shooting configuration box the zigzag option

This is the panorama generated by applying in the shooting configuration box the unidirectional option 

In neither case is any grid created from the images supplied.

Is there  another button that needs to be chosen to implement the grid tool?

[John Houghton]

Velson, Have you by any chance selected viewpoint correction for the images?  There might be vp parameters left over from a previous run.  If so, set the viewpoint option to Reset, for all images before running Align to grid.

John

--
You received this message because you are subscribed to the Google Groups "PTGui Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ptgui/cfbf7ebe-acb3-45dc-86ef-b90dcce58e3fn%40googlegroups.com.

[John Houghton]

Velson, On second thoughts, you would have to run the optimizer to reset the vp parameters, or set all of the vp parameters to 0 on the Image Parameters tab.

John

[Velson Horie]

John

Bingo! Yes the vp parameters were all active.  set them all to 0 and everthing went into place.

Thank you for bearing with me.

[John Houghton]

Velson, Excellent news!  You don't need to use viewpoint correction when you specify a huge focal length.

John