mmajurski
commented
3 years ago Can you clarify what you mean by "The stitched sets end up being different lengths"?
If you are stitching two different grids of images, I would expect that subtle differences in the input image would produce stitched images of different width and height. At the very least by a few pixels.
If you need the exact same image size, you could try assembling the second set with the translations computed from the first set.
Finally, if you acquired the images at two different times, no physical system is perfect, so I don't expect the tiles are going to be in the same 'localtion' with pixel level accuracy. There will often be single digit pixel shifts between images (if the stage is well calibrated) and larger shifts with less well calibrated equipment.
I have 2 sets of 6 images that were taken at the same locations at different times in an experiment.
Trying to stitch together each set and subtract the images as rasters. The stitched sets end up being different lengths, even if I set the horizontal overlap to be the same percentage for both sets (which they theoretically should be). Image_set_1.zip Image_set_2.zip
More details about my setup in MIST:
Filename Pattern Type: Row-Column Starting Point: Upper Left Grid Width: 6 Grid Height: 1 Column Start Tile: 4 Row Start Tile: 1 Timeslices: 0
Filename Pattern: img_r{r}_c{c}.tif
Blending Mode: Linear Subgrid: Use Full Grid
Horizontal overlap: 60% (estimated) Translation Refinement Method: Single Hill Climb Stitching Program: FFTW