Acrobat text conversion turns some pages upside down

I am scanning a very old book with lots of old cursive handwriting in it. Process is like this: 

photograph pages with dSLR

edit pages in LR, export as tif, clean up in PS

Import images as pages in ID 2023 using image directory to CSV script and data merge tool

Export PDF from InDesign 2023

Open PDF in Adobe Acrobat  Build: 23.1.20093.0 on Mac Pro

Use Scan and OCR tool, recognize text

The scan and OCR tool/recognize text reduces the file size for 100 pages from 250MB to 57MB, but some of the pages get turned upside down. I've tried it twice now. Out of 50 test pages, several get turned upside down in the PDF. Has anyone ever seen this?