Why renderImage in PDFBOX cuts off the background and leaves only the OCR layer?

Question

Java

When rendering an image via PDFBOX. The layer with the image itself is cut off from the file and only the OCR layer remains. When removing the OCR layer via Adobe Acrobat, pdfbox returns a blank white sheet. What can be done to render an image like a screenshot?
How it is extracted:

As it should: