I think it's worth evaluating [prinzfrank's pdf parser](https://github.com/PrinsFrank/pdfparser) as an alternative to smalot/pdfparser. It's got image handling built in, and seems to be more actively developed at the moment.