Imagemagick seems to be very violent with memory usage. A suggestion is to use -limit to limit the memory size convert will use, it should use a disk cache for anything else it needs. Read an explanation here:
http://www.imagemagick.org/pipermail/magick-users/2002-March/001285.html
And here's how to use the -limit option : http://www.imagemagick.org/script/command-line-options.php#limit
If not, you can do it like you describe, converting each image individually and then merging them with some other tool. Something like
for name in p*.png; do convert $name `basename $name .png`.pdf; done
should do it.
Basically it iterates over all the png files, runs convert on them, the only magic is the basename $name .png part, which will be replaced by the filename minus the .png part, so that when attaching .pdf extension to it, the filename makes sense.
Try it for yourself so you get a feel of what basename does:
basename something.png .png
echo `basename something.png .png`.pdf
Best Answer
for the second part of your question I would recommend djvusmooth.
for the first part of your question I would suggest pdfchain should cover your requirements
both are available from the Software Centre.
I personally use a command in terminal to extract text:
this will export the text of the pdf to a file meaning you can edit the text in a text editor.