If you open the bdf file with notepad (might need something else because of unix style line endings - I use PFE, but wordpad usually works) you will see the file header then the individual characters, each character has its own header (name, width, offset, etc) then the bitmap data in hex.
Each row will have a set of hex digits 0-F, convert them to binary (use windows calculator if you want - each charactrer is 4 bits '0'=0000, '1'=0001, '2'=0010 ... 'E'=1110, 'F'=1111). Then as Lear points out the easiest way of visualising the character is to represent 1s with '#' and 0 with '.' (and use a fixed pitch font to view it);
You might find that there is not enough data. this will be because the character header can have some offset type values to give the inter-character space, others have it explicitly.
Sorry to be a bit vague - I don't have the source here to have a real font to look at and I didn't have time last night - have a play and see if it starts to make sense.
Edit: I assume you have a build environment and the source (to get from bdf to Rockbox format) so you should look for a utility called something like bdf2bmp () - it has a mode where the whole font is dumped in .# bitmap format - that will make decoding bdf easier - I'm still not near my PC with all that stuff so I'm guessing the name