It has been a while since I worked with this, but I think these are the read IDs velvet is using internally.
In the output folder there should be a file called "Sequences". It is basically a fasta file with the input reads. The fasta header should look like this:
> M00298:13:000000000-A1BJE:1:1101:17340:1510 2:N:0:1 6 1
It starts with the original read ID followed by the vlevets internals read ID (6 in this case). I do not know at the moment what the last number means.