People mean many things by "CNV"- what size of event are you looking for and what resolution? You can use Cortex to assemble structural variants - it's best in the range SNPs-->few kb, although it does have a reference-based method for calling large (tens of kb) variants (mostly deletions).
De novo assembly and genotyping of variants using colored de Bruijn graphs. Z Iqbal(), M Caccamo(), I Turner, P Flicek, G McVean, Nature Genetics (2012)
latest paper, on microbes:
High-throughput microbial population genomics using the Cortex variation assembler. Z Iqbal, I Turner, G McVean, Bioinformatics 2012
Cortex is set up to allow the user to specify ploidy and is definitely not "tuned" for human. If you are interested in very large (tens or hundreds of kb) duplications then Cortex is not able to call these I'm afraid. (I'm one of the authors I should add.)
In addition, this recent paper (nothing to do with me) looks interesting
De novo detection of copy number variation by co-assembly
Jurgen F. Nijkamp1,2,3, Marcel A. van den Broek2,3, Jan-Maarten A. Geertman4, Marcel J.T. Reinders1,3,5, Jean-Marc G. Daran2,3 and Dick de Ridder1,3,5,*
the paper uses their new tool on haploids (yeast).
There are a lot of other CNV tools, but my experience of those is limited to human data, so i can't really comment.