Preparing cDNA libraries for high-throughput sequencing might be your best bet to get a comprehensive view of piRNA populations. There are some qPCR based approaches, but they can be tricky and only measure the abundance of one piRNA at a time.
As input you can use total RNA, depleted of ribosomal sequences, or RNA from immunoprecipitations of Piwi-clade Argonaute proteins. It all depends a little bit on your model system and organism.
There are kits available from the big suppliers, such as Illumina Truseq. I'd recommend to use more hands-on protocols such as:
piRNAs can be methylated at their 3' ends. At the moment, you have to either use NGS or prime an RT reaction with a panhandle adapter and use a Pi specific primer and a universal primer. Tailing, like we use in the miScript system will work (we've done this) but only on the piRNA that is unmodified at the 3' end, which is some unknown fraction. Not recommended.