Whereas I'd be stunned if there were anything like that out there because it's just not a trivial problem and has no single answer. Two objects representing database rows might have different instance ids but refer to the same row so they couldn't be deduplicated without inspecting meta information for the table and perhaps consulting the live data, for just one example. What constitutes a duplicate is part of the design of the object not just and not necessarily the data. Objects in Perl can be regular expressions, arrays, etc, too. Would one deduplicate /\w/ and /[^\W]/ or are they different? :P Deduplication may or may not need to consider metadata or the constraints of relationships to other data or objects or environment or temporal variables.

Simple hash based objects without any "fancy" stuff might be trivial depending on the information design and it sounds like you're already on the right track for that.