Melbourne Councils and Archives Race to Fix Duplicate Image Problem in Digital Collections This Week

Victorian archivists and local government library staff spent much of this week manually reviewing thousands of digitised photographs after a systematic problem with duplicate and mislabelled images surfaced across at least three major Melbourne collections. The issue, which has been quietly building since a rushed digitisation push between 2021 and 2024, came to a head when the City of Melbourne's library services flagged hundreds of duplicate entries in its Ozimage holdings earlier this month.

The timing matters. State and local governments across Victoria have been funnelling significant resources into digital-access projects, partly to meet updated Public Record Office Victoria guidelines introduced in late 2025 requiring councils to provide searchable online access to heritage photographic records by mid-2027. Errors at this scale, left uncorrected, would undermine those compliance targets and potentially skew historical research built on those collections.

What Went Wrong — and Where

The core problem is straightforward: when multiple vendors scanned physical photographic collections at different times, images were ingested into databases without consistent deduplication checks. The State Library Victoria on Swanston Street has been aware of overlap between its Pictures Collection and donated council archives since at least early 2025. This week, staff at the Yarra City Council library network — covering branches in Richmond, Fitzroy, and Collingwood — began a parallel audit after identifying roughly 340 images that appear in both their own digital catalogue and an externally managed repository managed under a Victorian Government digitisation grant.

The Fitzroy Library on Moor Street has been a particular focus, given that its collection of late-nineteenth-century neighbourhood photographs was scanned under two separate grant programs — one through Creative Victoria and a second through the former Regional Library Digitisation Fund — with neither program's contractor cross-referencing the other's output. Staff there have been working through the backlog since Tuesday, according to internal council communications circulated to heritage advisers this week.

The problem is not confined to photographs. The Melbourne Catholic Archdiocese Archives in East Melbourne flagged a smaller but related issue with scanned parish records in June, where batch-upload errors produced duplicate file references across different parish collections. That institution is handling its review internally and expects to resolve the indexing errors before the end of July.

The Practical Cost of Getting It Wrong

Duplication is more than a cataloguing inconvenience. Storage costs for public digital archives are real and ongoing. Amazon Web Services pricing for the S3 storage tier most commonly used by mid-sized Australian cultural institutions runs at approximately AUD $0.025 per gigabyte per month — modest on its own, but multiplied across tens of thousands of high-resolution archival image files, redundant storage adds up to a non-trivial budget line. Beyond cost, researchers relying on unique identifier numbers to cite digitised sources face citation integrity problems if the same image carries two different reference numbers in published academic work.

Public Record Office Victoria, which sets standards for how councils manage and preserve records, published updated guidance in March 2026 specifying that duplicate image entries in publicly accessible digital repositories must be resolved prior to any new ingestion from external collections. That requirement is now creating a bottleneck: at least two Melbourne councils have paused planned digitisation projects while they clear existing duplicates to stay compliant.

For anyone working with Melbourne's digital heritage collections right now — researchers, journalists, genealogists, or documentary filmmakers — the practical advice from archivists is consistent: cross-check any image sourced from an online council or library portal against at least one secondary catalogue before citing it, and note the unique file identifier, ingestion date, and collection name in full. The State Library Victoria's catalogue at slv.vic.gov.au currently allows side-by-side comparison of metadata fields, which can help flag when two entries describe the same physical object. Councils expect most audits to wrap up by September, after which the affected databases should be cleaner and more reliable than they have been at any point in the past three years.