Ballarat's cultural institutions are working through one of the less glamorous consequences of two decades of mass digitisation — enormous backlogs of duplicate images embedded in public collections — and the approaches being tried here offer a sharp contrast to what peer cities in Canada, Scotland and New Zealand have attempted.
The problem sounds mundane but carries real cost. When a regional gallery or heritage trust digitises a collection, the same photograph can end up filed under multiple accession numbers, scanned at different resolutions on different days by different contractors, then ingested into a database that has no automatic way to reconcile them. Staff time spent identifying and resolving those duplicates is time not spent on public programs or loan requests.
The Local Picture
At the Art Gallery of Ballarat on Lydiard Street North — one of the oldest and largest regional galleries in Australia, holding a collection of more than 6,000 works — digital asset management has been a running priority since at least 2019, when the gallery undertook a cataloguing overhaul tied to a broader Victorian Government regional arts investment round. The gallery has not publicly disclosed how many duplicate image records it has identified, but the challenge is consistent with what similar-scale institutions report across the sector.
The Sovereign Hill Museums Association, whose open-air museum on Bradshaw Street draws hundreds of thousands of visitors annually, maintains a separate photographic archive documenting more than 60 years of site reconstruction and living history programs. Sovereign Hill's education and collections teams have described — in publicly available annual reports — the ongoing work of reconciling physical and digital holdings as the site's archival footprint has grown.
The City of Ballarat itself, through its Libraries and Customer Experience directorate, holds digitised records from the former Ballarat City Library collection, some dating to the 1880s goldfields era. Duplicate image records arise there partly because the same historical photograph was donated by multiple families over different decades, each copy arriving with different provenance notes and each scanned independently before the collections were merged.
What Other Cities Are Doing
Dunedin, New Zealand — a city of roughly comparable size and with a similarly strong identity built around heritage tourism — committed in 2023 to a structured deduplication program at Toitū Otago Settlers Museum, using open-source perceptual hashing tools to flag visually similar images for human review rather than automated deletion. The distinction matters: automated deletion has caused irreversible losses at several North American institutions where near-identical images turned out to document subtly different historical moments.
In Stirling, Scotland — another gold-heritage and tourism-driven regional centre — the Stirling Council Archives adopted a vendor-managed solution in 2024 under a contract reported at approximately £180,000 over three years. The approach offloaded technical work but drew criticism from local archivists who argued it reduced staff capacity to make contextual judgments about which duplicates carried independent historical value.
Canadian mid-sized cities including Fredericton, New Brunswick, have leaned on federal digitisation grants through Library and Archives Canada to fund in-house deduplication roles — a model that has no direct equivalent in Victoria's current grant architecture, where arts funding rounds tend to reward new public programming over back-end collection maintenance.
That funding gap is part of why Ballarat's institutions have moved at a slower pace than some international counterparts. Victoria's Regional Arts Fund and Creative Victoria's Place-Based Initiatives have supported digitisation projects, but collection maintenance — including deduplication — has generally fallen outside eligible expenditure categories. Ballarat's institutions are not unique in navigating that constraint; it is a structural feature of how Australian arts funding is designed.
The practical consequence is that deduplication work here tends to happen in concentrated bursts when a staff member has capacity, rather than as a continuous program. The Art Gallery of Ballarat is understood to be one of several Victorian regional galleries in early-stage conversations with the Public Record Office Victoria about shared standards for image metadata that could make future deduplication faster and cheaper — though no formal agreement has been announced. Institutions watching that process should expect any shared framework to take shape no earlier than late 2027, given the typical pace of cross-agency standards development in Victoria. In the meantime, the most practical step available to smaller local collections is adopting consistent file-naming conventions at the point of scanning — a low-cost change that archives internationally credit with reducing duplicate creation by a significant margin before any cleanup is ever needed.