Ballarat's Digital Archive Faces a Familiar Global Problem: What To Do When the Same Image Appears Twice

Ballarat's cultural institutions are confronting a problem that has quietly consumed thousands of staff hours at museums and galleries from Glasgow to Guadalajara: duplicate images embedded deep inside digital archive systems, distorting catalogues, inflating storage costs, and muddying public access to heritage collections.

The issue has sharpened in 2026 as regional bodies across Victoria face renewed pressure to digitise physical holdings faster — partly driven by state government commitments to expand public access to cultural collections — while simultaneously managing the sprawling legacy databases those earlier digitisation rushes left behind.

What's Happening Locally

At the Art Gallery of Ballarat on Lydiard Street North, collection management staff have been working through a backlog of duplicate entries generated during successive digitisation rounds stretching back to the mid-2000s. The gallery holds more than 6,000 works, and digital records for some pieces exist in multiple catalogue entries — different scan resolutions, different metadata tags, different upload dates — creating confusion for both internal researchers and members of the public using the online collection portal.

The Ballarat Heritage Office, which coordinates documentation across the city's historically significant precincts including the Eureka Centre precinct and the Sturt Street civic corridor, faces a related but distinct challenge. Photographic surveys of streetscapes and building facades have been commissioned repeatedly over the years by different consultancies, often producing overlapping image sets with no systematic deduplication protocol in place. The result is redundant files occupying server space and complicating search results for heritage practitioners and planners.

Sovereign Hill, the open-air museum on Bradshaw Street that draws visitors from across Australia and internationally, runs its own extensive photographic archive for interpretive and marketing purposes. Staff there have described the problem in general terms as a known operational challenge — not unique to Sovereign Hill, but common across living museums of comparable scale globally.

How Other Cities Are Handling It

The comparison with similar-sized heritage cities is instructive. Ballarat's population sits at roughly 120,000 people, placing it alongside international counterparts like Stirling in Scotland and Ouro Preto in Brazil — both cities with significant historic built environments and active digitisation programs. Stirling Council's digital heritage unit completed a deduplication audit of its photographic holdings in 2024, reducing redundant image files by an estimated 34 percent according to public reporting from Historic Environment Scotland. Ouro Preto, working through Brazil's IPHAN heritage agency, embedded automated hash-matching tools into its ingest workflow as part of a federal digitisation grant finalised in early 2025.

Both approaches cost money upfront. Automated deduplication software licences for a mid-sized institutional collection typically range from around A$8,000 to A$25,000 annually depending on collection volume and the vendor, based on publicly available pricing from platforms such as ResourceSpace and NetX. Manual audits, by contrast, can run far higher in staff time — a single person reviewing a collection of 50,000 images at a careful pace may need six to twelve months to complete the work.

Regional Victorian institutions have generally lagged behind metropolitan counterparts partly because the State Library of Victoria's digitisation support programs have historically prioritised Melbourne-based collections. That has begun to shift. The Regional Public Libraries Infrastructure Fund and related programs administered through Creative Victoria have signalled broader eligibility criteria for 2026-27, though specific allocations have not yet been publicly confirmed.

For Ballarat, the practical path forward involves a combination of approaches already proven elsewhere. Institutions like the Art Gallery of Ballarat can apply perceptual hashing — a technique that identifies visually similar images even when file names differ — without replacing existing catalogue software entirely. Several open-source implementations exist at no licence cost. The harder task is governance: agreeing across the Ballarat Heritage Office, the gallery, and tourism bodies like Sovereign Hill on a shared metadata standard so that future digitisation work doesn't reproduce the same problem a decade from now.

A joint working group involving representatives from the City of Ballarat's cultural services unit and the gallery's collection team would be a logical first step — one that cities like Stirling have found worth the coordination effort, even when progress is slow and the meetings are unglamorous.