Melbourne Takes a Hard Look at Duplicate Images in Public Archives — and How It Stacks Up Against Cities That Moved First

The problem sounds mundane until you actually try to fix it. Melbourne's major cultural institutions collectively hold tens of thousands of digitised photographs, maps and documents — and a growing share of those records are exact or near-exact duplicates, filed under different catalogue numbers, sometimes in different collections, sometimes in different buildings. Getting rid of the wrong copy, or merging records incorrectly, can erase decades of curatorial context.

The issue has moved up the agenda in 2026 partly because of a Victorian Government push, flagged in last year's Creative State strategy refresh, to consolidate the state's fragmented digital heritage infrastructure before the end of the decade. That consolidation work has forced institutions to compare their holdings directly for the first time — and the duplication rate is proving awkward to discuss publicly.

What Melbourne's Institutions Are Actually Doing

The State Library of Victoria, whose main reading room on Swanston Street draws researchers from across the country, began a structured deduplication audit of its Pictures Collection in late 2024. The project uses perceptual hashing software — a technique that detects visually near-identical images even when file names and metadata differ — alongside manual review by trained cataloguers. The library has not publicly released figures on the scale of duplication found, but the audit is understood to be ongoing as of this month.

Further north, Darebin City Council's local history collection, held at the Darebin Libraries service headquartered in Preston, faces a more acute version of the same issue. The council digitised a large tranche of its photographic holdings in 2019 and again in 2022 under two separate grant programs, and the two digitisation rounds produced overlapping records with inconsistent metadata. Staff have been working through reconciliation manually, a process that archivists in the field describe as exceptionally time-consuming without dedicated software tooling.

Museums Victoria, based at the Melbourne Museum in Carlton, has further developed its own in-house collections management practices through its Collections Online platform, which publicly indexes more than 1.5 million records. The platform flags potential duplicate objects for staff review rather than automatically removing records, a conservative approach that reflects the genuine risk of false positives in automated systems.

How Melbourne Compares to Amsterdam, Toronto and Seoul

The global picture is instructive. Amsterdam's Stadsarchief — the city's municipal archive — completed a two-year deduplication project across its photographic holdings in 2023, ultimately consolidating more than 40,000 duplicate or near-duplicate image records and publishing a methodology guide that has since been downloaded by institutions in at least a dozen countries. The project cost approximately €380,000 and was funded jointly by the municipality and the Dutch Digital Heritage Network.

Toronto Public Library, which holds one of North America's larger local photographic collections through its Baldwin Collection, adopted a semi-automated workflow in 2022 that reduced manual review time by roughly 60 percent, according to figures the library published in its 2023 annual report. Seoul's National Folk Museum completed a comparable exercise in 2021 as part of South Korea's broader Digital New Deal infrastructure program, processing more than 200,000 image records over 18 months.

Melbourne's efforts are real but less coordinated. There is no single city-wide deduplication standard, no shared software platform across institutions, and no published timeline for when the State Library audit will produce public results. The Victorian Government's Public Record Office Victoria, based in North Melbourne, sets recordkeeping standards for government agencies but does not have direct oversight of library or museum collections.

Practitioners in the sector point to funding as the core constraint. The Amsterdam project's €380,000 investment is roughly equivalent to A$640,000 at current exchange rates — a figure that sits awkwardly against the discretionary budgets of most Victorian collecting institutions outside the major state bodies.

The practical stakes are higher than they might appear. Researchers relying on digitised collections for architectural history, First Nations documentation projects or migration heritage work can waste significant time chasing duplicate records that lead back to the same source material. For institutions trying to license images commercially, duplicate records with conflicting rights metadata create legal exposure.

The State Library is expected to share preliminary findings from its Pictures Collection audit at the Australian Library and Information Association conference later this year. That will be the clearest public signal yet of whether Melbourne is closing the gap on cities that have already done this work — or still finding the full scale of the problem.