Canberra's War on Duplicate Images in Government Records: How the Capital Stacks Up Against Cities Worldwide

Canberra's public sector holds one of the densest concentrations of digital imagery archives in the Southern Hemisphere. From aerial surveillance photographs at Geoscience Australia's Symonston campus to heritage documentation stored by the ACT Historic Places office on London Circuit, the volume of duplicate image files clogging government servers has become a measurable administrative problem — and a costly one. An internal review process under the Australian Government Records Interoperability Framework, active since late 2024, has pushed the issue onto the desks of chief information officers across multiple Commonwealth agencies.

The timing matters. Federal departments are under pressure from the Australian Public Service Commission to reduce data storage overhead as part of broader digital efficiency targets tied to the 2025-26 Budget's productivity agenda. Duplicate image files — often created when staff in different divisions photograph the same infrastructure, property or event without coordinating — consume server capacity, complicate Freedom of Information searches, and create version-control headaches that can delay policy work. For a city whose entire economic base is government administration, that is not a trivial inefficiency.

What Canberra Is Actually Doing

The National Archives of Australia, headquartered on Queen Victoria Terrace in Parkes, began piloting perceptual hashing software across a subset of its digitised photographic collections in the first quarter of 2026. Perceptual hashing allows automated systems to flag near-identical images even when file names, formats or metadata differ — a known weakness in earlier deduplication approaches that relied solely on file-size matching. The Archives has not publicly disclosed the scale of duplication found, but the pilot is understood to cover collections digitised between 2018 and 2024.

Separately, the ACT Government's Digital Strategy Division, operating out of offices in Canberra City, has been working through a deduplication audit of imagery held by Transport Canberra and City Services — including thousands of construction-phase photographs taken during the light rail Stage 1 build along Northbourne Avenue. Stage 2 documentation, covering the proposed route through the inner south toward Woden, is being structured from the outset with single-source image repositories to avoid replicating the earlier problem.

The Australian National University Library, on the Acton campus, runs its own digital asset management system for research imagery and has used open-source deduplication tools since 2022. University of Canberra's library team in Bruce adopted a comparable approach for its institutional repository around the same period. Neither institution has published comparative cost data, but open-source tools used in the sector, such as dupeGuru and PhotoPrism, carry no licensing fee — meaning the primary cost is staff time.

How Other Cities Are Handling the Same Problem

Canberra's approach sits somewhere in the middle of a global spectrum. Singapore's Government Technology Agency — GovTech — embedded automated image deduplication into its whole-of-government cloud migration program in 2023, applying it at ingestion rather than retrospectively. That upstream approach is generally regarded among archivists as more efficient, though it requires consistent metadata standards that many legacy systems lack.

Wellington, New Zealand — a useful comparison given its similar size, public-service-dominated economy, and parliamentary function — has handled the problem less systematically. The New Zealand Parliamentary Service acknowledged in its 2024 annual report that it was still managing duplicate digital records across multiple file-storage platforms, with a consolidation project scheduled to run through 2027.

Amsterdam's city archive, the Stadsarchief, completed a large-scale deduplication of its 20th-century photograph collection in 2022, removing more than 40,000 duplicate or near-duplicate image files from a public-facing database. That project used machine-learning classification and cost approximately €180,000, according to the Stadsarchief's published project documentation.

For Canberra residents and public servants, the practical upshot is straightforward. Agencies that complete deduplication work will respond faster to FOI requests involving photographs or visual records, since archivists will spend less time sorting through redundant files. Property developers seeking heritage clearances in suburbs like Braddon or Reid, where ACT Historic Places holds extensive photographic records, may notice shorter processing times as the audit progresses. The ACT government has not published a completion date for its current audit, but the Digital Strategy Division's work plan runs to the end of the 2026-27 financial year.