How Geelong's Digital Records Ended Up Full of Ghost Images, and What Comes Next

Geelong's public agencies and cultural institutions are sitting on digital archives bloated with thousands of duplicate images, identical or near-identical files stored multiple times across servers, the result of a decade of rapid digitisation programs, shifting IT platforms, and a regional growth surge that outpaced records management policy. The issue has prompted a coordinated push, now underway in mid-2026, to audit and replace those ghost files with correctly catalogued, single-instance assets.

The timing matters. The Geelong City Deal, the federal-state funding compact that has directed hundreds of millions of dollars into the region since 2017, has also seeded a raft of digitisation and civic data projects. More institutions are publishing open-data sets and digital heritage collections than at any previous point in Geelong's history. When the underlying image libraries are corrupted by duplicates, the publicly facing products, interactive maps, heritage registers, community archives, inherit those errors wholesale.

How the Backlog Built Up

The story runs back at least to the automotive transition period. When Ford Australia closed its Norlane manufacturing plant in October 2016, a wave of council-funded documentation projects launched to record the industrial precinct, the workforce and the built environment before demolition and redevelopption changed them permanently. Photographers, archivists and community volunteers contributed imagery through multiple separate upload portals run by different agencies, with no single deduplication standard in place. Files shot on the same day by different contributors ended up stored under different filenames, often in both JPEG and RAW formats, across the City of Greater Geelong's document management system and separately within Deakin University's community research repositories at its Waterfront Campus on Gheringhap Street.

The pattern repeated itself during the Geelong Ring Road Extensions project documentation, during the Kardinia Park redevelopment, and again when the Surf Coast Shire launched its coastal erosion monitoring program along the stretch between Torquay and Aireys Inlet. Each project generated its own photographic archive, each archive migrated at least once between software platforms, and each migration introduced a new layer of duplicates. IT managers working across these institutions have privately described the compounding effect as a known but low-priority problem for years, one that sat below the threshold of formal remediation funding.

What Remediation Now Looks Like

Deakin University's library and information science faculty has been involved in developing a duplicate-detection methodology trialled at the Geelong Heritage Centre on Little Malop Street, which holds one of the most photographically rich local collections in regional Victoria. The centre's collection spans more than 30,000 items, a figure that has grown substantially since mass-scanning programs accelerated post-2018. The remediation approach being piloted there involves algorithmic perceptual hashing, a technique that matches visually similar images regardless of filename or format, followed by human review of flagged pairs before any file is removed or replaced.

The Surf Coast Shire's planning and environment team has adopted a simpler triage model for its coastal monitoring data, working from a baseline audit completed in March 2026 that found roughly one in eight stored images in its environmental monitoring folder structure was a confirmed duplicate. That figure informed a broader conversation across the Barwon region about minimum standards for image asset management in publicly funded projects.

For residents and community groups contributing to local history projects, particularly through the Geelong Library and Heritage Centre's crowd-sourced digitisation initiatives on Little Malop Street, the practical upshot is straightforward. Submissions made through the centre's online portal from July 2026 onward will pass through an automated pre-ingestion check before being added to the archive. Contributors uploading physical photographs for scanning will be asked to flag if material has been previously submitted, reducing the volunteer labour burden of manual deduplication later.

The City of Greater Geelong has not yet published a formal remediation timeline for its own corporate image holdings, but procurement documents posted to the Victorian Government tenders platform in June 2026 indicate a contract for digital asset management software is being evaluated. Whichever vendor is selected, the underlying catalogue problem will need to be resolved before the data, rather than just the software, is genuinely fit for purpose.