Brussels Archives Are Drowning in Duplicate Images — and the Numbers Tell a Costly Story

Brussels municipal administrations collectively hold an estimated 2.4 million digital image files across departmental servers, and a significant share of them are exact or near-exact duplicates — a problem that is costing the city money and slowing the digital infrastructure that residents increasingly depend on. That finding emerged from a technical review conducted across several commune-level IT departments this spring, with Ixelles and Molenbeek-Saint-Jean among the districts that have begun internal audits in response.

The timing matters. The Brussels-Capital Region is midway through its Digital Brussels 2030 strategy, which commits €47 million over the decade to modernising the administrative technology used by the Region's 19 communes. Storage inefficiency is not a peripheral concern inside that programme — it directly eats into the budget headroom meant for front-line digital services, from online permit applications to open-data platforms accessible via the data.brussels portal.

What the Duplicate Problem Actually Looks Like in Practice

Duplicate image accumulation is a structural, not accidental, problem. When departments digitise planning documents, event photography, heritage records and press materials without a centralised asset management system, the same JPEG or TIFF gets saved multiple times under different file names across multiple folders. IT staff at Saint-Gilles commune identified more than 38,000 duplicate image files during a January 2026 internal clean-up — files that collectively occupied 214 gigabytes of live server space. That is storage capacity that carries a real cost: enterprise-grade cloud storage provisioned through the regional framework contract runs at roughly €0.023 per gigabyte per month, meaning those redundant files alone generated an ongoing annual overhead in the low thousands of euros per commune.

Multiply that across 19 communes, factor in the archives held by institutions such as the Institut du Patrimoine Bruxellois on the Rue du Chêne and the urban-planning directorate at Bruxelles Urbanisme et Patrimoine, and the aggregate figure becomes substantial. Sector specialists working in municipal records management put the proportion of duplicate or redundant files in unmanaged public-sector image libraries at between 15 and 30 percent — a range consistent with audits published by similar-sized European administrations in cities including Amsterdam and Vienna in recent years.

Automated Detection Tools Enter the Picture

The practical response being evaluated by the Brussels regional IT coordination body, Paradigm, centres on automated duplicate-detection software that uses perceptual hashing — a technique that identifies visually identical or near-identical images even when file names and metadata differ. A pilot programme launched at the Anderlecht commune offices on the Square Jacques Franck in March 2026 ran perceptual-hash scanning across 180,000 archived image files. Within 72 hours the tool flagged 22,400 files — just under 12.5 percent — as high-confidence duplicates, with a further 8,000 queued for human review because minor edits or compression differences placed them in a grey zone.

The Anderlecht pilot is scheduled to report final results to the Paradigm steering committee in September 2026, with a decision on region-wide rollout expected before the end of Q4. If adopted across all 19 communes and major regional institutions, the project scope would cover an estimated 3.5 million files by mid-2027, according to the programme's internal timeline documents.

For residents, the downstream effect of a cleaner digital estate is tangible. The open-data image catalogues that journalists, architects, researchers and neighbourhood associations use through the Brussels Environment and irisnet.be platforms load faster and return more accurate search results when duplicate entries are removed. A pilot clean-up of 60,000 heritage photos on the irisnet repository in late 2025 reduced average search-result loading time by 18 percent, according to technical notes published alongside the repository's December 2025 update log.

Communes that have not yet started their own audits would be well advised to begin mapping their image storage before the Paradigm report lands in autumn. The regional programme offers a technical assistance window — applications for that support close on 30 September 2026 — and departments that apply early gain access to the same hashing tools trialled in Anderlecht at no additional cost to their own IT budgets.