Brussels City Archive Flags Hundreds of Duplicate Images in Digital Heritage Push This Week

Brussels city archivists confirmed this week that an ongoing audit of the municipality's digital photograph collections has identified more than 400 duplicate image files spread across the Erfgoedbibliotheek/Bibliothèque du Patrimoine, the urban planning department's internal repository, and the publicly accessible Brussels Urban Development portal. The discovery, surfacing during a broader digitisation review that began in March 2026, has prompted an emergency working session scheduled for 9 July at the Stadhuis on the Grand-Place.

The timing matters. Brussels is currently two years into its Digitaal Erfgoed 2025-2030 programme, a cross-departmental initiative designed to consolidate the city's fragmented heritage collections under a single searchable platform. Duplicate images are not merely a storage problem — they distort search results, inflate reported collection sizes used to justify funding requests, and in some cases attach conflicting metadata to the same photograph, meaning two records of the same 1970s demolition site in Molenbeek can carry different dates, different credited photographers, and different usage licences. That has real legal and financial consequences.

The audit was triggered in late May when staff at the BOSA digitisation unit — the federal agency that provides technical support to Belgian public administrations — flagged an unusually high rate of near-identical files during a routine quality check on a batch of images transferred from the Archives de la Ville de Bruxelles on the Rue des Tanneurs in the Marolles district. Archivists then ran the same detection software across two additional datasets: the photographic holdings digitised under a 2022 agreement with the Université libre de Bruxelles, and a collection of urban survey images produced by the Brussels-Capital Region's own Perspective.Brussels planning agency.

What the Audit Actually Found

Of the 400-plus flagged files, roughly 60 percent are described by administrators as exact duplicates — byte-for-byte identical files that entered separate databases through parallel digitisation workflows with no cross-referencing. The remaining 40 percent are near-duplicates: slightly different crops, resolution conversions, or files that passed through different scanning equipment at different times. The combined collections under review contain approximately 38,000 digitised images, meaning the duplication rate runs to just over one percent — low by some international benchmarks but significant given the metadata conflicts attached to many of the flagged pairs.

The financial dimension adds pressure. The Digitaal Erfgoed programme received a confirmed allocation of €2.1 million for the 2026 fiscal year from the Brussels-Capital Region's budget, part of which funds external contractors to process and catalogue incoming batches. Administrators privately acknowledge — though no official statement has been issued — that paying to process and store duplicate files twice wastes a portion of that envelope. Exact figures on wasted expenditure have not been published and will form part of the 9 July working session agenda.

Perspective.Brussels confirmed this week that it is cooperating with the audit but declined to provide a specific count of flagged files from its own holdings before the July meeting. The Archives de la Ville de Bruxelles did not respond to a request for comment by publication time.

What Comes Next for the Collections

City administrators are weighing three options ahead of the 9 July session. The first is a manual review of all 400-plus flagged pairs by qualified archivists — slow but legally safest given the competing licence claims. The second is automated deduplication using perceptual hashing software already deployed by the Royal Institute for Cultural Heritage (KIK-IRPA) on its own collections; KIK-IRPA has signalled it could share the toolset under an existing inter-institutional agreement. The third option is a hybrid: automated flagging with human sign-off before any file is deleted or merged.

For residents and researchers who regularly use the publicly accessible image databases — including local history groups in Schaerbeek and Sint-Gillis who rely on the portals for neighbourhood documentation projects — the practical advice is straightforward: cross-reference any image found on the Brussels Urban Development portal against the Archives de la Ville de Bruxelles catalogue before citing or reproducing it. Until the deduplication process is complete, metadata attached to individual files cannot be fully trusted. A resolution is unlikely before September at the earliest.