Brussels Digitisation Drive Confronts Duplicate Image Problem in City Archives

Brussels city archivists formally launched the active deduplication phase of their digitisation programme this week, targeting an estimated backlog of redundant scanned images that has accumulated across multiple municipal databases since the first wave of document scanning began in 2019. The Urban Archives Service, based on Rue des Tanneurs in the Marolles district, confirmed the work is underway across at least three linked databases covering building permits, heritage listings and urban planning submissions.

The timing matters. Brussels is mid-way through a broader push to make planning and heritage records searchable online through the city portal Irisbox, the regional one-stop digital gateway operated by the Brussels-Capital Region. Duplicate image files — created when the same physical document was scanned more than once, submitted through multiple departments, or imported from legacy systems during software migrations — have slowed search performance and created conflicts in version control. A planning officer looking up a 1970s building permit for a property in Schaerbeek, for instance, might encounter three versions of the same page flagged as distinct records.

What Happened This Week

The deduplication work entered its most visible phase on Tuesday, July 1, when the Urban Archives Service took portions of the building permit database offline for rolling maintenance windows of four hours each. Users accessing Irisbox reported intermittent gaps in document retrieval for properties in the Saint-Gilles and Etterbeek communes, where scanning volumes were highest during the 2021-2023 period. The maintenance schedule is expected to run through July 18.

The project is being handled using open-source perceptual hashing tools adapted by the city's IT directorate, Digit.Brussels, which manages digital infrastructure across the 19 communes of the Brussels-Capital Region. Perceptual hashing compares image fingerprints rather than file names or metadata, meaning a document scanned at 300 dpi and the same page rescanned at 150 dpi are still identified as duplicates. According to internal project documentation circulated at a May meeting of the Regional Informatics Commission, the pilot run on a subset of Ixelles commune records found a duplication rate of roughly 18 percent — meaning nearly one in five image files was a redundant copy of an existing record.

That 18 percent figure, if it holds across the full archive, translates to a significant storage and retrieval problem. The city's digitisation contracts, awarded in stages from 2019 onward, priced scanning at a per-image rate. Redundant images inflated invoice totals and now occupy server space maintained at the regional data centre on Avenue des Arts in the European Quarter. Digit.Brussels has not published a total cost estimate for the deduplication effort.

What It Means for Residents and Developers

For residents and property developers navigating the permit system, the practical effect this week is straightforward: some historical document searches will return incomplete results until each commune's records are cleaned and re-indexed. The Urban Archives Service has advised users requiring urgent access to heritage or planning documents to submit requests by email directly to the relevant commune office rather than relying solely on the Irisbox portal during the maintenance period.

Architects submitting new planning applications through the Brussels Environment agency, known as Bruxelles Environnement, are unaffected — that submission pipeline runs on a separate infrastructure. The disruption is limited to archival lookups of historical records.

Once the deduplication is complete, city officials intend to republish the cleaned dataset under an open data licence via the regional open data platform, data.brussels.be, making it available for research and mapping projects. Urban historians and heritage organisations including ARAU, the Atelier de Recherche et d'Action Urbaines based in the Lower Town, have previously used similar datasets for neighbourhood surveys. The cleaned archive is expected to be available in full by the end of the third quarter of 2026, though the Etterbeek and Schaerbeek batches — the largest by volume — are not scheduled for completion until September. Residents with time-sensitive permit queries are advised to call their commune directly in the interim.