Berlin's Public Archives Push to Fix Thousands of Duplicate Images in Digital Collections This Week

Berlin's Landesarchiv and the Zentral- und Landesbibliothek Berlin launched a joint technical review this week targeting an estimated 14,000 duplicate image entries spread across their shared digital catalogue, a problem archivists say has quietly distorted search results and misled researchers for several years. The push, timed to coincide with a broader Senat-backed digitisation initiative, is the first coordinated deduplication effort of its scale across Berlin's public cultural holdings.

The timing matters. Berlin's digital infrastructure for public records has expanded rapidly since 2022, when the Senat allocated additional funding under the Kulturdigitalisierung programme to move physical holdings online. More collections, contributed by district archives in Mitte, Kreuzberg and Pankow, were folded into the same shared database without a unified image-fingerprinting standard. The result: identical photographs of Weimar-era streetscapes or postwar reconstruction sites were uploaded multiple times under different metadata tags, making systematic research unreliable.

What the Deduplication Actually Involves

The technical work, running from June 30 through July 11, uses perceptual hashing — software that compares visual structure rather than file names — to flag near-identical images. Staff at the Landesarchiv on Eichborndamm in Reinickendorf are manually reviewing flagged pairs before any deletion, a safeguard inserted after a 2024 pilot project in Hamburg's Staatsarchiv mistakenly removed a small number of genuinely distinct photographs during automated cleanup. Berlin archivists drew directly on that Hamburg experience when drafting their own protocol this spring.

The Zentral- und Landesbibliothek's digital team, based at the Amerika-Gedenkbibliothek on Blücherplatz in Kreuzberg, is handling the metadata reconciliation side — ensuring that once duplicates are removed, the surviving image record carries complete, correctly attributed information. Around 3,200 images flagged so far this week are linked to the Topographie des Terrors documentation centre's contributed collection, a sensitive subset requiring extra verification steps before any changes are committed.

Researchers who use the shared portal, known as the Berlin Digital Collections Gateway, noticed degraded search quality throughout 2025. A query for images of Potsdamer Platz before 1945, for instance, could return the same photograph six or seven times in a single results page, pushing genuinely different materials off the first screen. The Senat's cultural administration acknowledged the issue formally in a March 2026 written response to a parliamentary question from the SPD faction, though no specific remediation deadline was set at that point.

Why This Affects More Than Historians

The duplicate problem has practical consequences beyond academic research. Berlin's growing network of neighbourhood history projects — including the Stadtteillabor in Wedding and community memory groups in Neukölln — relies on the same digital gateway to source images for exhibitions and educational materials. Mislabelled or endlessly repeated images have on at least two occasions led community groups to present the same photograph twice in public displays, according to a notice circulated by the Bezirksamt Neukölln's cultural office in April 2026.

Startups in Berlin's tech sector have also begun licensing archival imagery commercially through a 2025 partnership scheme with the Landesarchiv. Licensing fees start at €45 per image for non-commercial use and rise significantly for print or broadcast rights. Duplicate entries complicate rights tracking, since the same image might carry different rights metadata depending on which upload it originated from — a legal exposure that the current cleanup is partly designed to close.

The full deduplication review is expected to conclude by July 11, with a public report on findings submitted to the Senat's cultural administration by the end of the month. Once the database is clean, archivists plan to implement an automated hash-check at the point of upload, so newly contributed collections from district archives cannot re-introduce duplicates. Researchers and community groups wanting to flag specific problem records before the July 11 deadline can submit corrections directly through the Berlin Digital Collections Gateway contact form, which archivists say they are monitoring daily through the end of the review period.