How Istanbul's Historic Archive Became a Battlefield Over Duplicate Images

Istanbul's municipal digital archive holds more than 1.4 million scanned photographs, maps, and technical drawings — a collection assembled over three decades by the Istanbul Metropolitan Municipality's Department of Cultural Heritage. A growing percentage of that collection is now flagged as containing duplicate images: identical or near-identical files uploaded multiple times across separate databases, in some cases linked to contested development permits along the Bosphorus waterfront and in the historic Fatih district.

The issue matters today because the municipality is mid-way through a 2024–2027 digitisation drive that is supposed to make heritage records publicly searchable for the first time. Duplicate and mismatched image files, archivists have warned in internal working documents, undermine that goal by creating conflicting reference points — a problem with direct consequences when those records are cited in planning decisions for earthquake-vulnerable neighbourhoods.

How the Duplicates Accumulated

The root of the problem stretches back to the early 2000s, when the municipality ran at least three parallel scanning programmes — one through the Istanbul Research Institute in Beyoğlu, one through the municipality's own Atatürk Library in Taksim, and a third contracted to a private firm — without a unified file-naming convention or a central deduplication protocol. Each institution kept its own version of scanned negatives, and when those collections were eventually merged into the shared İBB (Istanbul Büyükşehir Belediyesi) digital platform after 2019, the duplicates migrated with them.

The problem compounded after the February 2023 Kahramanmaraş earthquakes, which prompted an emergency audit of seismic risk assessments across Istanbul. Planners needed to pull historical photographic evidence of building construction dates in districts like Avcılar and Bağcılar, where older masonry stock is concentrated. Researchers found multiple versions of the same 1960s survey photographs tagged with different dates, different building addresses, and in several cases different GPS coordinates — inconsistencies that slowed the audit by weeks, according to a summary of the review circulated within the municipality's urban planning division.

The Istanbul Research Institute, which operates under the Yapı Kredi cultural umbrella and is based in Beyoğlu, had separately digitised roughly 80,000 images of the city from its own physical collection. When integration talks with the İBB archive began in 2022, the absence of a shared metadata standard meant that an estimated 12,000 to 15,000 of those images overlapped with files already held by the Atatürk Library — different scans of the same original prints, none of them formally designated the authoritative version.

What a Fix Would Actually Require

Archivists at the Atatürk Library have pushed for adoption of a perceptual hashing system — software that compares images at the pixel level to flag near-identical files regardless of filename — since at least 2021. A pilot involving 200,000 images from the Fatih district's Ottoman-era building stock ran through the first half of 2025. The results, presented to the municipal heritage committee last autumn, found a duplication rate of roughly 8 percent across that sample, with a smaller subset — around 1,100 images — containing outright conflicting metadata.

The Bosphorus development controversy has sharpened the stakes. Heritage activists monitoring permit applications for waterfront properties between Bebek and Rumelihisarı have pointed to cases where developers referenced archival photographs to argue a building predates modern zoning restrictions, only for opponents to produce what appears to be an identical scan tagged with a different year. Without a canonical image record, neither claim can be definitively resolved through the archive alone.

The municipality's current roadmap calls for a deduplication pass across the full 1.4-million-item collection to be completed by the end of 2026, with a public-facing search portal to follow in early 2027. The practical deadline is harder than it looks: the municipality must also contend with integrating records from the 39 district municipalities, several of which ran their own scanning projects after 2009 and have yet to submit files to the central platform. Archivists say that once deduplication software is fully deployed, the first public test will be a searchable photographic record of Sultanahmet — Istanbul's most documented square kilometre — made available to researchers and, eventually, residents filing heritage objections against demolition permits.