Istanbul's Duplicate Image Problem: The Key Decisions That Will Shape the City's Digital Heritage Archive

Istanbul's cultural heritage institutions are sitting on a problem that has quietly ballooned for years: tens of thousands of duplicate digital images stored across overlapping municipal, academic and state archives, consuming server capacity, muddying search results and slowing the work of preservationists trying to document neighborhoods at acute earthquake risk. The question now is not whether to act, but which approach wins out when the competing agencies finally sit down together.

The urgency sharpened after the February 2023 Kahramanmaraş earthquakes renewed public and institutional focus on Istanbul's own seismic exposure. Emergency documentation efforts run by groups including the Istanbul Metropolitan Municipality's Directorate of Cultural Heritage and the Boğaziçi University Kandilli Observatory and Earthquake Research Institute generated vast photographic datasets — often of the same structures, shot within days of each other by different teams. Those datasets were uploaded to separate repositories with no shared deduplication protocol, according to heritage professionals who work with the archives. The result is redundancy at scale, with no single authority holding a clean master record.

Where the Bottleneck Sits

The practical crunch point is in the historic peninsula and along the Golden Horn waterfront, where documentation intensity is highest. The Istanbul Archaeological Museums on Osman Hamdi Bey Yokuşu and the municipality's heritage unit at Saraçhane both maintain independent image libraries covering many of the same Ottoman-era structures in Fatih and Eminönü. Field workers and archivists who use both systems regularly encounter the same facade photographed from nearly identical angles, catalogued under different file names and different metadata schemas, with no flag linking them. That fragmentation makes automated AI-based deduplication — the approach most commonly used in European peer institutions — harder to implement cleanly, because the underlying metadata is inconsistent.

SALT Research, the archive and library operated by Garanti BBVA in Beyoğlu, took an early step toward harmonization when it digitized and opened roughly 1.7 million items in its own collection. That project, completed in phases through 2024, gave Istanbul one of its most searchable image databases. But SALT's holdings are largely historical, meaning the newly generated post-earthquake documentation material sits outside its scope and outside its deduplication workflow entirely.

The Decisions That Cannot Wait

Three choices now define what happens next. First, institutions must agree on a common metadata standard — the most contested decision, because it requires each body to accept that its existing tagging conventions will need partial revision. Talks under the Istanbul Deprem Hazırlık Koordinasyon Kurulu, the city's earthquake preparedness coordination board, have touched on digital infrastructure but have not produced a binding framework for image archives as of July 2026.

Second, the municipality must decide whether to fund a centralized cloud-based image repository or continue maintaining distributed systems with a lightweight syncing layer on top. The cost difference is significant. Cloud migration estimates for comparable municipal archive projects in cities like Amsterdam and Barcelona have run between €800,000 and €2.5 million depending on data volume, and Istanbul's aggregate holdings are larger than either. Budget constraints under current lira-denominated municipal financing make the higher-end option politically difficult, particularly given ongoing disputes between the CHP-run municipality and the AKP central government over resource allocation.

Third, and most immediately, the relevant directorates need to designate a lead agency with actual authority to reject duplicate uploads at the point of ingestion — something no single body currently has the mandate to do across all Istanbul's public heritage databases.

Heritage professionals monitoring the process say the next six months will be decisive. The municipality's 2026-2027 digital infrastructure tender cycle opens in September, and the specifications written into that tender will either bake deduplication requirements into any new contract or leave them out again. Advocates pushing for inclusion argue that the Tarihi Yarımada — the historic peninsula bounded by the old Byzantine walls — cannot wait for another documentation cycle to produce another round of duplicates before the system is fixed. The buildings being photographed are not getting more stable. The window for clean, reliable records is open now.