How Paris's Fight Against Duplicate Building Images Reached a Breaking Point

Paris city planners formally acknowledged last month that the municipal photographic inventory of the built environment — a sprawling archive used by architects, heritage bodies, and urban developers alike — contains tens of thousands of duplicate images, many generated during the accelerated documentation drives that preceded and followed the Paris 2024 Summer Olympics. The redundancy problem, years in the making, has begun to actively obstruct planning decisions tied to the Grand Paris Express expansion and the ongoing Seine riverbank regeneration project.

The timing matters. With the Grand Paris Express now in its most capital-intensive phase — extending new metro lines into the inner suburbs of Seine-Saint-Denis and Val-de-Marne — urban planners and environmental assessment teams are pulling heavily from the city's centralised image archive to model building envelopes, street profiles, and heritage-sensitive zones. Duplicate records slow that process, inflate storage costs at the Hôtel de Ville's data infrastructure, and in at least some cases have led to conflicting classifications of the same structure.

How the Archive Got So Cluttered

The roots of the problem stretch back to at least 2019, when the Ville de Paris launched a major street-level documentation campaign in anticipation of Olympic infrastructure works. That drive ran parallel to a separate photographic survey commissioned by the Atelier Parisien d'Urbanisme, known as APUR, which focuses on spatial planning and urban research. Neither database was designed to communicate with the other in real time, and duplication was essentially baked in from the start.

The 2024 Olympics accelerated the problem dramatically. Drone operators and contracted photographers swept through zones including the Plaine Saint-Denis — home to the Athletes' Village — as well as the Bercy Arena district and the reconfigured banks of the Seine between the Pont d'Iéna and the Pont de la Concorde. Multiple teams often photographed the same facades, intersections, and rooflines within days of each other. Post-Games legacy planning then required yet another documentation pass, this time focused on repurposing temporary structures and assessing permanent infrastructure left behind.

By early 2025, the Bibliothèque de l'Hôtel de Ville's digital asset management system — which consolidates records from APUR, the Direction de l'Urbanisme, and several arrondissement-level heritage bodies — had grown to more than 4.2 million image files. Rough internal estimates, circulated during a coordination meeting of Paris urban services in March 2026, suggested that somewhere between 18 and 25 percent of those files were functional duplicates or near-duplicates, meaning images of the same subject taken within a short timeframe and stored under different metadata tags.

The Cost of Doing Nothing

Storage is the immediate expense, but it is not the most serious consequence. In the 10th and 11th arrondissements, planners working on social housing rehabilitation along the Canal Saint-Martin corridor have reported that competing image records attached to the same buildings carry different condition ratings — a product of different teams assessing the same exterior at different stages of renovation. That inconsistency has complicated at least two dossiers submitted to the Commission Nationale du Patrimoine et de l'Architecture.

The rental and housing market tension already squeezing Paris renters gives the issue a practical edge. Accurate building condition data feeds directly into decisions about permitted rental values under the city's encadrement des loyers — the rent-capping regime that the Ville de Paris has enforced since 2019 and which now covers the entire city. If a building's classification is muddied by contradictory photographic records, appeals by landlords contesting rent-cap designations become harder to adjudicate cleanly.

City technicians are now piloting an automated deduplication protocol across a test batch of roughly 200,000 files drawn from the 18th arrondissement — Montmartre and the Goutte d'Or — as a proving ground before any city-wide rollout. The process uses perceptual hashing to flag near-identical images for human review. If the pilot, scheduled to conclude before the end of September 2026, meets its benchmarks, the Direction des Systèmes et Technologies de l'Information plans to extend it across the full archive through 2027, timed to coincide with the next major Grand Paris Express commissioning tranche.