Delhi's War on Duplicate Images: How the Capital Stacks Up Against London, Singapore and São Paulo

Delhi's three municipal corporations merged into a single entity — the Municipal Corporation of Delhi — in May 2022, and the administrative consolidation left behind one very unglamorous problem: tens of thousands of scanned property documents containing duplicate, mis-tagged or outright blank image files. Three years on, the MCD's GIS and Records Digitisation Cell, operating out of its headquarters on S.P. Mukherjee Civic Centre in Connaught Place, is still working through a backlog that staff estimate runs to more than 400,000 individual document images across the unified property tax database.

The issue matters now because Delhi's property registration system feeds directly into the Delhi Development Authority's master plan enforcement, the Pradhan Mantri Awas Yojana housing eligibility checks, and the Chief Minister's ongoing Yamuna floodplain encroachment audit. A duplicate image — two different properties assigned the same scanned file, or one property's record illustrated with another building's photograph — is not a clerical curiosity. It can block a loan, delay a mutation order, or create a legal loophole that lawyers in Tis Hazari District Court have been known to exploit for years.

What Delhi Is Actually Doing

The MCD's current approach relies on a two-stage process. First, an automated hash-matching tool flags files where pixel data is identical. Second, human reviewers at a verification centre in Rohini's Sector 7 offices manually assess flagged pairs to determine whether the duplication was a scanning error or a deliberate data entry shortcut taken during the rushed digitisation drives of 2018 and 2019. The Rohini centre handles roughly 1,200 image reviews per working day, according to the workflow schedule published on the MCD's public procurement portal in March 2026.

The Delhi State Spatial Data Infrastructure programme, run jointly by the Delhi government and the Survey of India's Northern Zonal office in Janpath, is separately building a master image registry that would assign each geo-tagged property photograph a unique identifier tied to its khasra plot number. Pilot testing for that registry began in Shahdara district in February 2026, covering approximately 18,000 properties in the first phase.

Compare that with Singapore's approach. The Singapore Land Authority completed a full de-duplication of its Integrated Land Information Service database in 2023 after a three-year project that used perceptual hashing combined with machine-learning classifiers trained on building typologies. The SLA publicly reported a final duplicate rate of under 0.3 percent across 1.4 million records. Delhi, by contrast, has not published an equivalent completion metric, and the MCD's digitisation progress dashboard — accessible via its public grievance portal — last updated its image-verification figures in November 2025.

How Other Cities Approach the Same Problem

London's Valuation Office Agency, which maintains property records for the 33 boroughs, moved to a centralised image-management system called CAIS — the Central Asset Image Store — in 2021. The VOA mandates that every property image carry an embedded timestamp and surveyor ID before it enters the live database, a pre-ingestion quality gate that Delhi's systems currently lack. São Paulo's municipal cadastre, the Cadastro Técnico Municipal run by the Secretaria Municipal de Urbanismo e Licenciamento, tackled a similar backlog between 2020 and 2024 by outsourcing de-duplication reviews to a consortium of three technology firms, cutting its unresolved duplicate rate from an estimated 8 percent to below 1 percent in roughly four years.

Delhi's challenge is compounded by scale. The city's land records span colonial-era patwari registers, mid-century municipal ledgers, and post-liberalisation digital files, all converted to image formats at different resolutions and under different administrative mandates. Old Delhi neighbourhoods like Ballimaran and Chitli Qabar have particularly dense and contested property histories, where a single narrow plot may have accumulated a dozen document scans across different agencies over 70 years.

The practical consequence for residents is straightforward: anyone applying for a property mutation at a sub-registrar office in areas covered by the Shahdara pilot should ask staff whether their khasra number has been enrolled in the new image registry. If it has, the duplicate-check runs automatically before paperwork is accepted. If it has not — and most of Delhi falls outside the pilot zone for now — the older, slower manual process still applies, and applicants should retain physical copies of every document submitted, since image mismatches have previously caused files to be returned without explanation. The MCD has said the Shahdara pilot is due for an evaluation report in September 2026, which will determine whether the registry rolls out citywide.