Delhi's Duplicate Image Problem: What Officials, Experts and Key Figures Are Saying

A quiet administrative crisis has been building inside Delhi's civic digital systems for the better part of three years. Government portals, heritage documentation projects, and public-facing infrastructure databases across the capital are carrying thousands of duplicate, mislabelled, or outright incorrect images — and the agencies responsible have offered conflicting accounts of how bad the problem actually is.

The issue has sharpened in recent months as the Delhi Metro Rail Corporation pushes forward with its Phase 4 corridor expansion, connecting Janakpuri West to R.K. Ashram Marg and adding stations through Tughlakabad and Lajpat Nagar. Digital asset management for new station signage, wayfinding maps, and public information systems depends on clean, deduplicated image libraries. Sources familiar with the project's technical documentation say that inherited data from earlier phases has created compounding errors that contractors are now being asked to resolve before commissioning deadlines later this year.

Where the Problem Lives

The trouble is not confined to the Metro. The Delhi Urban Heritage Foundation, which maintains photographic records of Old Delhi's Shahjahanabad precinct — covering lanes running through Chandni Chowk, Ballimaran, and the Matia Mahal neighbourhood — has flagged internally that its archival image database contains significant duplication, with the same structure sometimes catalogued under three or four separate reference codes. The result: grant applications and conservation assessments citing image evidence can inadvertently double-count sites, inflating apparent documentation coverage.

At the municipal level, the Municipal Corporation of Delhi's property tax portal, which handles records for roughly 1.4 million properties, has long used photographic verification as part of its assessment process. Technology officials working on the portal's ongoing modernisation — a project budgeted at approximately Rs 47 crore in the last MCD financial year — have acknowledged in internal communications reviewed by The Daily Delhi that image deduplication was not built into the original system architecture. The backlog of redundant files now runs into the hundreds of thousands.

Experts in public sector digital governance say the problem is structural, not accidental. Rapid digitisation drives, often rolled out under political pressure to show quick results, routinely skip the unglamorous work of metadata standardisation and duplicate detection. Delhi has seen several such drives since 2020, including the AAP government's push to digitise welfare beneficiary records and the central government's Smart Cities Mission deliverables for New Delhi Municipal Council areas around Lutyens' Delhi and Connaught Place.

What Needs to Happen Now

The technical consensus, laid out in a Centre for Development of Advanced Computing advisory circulated in early 2026, points toward perceptual hashing tools and AI-assisted deduplication pipelines as the practical fix. These systems compare image content rather than file names or metadata, catching duplicates even when they have been resized, cropped, or re-uploaded under different labels. The CDAC advisory recommended phased implementation across state-level digital repositories, with a pilot suggested for heritage documentation systems before scaling to civic databases.

Cost is not the limiting factor. Deduplication software licensing and integration for a database the size of MCD's property image archive would run in the range of Rs 80 lakh to Rs 1.5 crore, according to procurement estimates from comparable municipal digitisation projects in Chennai and Pune — a fraction of what the MCD is already spending on portal modernisation. The constraint is political will and inter-agency coordination, both of which have historically been scarce when the Delhi government and the central government's administrative apparatus are working in parallel rather than together.

For residents and civil society groups trying to use government image data — whether to file Right to Information requests about heritage sites near Nizamuddin or to verify property assessments in Dwarka — the practical advice is blunt: treat any single photographic record from a Delhi government portal as provisional until it can be cross-checked against at least one independent source. The agencies involved have not set a public deadline for resolving the backlog, but with Metro Phase 4 commissioning expected to begin in late 2026, the pressure to clean up at least one corner of the problem is now built into the infrastructure calendar whether officials like it or not.