How Delhi's Government Archives Ended Up Flooded With Duplicate Images — and What It Cost to Fix It

Delhi's Department of Information and Publicity quietly crossed a milestone this spring: after nearly 18 months of remediation work, technicians completed the first full audit of the capital's centralised photographic archive, a repository that had swelled to more than 2.3 million image files — with independent assessors estimating that roughly 40 percent of those files were functional duplicates, stored multiple times across incompatible servers maintained by at least seven separate agencies.

The audit matters because it marks a turning point in how the Aam Aadmi Party administration manages its visual public record. Duplicate images are not simply a storage nuisance. They distort official timelines, create legal ambiguity over which version of a document photograph is authoritative, and drive up cloud-hosting costs for departments that are already operating on stretched budgets. For a government that has made digital transparency — through platforms like the Delhi e-District portal — a core part of its administrative identity, a bloated and contradictory image archive undercuts that message directly.

How the Backlog Built Up

The problem has roots that stretch back to 2015, when the AAP government first took office and began digitising large volumes of paper records held at the Delhi Secretariat on ITO Road. The rush to digitise — accelerated further during the Covid-19 lockdowns of 2020 and 2021, when field inspections were halted and departments defaulted to scanning everything they had — meant that quality controls were sacrificed for speed. The Public Works Department, the Delhi Urban Shelter Improvement Board, and the Delhi Jal Board each built separate scanning pipelines, none of which talked to the others. A single inspection photograph of, say, a Yamuna flood-plain encroachment site near Usmanpur in northeast Delhi might be scanned by a field officer, uploaded by a department head, and then re-uploaded by a communications staffer preparing a press release — three copies, no metadata linking them.

The Delhi Metro Rail Corporation added another layer of complexity. As Phase 4 construction documentation began flowing in from sites along the Janakpuri West–R.K. Ashram corridor and the Aerocity–Tughlakabad line, thousands of progress photographs entered the broader municipal system without a standardised naming convention. By late 2024, an internal review commissioned by the Directorate of Information Technology found that server storage allocated to image files had doubled in three years, with costs running significantly above budget projections for that financial year.

What the Remediation Actually Involves

The current cleanup programme, managed through a tender awarded to a Delhi-based IT firm in January 2025, uses perceptual hashing — a technique that generates a numerical fingerprint for each image and flags near-identical copies — to identify duplicates without requiring a human to open every file. The process is running across data centres in Dwarka Sector 10, where much of the Delhi government's cloud infrastructure is hosted.

The practical challenge is deciding which copy to keep. Archivists at the Delhi State Archives on Shamnath Marg have been brought in to establish a hierarchy: the file with the earliest verified timestamp and the most complete metadata is designated the master record; the rest are flagged for deletion after a 90-day review window. Departments have until September 30, 2026, to contest any proposed deletion.

This is not a trivial exercise. Storage costs aside, the legal standing of government photographs in court proceedings — particularly in land-acquisition disputes near areas like Rohini and Narela, where municipal boundary documents are frequently contested — depends on there being a single, traceable master file. Courts have questioned the provenance of digital images in at least two Delhi High Court cases in the past three years, according to cause-list records reviewed by this reporter.

For ordinary Delhiites seeking certified copies of municipal photographs through the Right to Information Act, the practical payoff should arrive by early 2027: a single searchable portal, rather than the current situation where applicants must know which department holds which version of a record. The Directorate of Information Technology has described this unified portal as a target deliverable in its 2026-27 annual plan. Whether the September deadline holds depends on how many departmental objections pile up between now and autumn.