How Delhi's Government Archives Ended Up Full of Ghost Images: The Duplicate Photo Crisis Explained

The problem did not appear overnight. Delhi's municipal and state government digital archives contain tens of thousands of duplicate and misidentified photographs — images filed under wrong project names, wrong locations, wrong years — the accumulated result of at least four separate digitisation efforts run between 2014 and 2024 with no single agency coordinating the work. Officials at the Delhi Secretariat on ITO are now confronting a records backlog that stretches across departments and threatens the credibility of project documentation for everything from the Yamuna riverfront redevelopment to Phase 4 of the Delhi Metro.

The timing matters because transparency in public infrastructure spending is under acute political scrutiny. With AAP chief minister Arvind Kejriwal defending flagship welfare programmes against pressure from the BJP-led central government, any suggestion that photographic evidence supporting project completion claims is unreliable or duplicated feeds directly into an already combustible accountability debate. The image-duplication issue is not merely a bureaucratic nuisance — it is a paper trail problem with real political stakes.

How Four Digitisation Drives Created One Enormous Mess

The roots of the crisis go back to 2014, when the Delhi government's Department of Information Technology awarded its first bulk scanning contract to a Nehru Place-based vendor to digitise paper records from the Public Works Department. That drive produced roughly 1.2 lakh scanned files, many of them photographs of road works and drainage projects in areas including Rohini, Dwarka, and Shahdara. Because no standardised naming protocol was enforced, files were labelled by the scanning operator rather than by any departmental taxonomy. Duplicates entered the system from the first week.

Three subsequent digitisation projects — in 2017, 2020, and 2023 — imported new batches of photographs without first auditing what already existed. The 2020 drive, conducted during the pandemic period partly by a consortium linked to the Delhi Urban Shelter Improvement Board, added an estimated 40,000 images from resettlement colony documentation in areas like Seelampur and Sangam Vihar. Cross-referencing across drives never happened. By the end of 2024, the Delhi State Data Centre in Dwarka Sector 11 was hosting multiple versions of the same site photographs, some misfiled under entirely different project codes.

A Right to Information request filed in March 2025 by a civil society researcher at the Centre for Policy Research in Jor Bagh first brought the scale of the problem into public view. The response from the Department of Administrative Reforms acknowledged that no unified image registry existed across departments and that deduplication had not been carried out on the central server. The department's written reply confirmed the issue without providing a remediation timeline.

What the Cleanup Requires — and Who Is Responsible

Fixing the archive is not simply a matter of running deduplication software. Photographs linked to contract disbursement records — particularly those tied to the Pradhan Mantri Awas Yojana-Urban housing files administered jointly through the Delhi Development Authority on Vikas Minar — carry legal weight. Deleting a duplicate that turns out to be the only surviving copy of a site-inspection photograph could compromise an ongoing audit or court case. Archivists and technology consultants working on similar projects in cities like Mumbai and Bengaluru have recommended a minimum six-month review cycle before any bulk deletion is authorised.

The Delhi government's IT department put out an expression of interest in April 2026 for a vendor to conduct a full deduplication audit across five departments, with a project value ceiling of Rs 3.8 crore. The deadline for submissions was June 30, 2026 — four days ago. Whether a contract has been awarded is not yet confirmed through official channels.

For departments currently preparing documentation for Phase 4 Metro corridor inspections along the Janakpuri West–R.K. Ashram Marg line, the practical advice from records management specialists is straightforward: photograph every site with GPS-embedded metadata switched on, use a department-issued file naming convention from day one, and submit images to a single nodal officer before they are uploaded to any shared server. The workaround is unglamorous. It is also the only thing standing between the next contractor and another decade of ghost images.