How to Catalogue a Photograph Collection: A Step-by-Step Guide

Somewhere in a local history society, there are four cardboard boxes. Inside them: roughly 800 photographs, mostly black and white, many unlabelled, some with pencil annotations on the back that say things like “Harvest Festival — late 1930s?” Nobody can say with certainty who donated them, which parish they depict, or whether anyone has seen them since 1987. The photographs are deteriorating. They are entirely unsearchable. And they are, in all likelihood, irreplaceable.

This is not an unusual situation. Across the UK, museum stores, county record offices, family archives, and community heritage projects hold millions of photographs — prints, negatives, daguerreotypes, glass plates — that have never been properly catalogued. When someone does sit down to figure out how to catalogue a photograph collection, they often discover that the task is simultaneously more important and more complicated than it first appeared.

This guide will walk you through what a proper catalogue record looks like, which standards apply, and how modern AI-assisted tools can reduce a months-long project to something far more manageable — without compromising on quality.

---

Why Cataloguing Photographs Matters

The most obvious reason is discoverability. A photograph that has no catalog record is, for practical purposes, invisible. It cannot be found in a search, cannot be surfaced in response to an enquiry, and cannot be included in any finding aid that a researcher might consult. The archive holds it, but it might as well not exist.

There are also preservation arguments. The cataloguing process forces a physical or digital inspection of each item. Condition issues — mould, fading, tears, vinegar syndrome in acetate negatives — are identified and documented. Items that need conservation attention get flagged rather than sitting unexamined until the damage is irreversible.

For institutions seeking funding, a well-maintained catalogue is increasingly a prerequisite. Heritage Lottery Fund and National Lottery Heritage Fund grants routinely ask about collection management policies and cataloguing standards. Demonstrating that your collection is properly described and accessible strengthens any bid considerably.

And then there is the straightforward matter of institutional memory. Staff and volunteers change. The volunteer who knew that “Box 14 — river views” actually contained photographs of flood damage from a specific 1953 event will eventually leave. If that knowledge is not captured in a catalog record, it leaves with them.

---

What Information to Record: Building a Complete Catalog Record

This is where most guides get vague. They say “record the metadata” without specifying which fields matter and why. Drawing on general cataloguing guidelines used across the archival profession — including the approach taken by the Library of Congress Prints and Photographs Division — here is a practical breakdown of what a complete catalog record for a photograph should contain.

Title. A brief descriptive title you assign to the item. Describe what the image shows as specifically as the evidence allows. “View of Market Street looking north, showing Woolworths frontage” is useful. “Untitled photograph” is not.

Reference number. Every item needs a unique identifier — a digital ID that ties the physical object or digital image file to its catalog record and prevents confusion when two photographs look similar or when an item is temporarily separated from its housing.

Date. Record the date as precisely as the evidence allows, using a consistent format. Where you cannot give an exact date, use a range (e.g. “1920–1935”) and document what evidence informs your estimate. Undated prints and photographs should not simply be filed as “undated” if a reasonable estimate is possible from photographic format, costume, vehicle type, or building signage.

Location and place names. Record place names at the most specific level you can verify, using a controlled vocabulary such as the Getty Thesaurus of Geographic Names. Place names are among the most valuable fields for researchers working on local history — they drive discoverability and support geographic browsing across the catalogue.

Subjects and people. Who or what is depicted? For people, record names if known — surname-forename order, with birth and death dates where they can be confirmed. This is particularly valuable for portrait photographs, where the identity of the sitter is often the primary reason a researcher is searching. For objects and places, use subject headings from a controlled list. Catalogers assign index terms from sources such as the Library of Congress Subject Headings to ensure consistency across the catalogue.

Genre and physical characteristics. Record the format — albumen print, gelatin silver print, cyanotype, colour transparency, glass negative, daguerreotype, and so on — along with dimensions and any mount or housing. Genre and physical characteristics matter both for preservation decisions and for researchers studying the history of photography. For digital images, record the file format, resolution, and colour profile.

Condition. Note any damage, fading, tears, or deterioration. A simple controlled vocabulary (Good / Fair / Poor / Critical) is sufficient for most purposes, with a free-text field for specifics.

Rights and access. Who holds copyright? Are there any restrictions on reproduction or access? Copyright in photographs is a genuinely complex area in the UK. Document what you know and flag uncertainties for legal review. Leaving this field blank is not a neutral choice — it creates problems at the point of publication.

Provenance and acquisition. How did the item enter the collection? When? Who transferred it? This contextual information is often what makes the difference between a photograph being interpretable and being a puzzle.

Inscriptions and annotations. Record any text visible on the photograph or its mount — captions, stamps, pencil notes on the reverse, studio marks, or negative numbers. Transcribe inscriptions exactly as written, preserving spelling and punctuation. Note the location of the inscription (e.g., “pencil annotation on verso”) and distinguish between what is written on the object and what you are adding as a cataloguer. An inscription that reads “Aunt Edith, Whitby 1923?” is valuable evidence — but it is the donor’s or creator’s note, not a verified fact.

Linking digital surrogates to physical items. If the photograph exists as both a physical object and a digital scan, the catalogue record must make the relationship explicit. Record the physical location of the original (box, folder, shelf mark) alongside the digital file reference. If multiple scans exist — a preservation master and an access copy, for instance — note which is which. A catalogue record that describes only the digital image, with no reference to the physical original, creates a gap that becomes harder to close over time.

Handling Uncertainty in Photograph Cataloguing

Uncertainty is normal in photograph cataloguing — it is not a failure. Many photographs arrive without clear provenance, with ambiguous dates, unidentified subjects, or locations that can only be estimated. The key is to record uncertainty honestly and consistently, rather than leaving fields blank or guessing without flagging the guess.

Distinguish factual description from inferred description. A factual description records what is directly observable or documented: “Two women standing outside a shop front, one holding a basket.” An inferred description adds interpretation: “Probably Market Street, Hexham, based on the visible shop signage and comparison with known views of the town.” Both are valuable, but they should be clearly distinguished. Use qualifiers — “probably”, “possibly”, “circa”, “attributed to” — and document the evidence behind any inference.

Unidentified people and places. Do not leave these fields blank if any evidence exists. Record what you can observe — approximate age, clothing style, visible signage, landscape features — and note any partial identifications from inscriptions or donor information. “Unidentified woman, formal dress, possibly 1920s based on clothing” is far more useful than “Unknown” and gives future researchers something to work with.

Date estimation. Where exact dates are unknown, use a range based on available evidence — photographic process, costume, vehicle types, building features, or associated records. Record the basis for your estimate: “Estimated 1910–1925 based on gelatin silver print process and Edwardian-era clothing.” Consistent use of date ranges makes the catalogue searchable even when precision is limited.

Specimen Catalogue Record

To illustrate how these fields work together in practice, here is a specimen record for a single photograph:

Field	Value
Reference number	PH/2026/0142
Title	View of Market Street looking north, showing Woolworths frontage and pedestrians
Date	circa 1935–1950 (estimated from vehicle types and shop signage visible in image)
Creator	Unknown; attributed to J. Henderson based on pencil note on verso
Location	Market Street, Hexham, Northumberland
Subjects	Streets; Shops; Pedestrians; Woolworths
Genre / physical form	Gelatin silver print, 150 × 100 mm, mounted on card
Condition	Fair — slight fading at edges, minor surface abrasion upper left
Inscription	Pencil on verso: “Market St, Hexham — J.H.”
Rights	Copyright status uncertain; creator not confirmed. Flagged for review before online publication.
Provenance	Donated by Mrs A. Henderson, 2019. Part of Henderson family collection (ACC/2019/034).
Digital surrogate	Master: PH_2026_0142_master.tif (600 dpi, 48-bit colour). Access: PH_2026_0142_access.jpg
Physical location	Box 14, Folder 3, Local Studies Store, Room 2

This record is not perfect — the creator is uncertain, the date is a range, the copyright is unresolved. But every field contains what is known, every uncertainty is flagged, and a researcher encountering this record can assess its reliability and decide how to use it. That is what a good catalogue record does.

A complete catalog record using these fields transforms a photograph from an object into a document — one that can be cited, searched, shared, and interpreted by someone who has never seen the physical item.

---

Choosing a Cataloguing Standard

You do not need to implement a single standard in its entirety, but you do need to make a deliberate choice.

SPECTRUM is the UK museum collections management standard, maintained by Collections Trust. It covers the full lifecycle of a collection object, including cataloguing the collection from intake to disposal. If you are a museum or work closely with museum collections, SPECTRUM provides the procedural framework within which your catalogue structure should sit.

Dublin Core is a minimal, widely-used metadata standard. It is not photography-specific, but its simplicity makes it useful for interoperability — particularly if you plan to share records with aggregators like Europeana or the National Archives’ Discovery system.

EAD (Encoded Archival Description) is the standard XML format for archival finding aids. It is well suited to hierarchical archive collections where photographs form part of a larger fonds, and export in EAD3 format is increasingly expected for funded digitisation projects.

For most UK local history societies and smaller heritage organisations, a pragmatic approach is to catalogue using Dublin Core-compatible fields while ensuring your records can be exported in EAD3 if required for reporting or data sharing. The general cataloguing guidelines article covers this in more detail.

---

Manual vs AI-Assisted Cataloguing

There is an honest case for manual cataloguing. A trained archivist working methodically through a collection will produce richly contextualised records. They will recognise that the building in the background is the old town hall demolished in 1962. They will cross-reference the donor file and establish that the photographs almost certainly date from the 1950s.

The problem is time. A careful archivist, working at a realistic pace with a mixed photographic collection, might complete twenty to thirty catalog records in a working day. A collection of 5,000 photographs represents months of work, assuming the staffing exists to do it. For many organisations, the cataloguing backlog simply never clears.

AI-assisted cataloguing does not replace that professional judgement. What it does is handle the mechanical parts of creating an initial catalog record: reading the image, identifying likely subjects and locations, extracting any text visible in the frame, generating a draft description, and pre-populating metadata fields. A human archivist then reviews, corrects, and enriches those draft records — but starting from a populated form rather than a blank one.

The Archiver works this way — upload your photographs and it reads each image, identifies subjects, dates, and place names, and drafts a catalogue record for your review. You correct and approve; it does the heavy lifting.

The practical effect is that a collection which would take three months to catalogue manually might be substantially complete in two to three weeks. The records are not identical to fully manual work — AI-generated descriptions benefit from human review — but the coverage achieved is far greater, and the records are genuinely useful rather than a perpetually deferred aspiration.

---

Step-by-Step: Cataloguing a Photograph Collection

Step 1: Prepare the physical or existing digital material. Before cataloguing begins, take stock of what you have. Count items, identify series or groupings, and note any existing labels or annotations. If working from physical photographs, you will need to digitise them first or work alongside digitisation.

Step 2: Create the digital image files. Scan physical photographs at a minimum of 400 dpi for standard prints; 600 dpi or higher for negatives and items with significant fine detail. Digital images are often created as TIFFs for archival masters and JPEGs for access copies. If you already have digital images from a previous digitisation project, organise them into a clear folder structure before ingest.

Step 3: Upload to your cataloguing system. In an AI-assisted workflow, upload your digital images to the platform. The Archiver accepts bulk uploads and begins processing immediately — reading each image, extracting visible text, identifying subjects and place names, and drafting metadata fields including genre and physical characteristics.

Step 4: Review and correct AI-generated records. Work through the draft records systematically. Check dates against your contextual knowledge. Verify place names against local authority files. Correct any misidentifications. Add provenance information that the system cannot infer from the image alone. This is where professional expertise makes the decisive difference.

Step 5: Apply controlled vocabularies. Standardise subject headings and place names against your chosen authority files. This step is critical for ensuring that the information in the catalog record remains consistent and useful as the collection grows.

Step 6: Export in the required format. Once records are reviewed and approved, export in the format your institution or funder requires — EAD3 for archival finding aids, BagIt for digital preservation packages, or CSV for import into a collections management system.

Step 7: Establish ongoing maintenance. A catalogue is not a project with an end date. New acquisitions need to be catalogued as they arrive. Build catalogue maintenance into your regular workflow rather than treating it as a backlog-clearance exercise.

---

Common Mistakes to Avoid

Inconsistent date formats. Mixing “March 1954”, “03/1954”, and “c.1950s” in the same catalogue makes date-range searching unreliable. Choose a format and stick to it from the outset.

Free-text place names without authority control. “Brum”, “Birmingham”, and “Birmingham City” all mean roughly the same thing, but a search for any one of them will miss the others. Use a standard form and record variants as alternative names.

Skipping the rights field. It is tempting to leave copyright information blank when it is uncertain. Do not. Record what you know, note what is uncertain, and flag items that need legal review.

Over-relying on AI output without review. AI systems are excellent at identifying common subjects and extracting visible text, but they will misidentify unfamiliar local landmarks and struggle with heavily degraded images. Human review is not optional — it is the stage that makes the catalogue accurate rather than merely fast.

No unique identifiers. If your catalog records do not have persistent, unique reference numbers tied to the physical or digital objects, the catalogue and the collection will gradually drift apart. Assign reference numbers before you start.

---

Start Cataloguing Your Photograph Collection Today

The four boxes in that local history society are not unusual — they are the norm. But the barrier to addressing them has dropped significantly. For a deeper look at the standards and workflows that underpin good practice, see our complete guide to archival cataloguing and the features overview for a full picture of what AI-assisted tools can do. If your collection includes objects alongside photographs, our guide to cataloguing a museum collection covers the specific fields and SPECTRUM requirements. For museums, the dedicated museums page explains how AI-assisted cataloguing fits museum workflows, and our community archives page covers tools suited to volunteer-run collections.

The Archiver reads your photographs, extracts dates, subjects, and place names, drafts catalog records, and exports in EAD3, BagIt, or CSV format. You review, correct, and approve — the platform does the heavy lifting.

If you are sitting on an uncatalogued collection, there is no better time to start. Request early access to try The Archiver on your own collection.