Photo Restore LabBring old photographs back to life
Get Photo Restore free

Bring old photographs back to life

How to clear duplicates out of a scanned photo archive

A box scanned over several sessions ends up with the same photograph saved two or three times, and sometimes a better rescan that only looks like a duplicate. Here is a safe way to find them, tell the two kinds apart, and clear them out without losing anything.

By Fable · 11 November 2026 · 9 min read

Photograph: Suzy Hazelwood / Pexels

A box of four hundred photographs rarely gets scanned in one sitting. It gets scanned across a few weekends, sometimes a few months apart, and somewhere in the middle of that the same photograph goes through the scanner twice. Nobody did anything wrong. A folder just quietly filled up with two or three copies of the wedding photo, the christening photo, and the one of the dog on the porch that everyone seems to have taken home a print of.

Clearing those out feels like a five minute job and it is not, because two very different things both look like a duplicate on a thumbnail, and only one of them is safe to throw away.

Exact duplicates and near duplicates are different problems

An exact duplicate is the same file, or an identical copy of it. Same pixels, same resolution, same everything, usually because a scan got saved twice, copied into two folders during an import, or backed up and then merged back into the working folder. These are genuinely safe to delete, because removing one changes nothing about what you own. The information is not reduced by one bit.

A near duplicate is a different capture of the same photograph. The same print scanned twice at different settings. A second original print from the same negative, slightly different in exposure or wear. The same photograph photographed once and scanned once, because somebody was not sure which method they had already used. These look the same at thumbnail size and are not the same file, and the two copies are very often not equal in quality.

Treating a near duplicate like an exact one is the mistake that costs something later. If you delete whichever copy a tool happened to list first, you have a genuine chance of keeping the worse scan and losing the better one, with no way to tell afterwards that a better version ever existed.

Why a rescan can look exactly like a duplicate

This is the case that matters most and the one automated tools handle worst. Say a box was scanned early on at a low setting, before anyone had checked our guide on what DPI to scan old photos at. Months later, the same prints got rescanned properly at a higher resolution, maybe after the DPI advice sank in or a new scanner arrived. Both scans show the same picture. One of them actually has far more information in it.

Any tool that matches images by how they look will flag these two files as duplicates, because they are visually the same photograph. It has no way of knowing that one file is 1200 pixels on the long edge and the other is 4800, or that one was scanned in colour to keep the paper tone and the other was scanned in greyscale and threw it away, a choice our guide on scanning black and white photos in colour or greyscale covers if you want to check which one you are holding. A tool that auto-deletes on a visual match will happily remove the better file and keep the worse one, because file size and resolution are not what it is comparing.

This is the single reason to never let software decide and delete in the same step. Finding candidates automatically is fine. Choosing between them needs a person looking at both files at full size, for at least long enough to check which one is actually larger and sharper.

A safe order of operations

The order that avoids an accidental loss is always the same, whatever tool you use to find the candidates in the first place.

1. Quarantine, never delete directly. Move suspected duplicates into a holding folder rather than deleting them on the spot. Call it something obvious like _review-duplicates and keep it inside the main archive, not in a separate place you will forget about. Nothing in this folder is gone, it is simply out of the way while you decide.

2. Review each group at full size. Open both files, not the thumbnails, and check pixel dimensions and file size before deciding anything. A quick way to sanity check which scan should have won is our scan output calculator, which shows what resolution a given print should produce at a given setting, useful for confirming that the larger file really is the better scan and not just a bigger export of the same thing.

3. Keep the better file, and only the better file. Higher resolution almost always wins. Colour almost always beats greyscale for a black and white print, because the paper tone and any fading are information you cannot add back later. If genuinely nothing distinguishes the two, keep either one and move on rather than agonising over a coin flip.

4. Delete from the quarantine folder last, and only after your backup has caught up. This is the point where loss actually becomes permanent, so it is worth not rushing it. If your backup runs on a schedule rather than manually, as it should under the plan in our guide on backing up scanned family photos, wait for one backup cycle to complete before clearing the quarantine folder, so a mistake in step three is still recoverable from an earlier copy.

Quarantine first, decide second, delete last, and never collapse those into one step no matter how confident the tool’s matching looks.

What automated tools are actually good for

Said plainly, because it would be dishonest to pretend otherwise: for a few hundred or a few thousand files, looking at every photograph yourself to spot duplicates is not realistic, and automated matching is the only practical way to surface candidates. That part of the job is worth automating.

What is not worth automating is the decision. Treat a duplicate finder as a tool that produces a shortlist for you to review, the same way a smoke alarm does not put out the fire, it tells you where to look. Review the shortlist against the order of operations above, and the tool has saved you real time without costing you a single irreplaceable file.

The photographs this never catches

A faded, water-stained or damaged photograph is not a duplicate of anything, but it is worth mentioning here because the same instinct that clears duplicates can tip into clearing damage too quickly. A print that looks washed out next to a crisper scan of the same picture is not automatically the worse file to keep, if the crisper scan is actually of a different, later print. Our page on faded and darkened photos is worth a look before you decide a dull-looking scan is disposable, since fading is frequently recoverable and the original scan is the only copy a recovery can ever be made from.

The one habit that prevents this next time

Duplicates mostly happen because a scanning session has no record of what was already done. The fix is not a better tool, it is a running note, even a plain text file, listing which folders or boxes have been scanned and when. Five minutes at the start of each session checking that note against what is in front of you costs far less than an afternoon sorting duplicates out of four hundred files later.

A scanned archive is worth the same care as the box it came from. Clearing it of genuine duplicates makes it easier to use. Clearing it carelessly can quietly undo scanning work that took a whole weekend to do properly the first time.

Common questions

How do I find duplicate photos in a scanned archive?
Most photo managers and a few dedicated tools can group visually similar images automatically, which is the fastest way to surface candidates in a few hundred or a few thousand files. Automation is good at finding candidates and bad at deciding what to do with them, so treat every result as something to review, not something to delete on the tool's say so.
Is it safe to automatically delete duplicate photos?
Not for a family archive. Automatic deletion works well for a camera roll full of accidental bursts, but a scanned collection mixes exact duplicates, a double print from the chemist, and rescans of the same photograph at a better setting that only look identical at a glance. Deleting automatically treats all three the same, and only the first one is actually safe to lose.
What is the difference between an exact duplicate and a near duplicate?
An exact duplicate is the same file, or a copy of it, with nothing different about the image itself. A near duplicate is a different capture of the same photograph, a rescan, a second print from the same negative, or the same print photographed twice. Near duplicates often differ in resolution or cropping even when they look the same on a thumbnail, so the better one is worth keeping rather than whichever one a tool found first.
Why would the same photo be scanned twice?
Usually because a box was scanned across more than one session, sometimes months apart, and nobody kept a record of what had already been done. It also happens on purpose, when an early scan was done quickly at a low setting and a later rescan corrects it. That second case is the one worth knowing about before you clear anything, because the "duplicate" is actually the better file.