How to sort trackday photos by rider in minutes

Every motorsport photographer knows the trap. The shooting is the easy part. The evening is where the money leaks away: culling thousands of near-identical frames, working out which blur of leathers belongs to which rider, and building a folder for each one. By the time the galleries go up a day or two later, the riders have driven home and the buying urge has cooled.

This guide walks through a workflow that closes that gap, so you can deliver the same day while riders still have the adrenaline in their veins. It works for motos, cars and karts, and for both stills and video.

Why sorting by rider is the hard part

Culling for quality is a solved problem. Selecting sharp frames, dropping the throwaways, and applying a preset is something most photographers already do quickly. The bottleneck is identity: deciding which rider is in each frame so the right person gets the right folder.

The obvious idea is to read the number on the bike or the bib on the leathers. In practice that breaks exactly where racing lives. The number is often hidden behind another rider, smeared with mud, motion-blurred at the apex, turned away from the camera, or simply not on the bike at a trackday. Any workflow that depends on reading a number inherits every one of those failures, and a single misread sends a photo into the wrong rider's folder, which is worse than no sort at all.

The alternative is to match on appearance: the helmet design, the livery of the bike, the colours and cut of the riding kit. Those stay visible and consistent across a whole session even when the number does not. This is the approach RaceLabs uses, and it is why the sort keeps working when a number-reading tool would give up. You can read the measured accuracy, including mean Average Precision of 0.99 and 99.6% same-rider precision on riders the system had never seen, on the research and validation page.

The step-by-step workflow

1. Shoot as normal and back up the card. No special settings, no markers, no asking riders to hold a number to camera. Copy the card to your working disk first so you always have the originals untouched.
2. Do a fast quality pass, or skip it. If you already cull for sharpness in Lightroom or Photo Mechanic, do it now. If you would rather sort first and cull inside each rider's folder, that works too. The sort does not care about order.
3. Send the folder through RaceLabs. Point the app at the folder of photos and video. It compresses the frames, sends roughly one percent of the bytes to datacenter GPUs for the heavy analysis, and keeps your full-resolution originals on your own disk. There is no upload of the whole shoot and no locked online gallery. The mechanics are covered on the performance page.
4. Let it group by rider. The system clusters every frame by appearance and writes one folder per rider. A photo with several riders in it lands in each of their folders, so nobody misses a shot they are in. Frames it is genuinely unsure about go to a review bin instead of being guessed into the wrong folder.
5. Review the bin and deliver. Glance through the small review bin, drag any edge cases into place, and you have a clean folder per rider on your disk. Deliver them however you already do: a shared drive, a gallery host, or a direct handover. Because it is plain folders, nothing locks you into one platform.

Video is sorted the same way. RaceLabs analyses the footage and cuts a clip per rider, so a rider can buy their stills and their on-track video from the same delivery.

How fast, and what it costs

Speed is the whole point of same-day selling. RaceLabs analyses up to 7,000 images per minute, so a typical trackday card is sorted in minutes rather than an evening. More than 6,000,000 photos have been sorted through it to date at 99% sorting accuracy.

Pricing is pay-as-you-go and public: from €0.01 per photo, dropping to €0.004 as your volume grows over the year. A typical trackday of three to five thousand photos lands around €30 to €50. Billing is per photo processed, counting every frame the engine analyses, and your first event is free so you can run a real card through before you pay anything. The full breakdown is on the pricing page.

Reading numbers versus matching appearance

If you are comparing tools, this is the distinction that matters most, because it decides how often the sort quietly fails on the frames you most want to sell.

• Number-reading (OCR) tools depend on a clean, readable plate or bib in the frame. They do well on posed grid shots and struggle on action: hidden numbers, mud, blur, riders facing away, or trackdays with no numbers at all. A misread is silent and puts a photo in the wrong folder.
• Appearance matching reads the helmet, livery and kit, which stay visible through a full session. It holds up on exactly the action frames where numbers disappear, and when it is unsure it sends a frame to review rather than guessing.

Neither approach is magic, and appearance matching has its own hard cases, such as two riders in near-identical rental leathers or a rider who swaps bikes mid-day. The difference is that those cases surface in a review bin instead of ending up silently misfiled. For photographers shooting real track action rather than static grids, matching on appearance is the more reliable base to build a same-day workflow on.

Frequently asked questions

Do riders need to show their number to the camera?

No. RaceLabs matches on the helmet, bike livery and riding kit, so it sorts photos even when the number is hidden, blurred, dirty or absent. Trackdays with no numbers at all work fine.

How long does a full trackday take to sort?

Minutes. The engine analyses up to 7,000 images per minute, so a typical card of a few thousand frames is sorted while you pack up, ready to deliver the same day.

Where do my photos go, and do the originals leave my disk?

Your full-resolution originals stay on your own disk. Only a small compressed fraction of the data is sent to EU (Germany) GPUs for analysis, and the result is written back as normal folders. There is no locked online gallery.

How is it different from tools that read the number plate?

Number-reading tools fail whenever the plate is not cleanly visible, which is common in action photography, and a misread silently misfiles the photo. RaceLabs matches on appearance, which stays visible across a session, and sends genuinely unclear frames to a review bin instead of guessing.

What does it cost?

Pay-as-you-go from €0.01 per photo, dropping to €0.004 at volume, billed per photo processed. A typical trackday is around €30 to €50, and your first event is free.

← Back to the RaceLabs blog