Cameras disagree with each other and with themselves. Two cameras see the same leaf as two different greens, one camera sees it differently on Tuesday than on Monday, and nothing in a bare photo says how big anything is or which bench you were even looking at. The Station Card is the cheap, physical answer: one printed-and-assembled card that sits at the edge of every frame and carries the references and the station's identity together, so any frame from any camera can be read, scaled, color-corrected, and filed to the right place.
01The idea: put the truth in the frame.
Everything a photo does not know about itself, the card supplies from inside the picture. Which station is this, how many millimeters is a pixel, what does true white look like under this light, today. Because those answers travel in the frame, they survive everything that normally breaks a photographic record: the camera gets bumped, replaced, upgraded, or swapped for the phone in your pocket, and the card still tells each frame how to be read. The expensive way to get comparable data is to buy identical calibrated cameras and never touch them. The cheap way is cardstock.
This is the same thinking as the rest of the imaging shelf: the discipline lives in the method, not the hardware, and the method is open. And it is deliberately one concept applied everywhere a lens looks: the same card rides the phone walk-through, the fixed plant station, the field stake, and the views of cameras you already own. Each build page on the Configs shelf has a "Station Card here" section saying exactly how the card applies there, including the two honest exceptions: the witness tier, where it is optional, and thermal, which cannot see a printed card at all and carries the reference job on wet and dry reference leaves instead.
02Five jobs, one card.
One card, by deliberate decision: the identity and the references stay together, so there is nothing to mismatch and nothing to forget. What it carries:
| A machine-readable marker | identity + geometry | A high-contrast printed marker (the kind robotics uses) that names the station and lets software work out exactly where the camera was relative to the card. Move the camera and the frames still line up. |
|---|---|---|
| A scale bar | size | A printed ruler pattern, so every frame carries its own millimeters-per-pixel. The resolution rule becomes checkable per frame instead of taken on faith. |
| Color patches | color | Known printed colors, so white balance and exposure can be corrected after the fact. This is what lets a color change in the crop be separated from a mood change in the camera. |
| An infrared reference patch | NIR | A patch of known infrared brightness, for stations doing near-infrared work. It is a piece of material, not printed ink, for a reason worth a whole section below. |
| Human-readable label | people | The station name, in letters, because the first user of any record is a person squinting at a thumbnail. |
One deliberate omission: the wet and dry reference surfaces that make thermal imaging honest are not on the card, because a card cannot stay wet. They are their own small build, a pair of artificial reference leaves, described where the thermal story lives: the Crop Water Stress Index.
03Why this makes your phone an instrument.
A current phone has a better sensor, better optics, and better processing than anything in the affordable fixed-camera class, and you already own it. What it lacks is everything the card supplies: it does not know which station it is pointed at, what scale it is seeing, or how its heavy automatic processing colored the frame today. Card in frame, those gaps close, and the walk-through you were doing anyway becomes a data collection round.
That changes what to buy. Fixed cameras buy you the unattended interval, the frames captured while nobody is there. The phone buys you resolution on demand, the close-up when something looks wrong. They are complementary, and the card is what lets both feed one comparable record. The cheapest instrument in this whole catalog is the phone in your pocket and a piece of card.
04The materials, and the two traps.
Most of the card is ordinary printing on cardstock. The infrared patch is where two traps live, and they are the two places a homemade reference quietly lies to you.
- Printed ink is not infrared-neutral. A gray that looks perfectly neutral to your eye can be wildly bright or dark at near-infrared wavelengths, because printer inks behave completely differently out there. Print the infrared patch and your calibration is measuring the printer, not the plant. The patch must be a material with known infrared behavior, mechanically held in a window in the card. Plain white PTFE sheet, the slippery white plastic sold as sheet stock, is the working choice: it reflects strongly and evenly far past the visible range, it cuts with a knife, and it is cheap. It has one personality quirk that matters: nothing glues to it, so the card design holds it in a frame instead of pretending an adhesive will.
- Thin PTFE is see-through to light. The white plumber's tape on the same shelf is chemically right and physically wrong: it is so thin that light passes into it and out the back, so a few wraps of tape measures the cardstock behind it. The patch needs real thickness, millimeters of solid sheet, before it reflects like a reference instead of a film.
05A card is a consumable.
In a greenhouse, everything collects dust, algae, and overspray, and a reference that has drifted is worse than no reference, because it corrupts the record while looking official. So the card is designed to be replaced, not preserved: the printed part re-prints for pennies, the infrared patch pops out of its window and a fresh piece drops in, and the marker carries the station identity forward so a replacement card is the same station, continued. Plan on swapping cards the way you swap gloves, on a schedule you will actually keep.
06What the card cannot fix.
Honest limits, so the card earns trust where it does work:
- It cannot add resolution. If the pest is smaller than a pixel, no reference recovers it. The calculator tells you what your framing can see; the card just proves it per frame.
- It corrects the correctable. Exposure, white balance, and scale are the fixable errors. Two different sensor models still see color slightly differently at a depth no card can equalize, which shows up as a small step in the record the day a camera is replaced. The card makes that step small and visible instead of large and invisible.
- It does not do thermal. Thermal reference surfaces have to transpire and evaporate to mean anything, which paper cannot. That job belongs to the reference leaves.
07Where this stands.
The design is published; a printable, marked draft you can cut and assemble is the next artifact, and it is not on this page yet. We have not run cards through a season, and when we have, what we learn lands here. The design is open under CC BY, like the rest of the site: print it, improve it, tell us what failed.
Frequently asked questions.
How do I make photos of my plants comparable over time?
Fix the geometry, lock the camera's automatic settings, and put a reference in every frame. A printed card carrying a scale bar, known color patches, and the station's identity lets any frame be corrected and compared afterward: the scale says how big things are, the colors say what the light did, and the identity says where it was taken. Without a reference, day-to-day differences in exposure and white balance swamp the changes you actually care about.
Can I use my phone for plant monitoring instead of buying cameras?
Yes, and for close-up diagnosis it beats most affordable fixed cameras, because phone sensors and optics are better and you already own one. Its weaknesses are that it needs a person holding it and that its heavy automatic processing varies between shots. A reference card in the frame corrects most of that. Fixed cameras still earn their place for the unattended interval, the frames captured on schedule while nobody is walking the rows.
What is a color reference card for in photography?
It puts patches of known color in the picture, so software can work out what the light and the camera did to the frame and undo it. Corrected against the card, a photo's colors reflect the subject rather than the lighting or the camera's automatic guesses, which is what makes color comparisons across days, cameras, and locations meaningful.
Why can't I just print a gray reference card at home?
For visible-light color correction, a well-printed card genuinely helps. The trap is infrared: printer inks that look neutral gray to your eye behave unpredictably at near-infrared wavelengths, so a printed patch calibrates your printer instead of your camera. An infrared reference has to be a material with known infrared behavior, such as a thick piece of plain white PTFE sheet, held mechanically in the card.