What the results mean

Every photo goes through 3 checks, one after the other. The answer shows the result of each check. This page explains them in plain words.

Photo check: is the photo good enough?
Crop check: which crop is it?
Problem check: what’s wrong with the plant?

If something goes wrong in a check, the answer tells you why in errors, and crop_health is unknown.

The answer at a glance

A shortened example for a photo of a pepper plant with thrips:

{
  "image_feedback": { "focus": "good", "distance": "good" },
  "crops": ["pepper"],
  "crop_health": "unhealthy",
  "predicted_diagnoses": [
    {
      "common_name": "Chilli Thrips",
      "scientific_name": "Scirtothrips dorsalis",
      "diagnosis_likelihood": "very_likely",
      "pathogen_class": "insect",
      "eppo_code": "SCITDO",
      "peat_id": 600136
    }
  ],
  "errors": [],
  "plantix_trace_id": "1-67891233-abcdef012345678912345678"
}

1. Photo check

Plantix first looks at the photo itself: is it sharp, and was it taken close enough to the plant?

image_feedback grades the photo with good, mediocre, bad or very_bad:

Field	What it means
`focus`	How sharp the photo is.
`distance`	Whether the photo was taken close enough to the plant. Empty when no plant was found.

What can go wrong here:

`errors[].type`	What it means
`image_blurry`	The photo is blurry.
`distance_too_far`	The photo was taken too far away from the plant.
`bad_image`	The photo may have other problems. Check `image_feedback`.
`non_plant`	There is no plant in the photo.

2. Crop check

Next, Plantix finds out which crop is in the photo, or uses the crop you sent.

Field	What it means
`crops`	The crop: the one you sent, or the one Plantix recognised. When crops look very much alike, you get a crop group instead.
`hosts`	Part of each problem found: the crops this problem affects. Useful when you get a crop group.

What can go wrong here:

`errors[].type`	What it means
`crop_not_supported`	A crop was found, but the API doesn’t support it yet. See Supported crops.
`ornamental_plant`	The plant is an ornamental plant.

3. Problem check

Finally, Plantix looks for diseases, pests and nutrient deficiencies.

`crop_health`	What it means
`healthy`	No problem found. The answer can still list possible problems, so always look at how sure we are.
`unhealthy`	At least one problem found.
`unknown`	Something went wrong in one of the checks. See `errors`.

Each problem found (predicted_diagnoses) has:

Field	What it means
`common_name`	The everyday name of the problem, in the language you asked for.
`scientific_name`	The scientific name.
`diagnosis_likelihood`	How sure we are, see below.
`pathogen_class`	What kind of problem it is, see below.
`symptoms`	What you see on the plant.
`symptoms_short`	The same, as a few short points.
`trigger`	What causes the problem and how it spreads.
`preventive_measures`	How to prevent it.
`treatment_chemical`	Chemical treatment.
`treatment_organic`	Organic treatment.
`image_references`	Example pictures of the problem, to compare with your photo. Available in different widths, see Image sizes.
`peat_id`, `eppo_code`	Codes to link your own products or information to the problem, see PEAT IDs & EPPO codes.
`hosts`	The crops this problem affects.

For a healthy plant, trigger, preventive_measures and the treatments are empty (null).

How sure we are

Plantix compares the photo with known cases of the problem. The closer the match, the higher the level.

The five levels, from most to least sure:

`diagnosis_likelihood`
`very_likely`
`likely`
`possible`
`unlikely`
`very_unlikely`

What kind of problem

`pathogen_class`	Examples
`fungi`	Powdery mildew, apple scab
`virus`	Tomato spotted wilt virus
`bacteria`	Bacterial canker, bacterial wilt
`weed`	Weeds (none in the crop problems list today)
`insect`	Aphids, thrips
`mite`	Mango leaf coating mite
`deficiency`	Nitrogen deficiency, potassium deficiency
`additional`	Sunburn, herbicide damage, fruit cracking
`others`	Healthy plant, others

What can go wrong here:

`errors[].type`	What it means
`unknown_disease`	Plantix couldn’t recognise the symptoms. Most likely this problem isn’t supported yet.

Reference number

Every answer has a plantix_trace_id. Please include it when you contact us about a result.

← All guides