Plant.id Health Assessment

Identify plant diseases with Plant.id Health Assessment

Plant Health Assessment is a feature of Plant.id API. It provides information about the overall health of the plant (is_healthy binary classification) as well as diagnostics of a specific problem (diseases classification).

With Plant.id Health Assessment you can recognize:

  • 39 classes of fungal diseases
  • 15 classes of abiotic disorders
  • 14 classes of pests
  • 9 classes of chromista
  • 7 classes of viruses
  • 6 classes of bacterial diseases

Model specification

100 diseases classes

We can identify more specific issues (such as thrips, sooty mold, and nutrient deficiencies) and generic ones (such as abiotic and fungi). The model has been trained mostly on houseplants.

49.5% accuracy

The precision of the binary is_healthy classification is 95%, the recall is 75.5%. The model for 100 classes returns an exact match in 49.5% of cases, the TOP3 accuracy is 57.1%.

{
   "is_healthy_probability":0.001715570500000041,
   "is_healthy":false,
   "diseases":[
           {
         "classification":[
            "Abiotic",
            "Water-related issue"
         ],
         "name":"Water excess and/or uneven watering",
         "probability":0.910850346
      },
      {
         "classification":[
            "Abiotic",
            "Water-related issue"
         ],
         "name":"Root rot",
         "probability":0.1706690415
      },
   ],
   "diseases_simple":[
      {
         "name":"Abiotic",
         "probability":0.9738712015
      }
   ]
}

Pricing

The base price is €0.1 per request. There are significant discounts for higher volumes of identifications. Tell us what you need on business@plant.id.

We provide prepaid and retroactive models. You can pay with a card, wire transfer, or PayPal.

Future plans

We are continually feeding the model with fresh photos from our annotation pipeline and will be releasing new models on a monthly basis in 2021. We are planning the following improvements in the next releases:

  • Accuracy improvements; especially regarding garden plants.
  • More classes and finer level of health disorder detection; for example Glomerellaceae will be returned instead of just Fungi.
  • Tree-like structure of the results; for example, “Fungi > Sordariomycetes > Glomerellaceae > Colletotrichum”.
  • Adding content (e.g. URL or description per issue or issue category)