This is a beta version of the service and it may undergo changes that may not be compatible with the current version.
\nCrop.health is a machine learning-based service to identify crop diseases and pests from images. To use the API you need to get an API key.
\nTo use API from python, use SDK: kindwise API client.
\nThis is a description of an important part of the API. The specification of the endpoints is given below. In the right column, you can find example calls of these endpoints (on the right you can toggle between different examples).
\nYou can also open this documentation directly in Postman, where you can see endpoints and examples and try it out (after providing your API key).
\nMore implementation examples can be found here.
\nHere is openAPI definition of this api.
\nYou should provide the Api-Key header with your call. You can request the API key here.
There are two methods to create a new identification:
\nusing application/json request - all attributes including images are encoded in json body - see the example
using multipart/form-data request - attributes are sent in text fields; images are sent as files - see the example
Attributes:
\nimages
application/json: a list of strings (required) - one or more images (encoded as strings in Base64) of the crop you want to identify
multipart/form-data - all images in file fields (names of file fields are not important)
latitude - float (optional) - geographic coordinate, increases the identification accuracy
longitude - float (optional) - geographic coordinate, increases the identification accuracy
similar_images - boolean (optional) - if true, the model also selects images similar to the provided images for each suggestion
custom_id - int (optional) - a unique identifier, can be used to access the identification results
datetime - str (optional) - to improve results, e.g. for images taken at different seasons, specify the date when the images were taken in ISO format with variable precision (e.g. 2023-06-22 for days or 2023-06-22T11:28 for minutes)
Here we comment on important parts of the result. For a complete example of the results see the response examples below.
\naccess_token - a unique identifier of the identification, can be used to access the identification results
result->crop->suggestions - predictions of the model containing plant taxa suggestions and additional information; each suggestion consists of
id - the id of the suggested taxon, which doesn't change over time
name - English name of the taxon (e.g. potato) - this name can change over time
scientific_name - Latin name of the taxon (e.g. Solanum tuberosum) - this name can change over time
probability - the level of certainty that suggested taxon is the one in the image
similar_images - if similar_images=true is in request - list of representative images of the identified taxon carefully selected by the model to resemble the input image; the images may also contain license data that should be cited when using the image
details - additional info you requested, see below
result->disease->suggestions - predictions of the model containing suggestions including diseases, pests and healthy categories. Additional information is similar to crop suggestions and consists ofid, name, scientific_name, probability, similar_images, details.
There is a list of additional information about crops and diseases that you can get in the identification result. To get specific information, include the detail's name in the GET parameters.
\nThe default language is English; to get details in a different language include the language parameter in the GET parameters.
\nSome details are licensed. If you want to use this detail, you should also cite the source.
\nSome details are not available for all entities, if unavailable, null is returned.
This is a basic description of the available details. To see examples of return values and their exact format, see this example with all details for a single language or this example with multiple languages.
\nCrop details
\ngbif_id - int - id in GBIF database
image - string - with licence - url of representative image
images - list - with licences - list of more urls of representative images
Diseases details
\ntype - string - high-level taxonomical category (e.g. Fungi)
common_names - list of strings - localized - local, non-scientific names
taxonomy - dict - scientific taxonomy
eppo_code - str - id in the EPPO database
eppo_regulation_status - dict - quarantine status from EPPO database
gbif_id - int - id in the GBIF database
image - string - with licence - url of representative image
images - list - with licences - list of more urls of representative images
wiki_url - string - localized - link to disease profile page; if localized page not available, English one is provided
wiki_description - string - localized, licensed - short description from Wikipedia
description - string - short description, by kindwise
treatment - dict - treatment instructions in biological, chemical and prevention categories, by kindwise
symptoms - dict - description of symptoms, by kindwise
severity - string - static description of seriousness of the disease, by kindwise
spreading - string - description of how the disease spreads, by kindwise
Send images for identification. See above for description of response and request body.
\n","auth":{"type":"apikey","apikey":{"basicConfig":[{"key":"key","value":"Api-Key"},{"key":"value","value":"your_api_key"}]},"isInherited":true,"source":{"_postman_id":"b067e857-2490-4875-8b38-1b6bba5d3a92","id":"b067e857-2490-4875-8b38-1b6bba5d3a92","name":"Crop.health","type":"collection"}},"urlObject":{"path":["api","v1","identification"],"host":["https://crop.kindwise.com"],"query":[{"disabled":true,"description":{"content":"comma separetad list of requested crop and disease details (optional)\nsee detail description above
\n","type":"text/plain"},"key":"details","value":"type,common_names,eppo_code,wiki_url,taxonomy"},{"disabled":true,"description":{"content":"two-letter ISO 639-1 language code (optional, default en) - language of localized details\ncan be multiple (up to 3), separated by comma - i.e. (en,de,sv)
boolean (optional) - if true: request does not wait for identification to complete and response does not contain result section; result can be later retrieved by next endpoint
Can be used to
\nasync=Truecomma separetad list of requested crop and disease details (optional)\nsee detail description above
\n","type":"text/plain"},"key":"details","value":"type,common_names,eppo_code,wiki_url,taxonomy"},{"disabled":true,"description":{"content":"two-letter ISO 639-1 language code (optional, default en) - language of localized details\ncan be multiple (up to 3), separated by comma - i.e. (en,de,sv)
access_token retrieved on identification creation (or provided custom_id)
access_token retrieved on identification creation (or provided custom_id)
Can be used to get API key statistics. To:
\nactive)credit_limits for day, week, month or total)used by your API keycan_use_credits => hasn't reached the limit and has enough credits left)remaining credits of your API keyCan be used to send a user feedback to an identification.
\nSpecifics:
\nrating - a number from 0-5 or comment - a text comment from the user. Or include bothFeedback can help us improve our model, so thank you for sending it.
\n","auth":{"type":"apikey","apikey":{"basicConfig":[{"key":"key","value":"Api-Key"},{"key":"value","value":"your_api_key"}]},"isInherited":true,"source":{"_postman_id":"b067e857-2490-4875-8b38-1b6bba5d3a92","id":"b067e857-2490-4875-8b38-1b6bba5d3a92","name":"Crop.health","type":"collection"}},"urlObject":{"path":["api","v1","identification",":access_token","feedback"],"host":["https://crop.kindwise.com"],"query":[],"variable":[{"description":{"content":"access_token retrieved on identification creation (or provided custom_id)