Crop.health: curing crops, empowering farmers

While winter is a time when not much happens in nature, Kindwise has been anything but dormant. We have been busy preparing a new service to help farmers and gardeners around the world protect their crops.

We are happy to introduce our latest product - crop.health, an innovative solution designed to identify diseases and pests affecting crops from images. It was born directly out of the imagination of entrepreneurs developing technology for farmers, who tested our current solutions and wanted more. Thanks to this collaboration, we've ensured that crop.health is not only technically sound but also practical and easy to integrate into existing farming practices.

Try our demo!

High precision, comprehensive insights

Even in its beta version, crop.health can identify 288 diseases and pests on 23 major crops with 93% top-3 accuracy and 85% top-1 accuracy, on our validation dataset. On an independent dataset of images collected from Twitter, the model performed a 66% top-3 accuracy rate and 48 % top-1 accuracy.

See our public collection to get a quick idea about the capabilities of crop.health. 

Here are four key highlights of crop.health:

• Rich insights: In the API response, you get not only the diagnosis of the diseased crop, but much more, including treatment instructions, symptoms, and disease severity. See our documentation for full details.
• Technical wizardry: Merging years of experience in machine learning, plant disease identification, and API service delivery ensures our product's high accuracy and reliability.
• Free demo and affordable pricing: Explore our service for free and check out our bulk discounts. 
• Continuous Improvement: We are ready to improve crop.health based on early feedback. What's more, we have big things planned for the future, including adding more classes, focusing on more crops, and implementing large language models. See the full roadmap.

From common pests to deadly diseases

The API covers nearly 288 classes of health issues, including:

• Pests (such as Colorado potato beetle, aphids, and grasshoppers)
• Bacterial diseases (such as bacterial spots and speck, Pierce's Disease, fire blight)
• Fungal diseases (such as early blight, grey mold, powdery mildew)
• Viral diseases (such as Apple mosaic virus, Potato Virus Y, Banana bunchy top virus)
• Nutrient deficiencies

Our initial  23 crops have been carefully selected based on the needs of our customers and global crop production data. Here is the full list:

• Apple (Malus domestica)
• Banana (Musa spp.)
• Barley (Hordeum vulgare)
• Cassava (Manihot esculenta)
• Citrus (Citrus spp.)
• Cocoa (Theobroma cacao)
• Coffee (Coffea spp.)
• Corn (Zea mays)
• Cotton (Gossypium spp.)
• Cucumber (Cucumis sativus)
• Eggplant (Solanum melongena)
• Garlic (Allium sativum)
• Grapevine (Vitis spp.)
• Oil palm (Elaeis guineensis)
• Onion (Allium cepa)
• Potato (Solanum tuberosum) 
• Rice (Oryza spp.)
• Soybean (Glycine max)
• Sugarcane (Saccharum officinarum)
• Tea (Camellia sinensis)
• Tobacco (Nicotiana tabacum)
• Tomato (Solanum lycopersicum)
• Wheat (Triticum spp.)

For farmers and gardeners

Our API can be easily integrated into any app or other solution, with the end users being anyone who takes care of plants that are produced mostly for consumption. Here are 2 main end-user use cases we thought about when designing this service. 

1. Small farmers in developing countries

Suppose you are a smallholder farmer in a developing country, how do you get information on how to take care of your crops? Sometimes it is difficult to get the right information or to find an expert. Fortunately, access to smartphones and the Internet has become recently quite common. This is why online plant disease detection is a perfect fit for the farmer's needs and can help them identify and treat the health problem in time. 

2. Gardeners around the world

Growing food is a means for many but a source of joy for others. The ability to have a complex solution to help them understand how to care for it and treat and prevent disease can be just the thing to make gardening easier.

‍

Kindwise already provides a plant disease identification service, so what exactly is the difference? 

Plant.id Health Assessment is not focused at any specific group of plants but indoor plants have been most frequently asked about. With crop.health, we focus more on edible plants. Here are three main differences:  

• The list of diseases for crop.health has been carefully selected based on their occurrence in each of the crops. It focuses more on pests and infectious diseases, whereas types of abiotic diseases are common in the Plant.id Health Assessment , reflecting how common these issues are in different plant groups.. 
• The information provided in crop.health is tailored to the needs of farmers, including internationally recognized databases like the EPPO database.
• User-base: crop.health has a diverse user base, including small farmers and gardeners, as well as larger agribusinesses. Plant.id Health Assessment may be more appropriate for individual plant enthusiasts.

‍Here is a detailed table showing the difference between these two products:

Example of a typical identification in crop.health and Plant.id Health Assessment:

Discover crop.health

One of the benefits of using Kindwise products is our easy-to-use, cross-product platform. Log in to our admin panel, create a new API key, and get started.

Start exploring crop.health today and unlock a world of possibilities in crop disease detection and management. We are happy to provide you with free credits for testing - let us know at business@kindwise.com.

Start the journey now!

‍