Crop Diagnosis & Product Recommendations Through AI

Millions of dollars are spent annually by farmers to manage crop diseases, yet until recently, they frequently lacked access to reliable diagnostic equipment. Numerous crop disease strains can be eliminated using targeted therapies, but it's crucial to use the right one. 

Farmers may turn the best treatments ineffective if they are armed with erroneous diagnoses.

Crop disease is extremely difficult to identify without extensive subject matter knowledge. Only when the disease being treated has been appropriately diagnosed, treatment advice can be useful. But the problem is that pathologists can not visit every field.

The solution is to bring the data to the pathologists. For this, a diagnostic process has been designed that occurs through the use of advanced image analytics powered by artificial intelligence and machine learning.

Google's ML Engine's sophisticated deep learning capabilities were used to build disease classification capabilities. To train the neural network, Google's high-performance platform and more than 50,000 photos are utilized. The neural network's training time is significantly shortened using Google's Tensorflow Processing Units (TPUs), enabling quick and affordable model updates when new photos are gathered and filtered.

A farmer can use the smartphone application to snap pictures of the infected leaves on their crops. The GCP-hosted ML services get these photos, and they promptly provide the farmer with a diagnostic. Scalability is offered by ML Engine, a hosted, serverless platform, without the requirement to manage a number of servers.

All in all, a digital solution was built that allowed accurate product recommendations to be given to farmers all over the world without requiring to send plant pathologists out into the field each time a diagnosis is needed.