Overview / Usage

One of the many ways in which computerization in agriculture has made huge strides is in the detection of different plant diseases. Almost every country is now focusing on mechanizing agriculture to make it more accurate and precise and to keep up with the rising demand for food. Plant disease detection is one of the hardest parts of farming, and it has a big effect on how much food is grown.

Plant diseases hurt the quality of vegetables, organic goods, grains, and vegetables in a big way. Since losses in money are closely watched, it is important to come up with quick and effective ways to find and evaluate plant diseases. This project looks at how machine learning models can be used to improve the early stages of plant disease detection in order to make the grain supply more secure and the agro-biological system easier to manage.

In this solution a neural network built with Tensorflow scans images of plant leaves. The inference system based on the neural network model detects plant diseases. The disease is then automatically looked up to get the measures that can be taken against it.

Methodology / Approach

This is a solution for combating plant diseases by using a neural network to scan images for plant diseases. A farmer uses a mobile device to scan a leaf using a camera, the model picks up the image and is able to scan the image.

This is a search for a solution that is very efficient even in places with little or no internet.

Tensorflow, a model development platform was used to come up with a model. The model has been compressed for edge devices using tensorflow lite

Technologies Used

• Tensorflow
• Python 3
• Kotlin
• Android (AOSP)
• Intel Arc

Repository

https://github.com/oyamo/plant-disease-neural-net