Agriculture is the primary source of livelihood for about 58 per cent of India’s population. Gross Value Added by agriculture, forestry and fishing is estimated at Rs 18.55 lakh crore (US$ 265.51 billion) in FY19(PE).
Apart from being an industry, agriculture is the main source of income for about 70 million farmers in India. Undoubtedly, Agriculture is crucial to Indian economy. What's more important, is that Indian farmers are feeding 1.4 billion people, which is one-fifth of the total world population.
However, Indian agriculture is threatened by pests that can ruin their crops and their livelihoods.
According to a study by the Associated Chambers of Commerce and Industry of India, annual crop losses due to pests and diseases amount to Rs.50,000 crore ($500 billion), which is significant in a country where at least 200 million Indians go to bed hungry every night.
With so many tonnes of crops being destroyed every year due to pests and diseases, it is important to find a solution and implement it to avoid any more waste of crops due to diseases and pest infections. Which is why, we have come up with our solution which aims to reduce crop destruction by pests. In this way, we can feed atleast a portiion of the 200 millions who go to bed every night!
If 30% of crops produced in India are being affected by dieseases, imagine the global scenario. We believe, our Leaf Disease Identification project can tackle various problems such us Global Hunger, Responsible Production, and more.
We have developed a mobile application using Flutter that uses Machine Learning to identify more than 35 common diseases in crops.
The Machine Learning model is trained with over 6.5 Lakh images and has an accuracy of 97.53%
The mobile application is very easy to use. The farmer will first capture an image of an infected leaf or can also select an image from their mobile storage. Next, the leaf image will be analyzed by our Machine learning model, which will predict the leaf disease instantly. The application will predict the disease name, and will also provide the causes and possible solutions to counter the disease.
The application is divided into different modules:
The home screen contains simple instructions on how to use the application to identify leaf diseases. It also stores and displays past history of identified leaf diseases.
The Navigation screen contains options for the farmer to translate the application into a language of their own choice. The farmer can capture an image of a plant or select an image from their storage to know the condition of their plant.
The Language translation screen will allow the farmer to use the application in whatever language they are comfortable with.
The Suggestions screen is the main screen which returns the name of the disease affecting the crop. It will also return possible causes and solution for the particular disease.
That is why we have developed a language translation module that will translate our application into 15 widely spoken languages in India. We will soon expand it to 22 languages that are commonly spoken in India. This will make our application more accessible for the farmers.
Updated.Leaf.disease.identification.by.Team.Introspectors.mp4
Here is the TOC this project.
We have built a mobile application that is integrated with a robust machine learning model that is trained with a very large dataset.
The mobile application allows a farmer to capture an image of an infected leaf and select an image from device storage. The application will then predict the disease affecting the crop and will also provide the solution to that disease.
In this way, a farmer can detect a plant disease very easily.
To predict the solution, we have integrated a well trained Machine Learning model into our application which stands with an accuracy of 97.53%.
Talking about the technologies that we have used to build the product, we have used multiple technologies to develop our project.
Firstly, we developed our mobile application using flutter. We have used flutter to develop cross-platform mobile applications.
Next, we have used Tensorflow to develop a robust Machine learning model with an accuracy of 97.53% and trained it with over 6,55,000 images and used Keras to create a layered neural network to distinguish over 30 types of common plant diseases.
If we help to identify pests and diseases in their early stages, a farmer can take necessary actions to save their crop. This will contribute in feeding atleast a portion of the 200 million Indians who go to bed hungry every night.
We are looking to use Technology to tackle challenges in Agriculture. And we believe, our Leaf disease identification mobile application is going to play a crucial role in revolutionizing farming in India.
In order to get started with the application, fork the repository and create a local workspace.
Install Flutter SDK, Android Studio/ VS Code
Set the environment variables if necessary
- Clone the repo
git clone https://github.com/Dantusaikamal/Leaf-disease-identification.git- Install Flutter packages
flutter pub get- Run the application
flutter runWe have created a mobile application in which we need to capture an image of an infected leaf. And it will predict the disease affecting the crop and will also provide the solution to that disease.
In this way, a farmer can detect a plant disease very easily.
In order to predict the solution, we have integrated a well trained Machine Learning model into our application which stands with an accuracy of 97.53%.
Talking about the technologies that we have used to build the product, we have used multiple technologies to develop our project.
Firstly, we developed our mobile application using flutter. We have used flutter because we can develop cross platform mobile applications.
Next, we have used Tensorflow to develop a robust Machine learning model with an accuracy of 97.53% and trained it with over 6,55,000 images and used keras to create a layered neural network to distinguish over 30 types of common plant diseases.
Please check Contribute.md to know how you can contribute to this project.
If you want to sponsor the project, please donate your money to any relief funds for COVID-19 like WHO Relief fund or the organizations in your own country like PM cares fund
If you find an issue in the project, open a new issue or a raise a pull request in this repository.
You can also mail to [email protected] incase of any issues.
Check out the LICENSE file in the repo.