What it does
AGROTIS is mainly intended for the farmers and the agricultural authorities. It comes with a web application, mobile application and an IoT device. The device consists of sensors that lively records the farming parameters like humidity, temperature, soil moisture, sunlight, dew, wind speed, and stores it in a server. This information is used by the web application and the mobile application. The web application is only for the authorities. The data from the server helps them to track the status of the device, condition of the soil, climatic change of that particular place where the device is embedded. This data also helps the authorities to predict any future disease outbreak and notify the users of that particular region. It also has a feature to predict the disease acquired by the plant by uploading a photo of the affected leaf. The mobile application is mainly meant for farmers. The data from the server is analysed and gives suggestions to the farmers for better yield. The application has a platform for farmers to click a picture of the disease affected plant leaf and the application would predict the disease. The mobile application also has a voice assistance in Malayalam to guide the farmers. The web application and mobile application has a feature in which the weather forecast for the next 7 days is shown to the respective users. There is an expert forum option in both these applications where the queries from the farmers can be posted in the mobile application and the authorities can reply to it from the web application portal.
How we built it
IoT DEVICE The device is embedded in the soil, uses sensors, and collects the data like humidity, temperature, moisture content, light intensity(sunlight) of that place, and also the moisture of the soil. The device records the continuous data and stores it on the server. Only the current measurements of the above-mentioned parameters are shown to the farmers. The recorded data is used for analysis by the authorities. Every 1 hr the data from all the devices is sent to the server for analyzing the chance of disease outbreak.Sensors used for data collection in the stage of prototype development are
- LDR Sensor
The device will take the Nutrient parameter of that particular location using ISRIC ‘s API.
WEB APPLICATION The web application is designed for agricultural department authorities in which they can login and access all its features. Later it can be accessed by users by mentioning rules. The features of the web application include:
- Dashboard: Gives the overview of online and offline devices, outbreaks found, pending and published soil test results & the total number of queries posted.
- Real-Time weather forecasting: Region-wise weather conditions can be monitored and alerts about the adverse weather conditions.
- Data History: Gives detailed statements of the device readings and weather forecast with graphical representation.
- Device List: This page displays the details of all the Agrotis devices along with the readings from the sensors which include temperature, humidity, moisture, pH value, light intensity. It also shows the current status of the device, that is, active or inactive. The device can also be searched and viewed on a map.
- Soil Map: This feature allows us to view the soil properties of a particular location marked on the map that is displayed on the web page. The best crop suitable for that particular area is also predicted and shown by clicking on the “Prediction” button.
- Disease Prediction: By uploading the image of the disease infected leaves, the possible diseases are predicted and listed with the cure.
- Disease Outbreak: Using weather parameters like temperature, humidity, precipitation, light, and wind, the disease that is likely to occur in an area can be predicted and an alert notification regarding this can be given to the farmers in that particular area by simply drawing or Geofencing the area on the map using the provided tools.
- Soil Test Result: The sample of the soil to be tested can be given to the authorities and the test result will be displayed on this page.
- Expert Forum: Queries posted by the farmers using the mobile application can be answered by agriculture officers through this web portal.
- Weather Forecast: It also has a feature to predict the weather for the next 7 days and display it to the users.
- Publish News: It publishes news or notifications regarding new updates and also allows us to view the already published news/notifications.
MOBILE APPLICATION The mobile application is designed for farmers in which they can register their phone numbers and avail of facilities like
- Weather forecasting: Detailed weather report of the device installation area.
- Crop selection: Farmers can select multiple crops and can view crop details/requirements easily from the home page.
- Crop suggestions: The data from the device and location-based soil APIs & weather APIs are compared with the parameters of the crop selected by the farmer and gives them details about the variations and suggestions to increase the yield.
- Disease prediction: Offline ML model is developed in the app which predicts the disease affected in leaves and stems of the crop. If the model fails to identify the disease, it can be forwarded to the web, so that the web model identifies the disease if possible.
- Expert forum: If a farmer has any queries he can post them on this platform where the authorities will reply to those queries through the web portal.
- Weather forecast: A forecast for the next 7 days is shown by this feature
- Voice assistance: A voice assistance in Malayalam is incorporated in the application. This will help farmers to understand the basic functionality of the application and make it easy to use.
- Crop finder: With the data from the server the application would suggest the crop that is best suitable for that soil.
Disease outbreak alert Geofence areas on the map using various shape tools that are available and the users who have devices registered inside that particular geofenced area will get instant notifications regarding any disease outbreaks provided. The possibility of disease outbreaks can be found from graphs of various parameters displayed in the Dashboard.
The server will alert the authorities if there is any chance of a disease outbreak based on the device reading collected in the past 24 hours. All the parameters like temperature, humidity, moisture, relative humidity, light, will be analyzed by the server every hour. A graphical visualization is done on the web to see the statistical analysis for the disease outbreaks. Based on the analysis the system will suggest the chances of disease outbreak in different regions. If chances of disease outbreak occurs, officers can geofence that particular location using geofencing features above mentioned
The device consists of some sensors like moisture sensor, humidity sensor, Temperature sensor, Light intensity sensors. After embedding the device into the soil the data from these sensors are pushed into the database (firebase). The sensor collects data on an hourly basis and only the current value is displayed to the user the rest is stored in the firebase. This data is then analyzed in both the web and mobile applications. The farmer completes the authentication process in the app by providing the phone number used while buying the device (device Id). Firebase is used for authentication purposes and the real-time monitored data retrieved from the device is pushed into the Firebase. After the authentication, the application leads to the dashboard where some of the crops are listed out based on the data from APIs and ‘Kerala University Agriculture Cell’ for that particular place. From the crops given on the dashboard the farmer can select a crop and is directed to a page where the suggestions are given by comparing the data from the database, weather APIs, and a government portal - ‘Kerala University Agriculture Cell’. This data from the database gives details about the optimum temperature, moisture, humidity, etc for a particular crop. After the comparison, it displays the suggestions
One of the features of the application is disease prediction by uploading pictures of the infected plant. It is predicted with the help of an ML model. An offline model is deployed within the mobile application itself to test and predict the disease and in case it fails, a prediction is done by requesting the help of an online disease predictor deployed within the web application. In this manner, farmers can also get role-based authentication to the website which was originally meant only for the agriculture department officers. When the disease is predicted a cure for it is also suggested. The ML also procures the intensity of the disease. The image here can either be uploaded from the gallery or by clicking a picture of the leaf. The picture undergoes automatic cropping (260X260).
To make this app more user-friendly the application is incorporated with multi-language support also with voice assistance (Malayalam voice assistance).
Challenges we ran into
We don't have any data set regarding the nutrient level of various crops under various parameters. So we synthetically prepared the data.
Accomplishments that we're proud of
Agrotis have collaborated with Kerala State Agricultural Department. Authorities approved financial support for developing this product. They provide us the industrial experts and agricultural experts to work on this product. We were selected by certain investors also.
What's next for Agrotis
As we moving to develop our first MVP with the help of the authorities and government in this year itself.