The name Agri-Wiess is inspired by the German term ’ich weiß’ / ‘Ich Weiss’ which in English translates into the term ‘ I know’. Knowledge is power, and when farmers or anyone have the right information in hand, making decisions is so much easier. The ability to make informed decisions allows one to save time and money and reduces the stress of figuring out how to deal with situations where the outcomes may be unknown.
This project is inspired by the dream of a sustainable environment. The consumption of water in Agriculture can be analyzed and water can be used more efficiently.
What it does
This tool predicts the soil moisture within a certain soil profile. The soil moisture is the volumetric water content that a particular portion of the soil can hold. This is important to farmers, golf course managers, and anyone that works with soil. I.e being able to determine the VWC of a piece of land could determine how effective irrigation has been / or a water profile.
How we built it
We used Azure ML Studio to create and deploy a model as a service. The data used to train the model was collected from industrial-grade soil sensors. The Arduino sends data to an MQTT broker, the data is then sorted out, used to call a Weather-API and all the data is then saved into the Database. In order to do this, we had to make work with AWS IoT Core for the sensors to interact with the MQTT broker, AWS Lambda to trigger functions such as API calls and data entry.
Challenges we ran into
The biggest challenge was being able to successfully send sensor data to the Azure IoT hub.
Accomplishments that we're proud of
What we learned
Being able to deploy an Azure ML model as a service and test the accuracy of the model. Using Microsoft Azure ML studio is highly satisfying and has prompted me to work on completing the 30-day code challenge. Being able to be a part of this hackathon and submit our project.
What's next for AgriWeiss
Test and improve the model by feeding more data Find actual users to test the product