Urban Organic Farming

Hardware components
Hardware integration
KNN Model
Prediction model
Accuracy level

Inspiration

A lot of people today are interested in farming and its practices. Most people do not have enough land to do large-scale farming. As a result, many are shifting toward organic greenhouse farming. However, they lack specific expertise of this discipline and lack the time to maintain it. So our software assists them by providing facts on what may be put on a specific soil and how to properly care for the crop that has been planted by evaluating the soil using sensors and displaying all of the details on an app or a website the data is sent wireless using ESP32 sensor.

What it does

It helps people who are interested in urban farming by providing details on what crop can be planted by analyzing the soil of its nutrition content, past rainfall patterns of the area, past and future weather data of the location we are about use for farming, the type and amount of area the user is going to use for planting we combine all these details and use our machine learning model to predict which type of plants can be grown and can be maintained according to the data from NPK, DHT11, moisture and MQ-135 sensors and our API for weather prediction and past weather data monitor soil moisture, water, and plant's condition 24/7 using NPK, DHT11, moisture and MQ-135 sensors and give information to the user on how the plants are and also help them by telling them about when to water plant or any other activities to be done base on the 24/7 monitoring data we use an website or an app to display all the monitoring details and any information to the user is also conveyed

How we built it

We constructed the prediction model using the knn algorithm (machine learning model) since it had a higher accuracy level than the other models. To monitor the soil and Arduino, we utilized NPK, DHT11, moisture, and MQ-135 sensors. To display all of the information to the users, we used a website or an app. The app was built in Android Studio, and the website was constructed in React and Javascript.

Challenges we ran into

The most difficult aspect was gathering all of the data for training the machine learning model; it was tough to gather diverse types of data in huge amounts for varied scenarios.

Accomplishments that we're proud of

We are happy that we are able to build a machine learning model that has an accuracy of 97%

What we learned

We realized that utilizing sensors and machine learning models is a challenging endeavor, but if we can complete it correctly, we will be able to achieve positive outcomes.