farm.ai

Intuitive Crop Health Monitoring System

Comment

Introduction

Globally, over 45% of crop loss is due to issues caused by weather and disease. This crop loss comes with a significant cost of over $220 billion a year. As climate change further impacts global production, the risk of crop loss continues to rise. However, water-smart strategies, plant disease detection, and a focus on soil health can help mitigate these losses. This is why we created farm.ai.

Farm.ai is an intuitive early detection system that shows when certain parts of a crop might be dying and subsequently notifies the farmer before the spread of disease and weathering. Farm.ai has various sensors to detect internal and external factors such as temperature, humidity, wind speeds, and soil moisture content.

Inspiration

Farm.ai was created in response to the alarming number of crop losses and costs that occur every year. These issues will impact our future and the future of our families and we believe it is important to have useful preventative measures in place as soon as possible. While these issues are a consistent problem, there is much we can do to help offset the damage. For example, Farm.ai can help maximize the productivity of irrigation systems. In real-time, the farmer will see an unfavorable output related to soil moisture and humidity which informs them that irrigation times should be adjusted. These adjustments can make a lasting impact on the health of our ecosystem.

Feature Breakdown

Two hardware devices that collect soil moisture level of plants, external temperature and humidity, wind speed and direction data. Crop health visualization portal that allows farmers to view the data collected by the sensors in a smart and intuitive way. From satellite or drone photographs can determine the health levels of each plant in the photo and return a visual representation their health levels.

How we built it

The first device this project was built with is an Arduino Uno. Connected to the Arduino Uno is a capacity soil moisture sensor and a DHT11 temperature and humidity sensor. The moisture sensor is used to detect when the plant's soil contains too much water or needs to be watered. This could be connected to another system that automatically activates an irrigation system once a certain moisture content has been passed. The temperature and humidity sensor is used to gather data on the outside environment of the plants to determine whether their current environment is suitable for their growth. Our second device is a wind speed and direction detector. Our device was created by utilizing an Arduino Mega attached to a standard computer fan. Connected to the computer fan are bamboo skewers and condiment cups. The purpose of these supplies is so that wind is caught into the condiment cups. When it catches the wind, the fan spins. When wind causes the fan to spin the speed can be measured by taking the amplitude of the voltage outputted by the fan and convert it to miles per hour.

We used Django for the front-end/back-end and used Plotly's Dash to make the graphs and the appearance.

Challenges we ran into

Kari: Trying to convert my traditional art skills into vector art was really difficult for me! I did not realize just how different the programs were. It took me a while to get a design that felt cohesive to my original ideas. I also ran into a bit of trouble trying to create the gif. Albeit simple, I had to make sure each frame was properly captured within Inkscape.

Cole: This was my first time using Django. It was definitely different than React, which is what I normally use. One of the most difficult portions of creating Farm.ai was the AI portion.

Accomplishments that we're proud of

Kari: Learning to create a simple design using vector images in Inkscape was awesome! The end product was much different than the beginning, but creating a simple and cohesive logo for the first time was fun! I felt super excited to learn a bit about Python and Github!

Cole: I am proud that we were able to get the OpenCV detection of healthy vs unhealthy plants working. I was pretty worried that was going to be a bust, but it works!

What we learned

Kari: I learned how to do some work in Inkscape and how to create vector images. I also learned how to create a gif using Inkscape. There was a lot of information about crop health and the consequences of global warming that I was able to learn while researching for this project. When it came time to create the video, I learned how to record audio and take out background noise. I was able to learn a bit of Python and how Github works!

Cole: This was my first time doing any project with OpenCV so it was interesting to try and learn about that over the weekend. I also learned how to tutor someone properly about development workflow.

What's next for farm.ai

Farm.ai aspires to become more functional and available in order to help combat the challenges associated with climate change. On a more technical side, we would like to get the NDVI image to be within the web application so users can upload their images and have them displayed right on the webpage.

Share this project:

Updates