Irrigation mechanism built using Arduino hardware
VB.net inferface displaying plant characteristics and live feed of webcam
VB.net interface displaying stored information for plant from the mySQL database
I was inspired by my numerous trips to India, where I saw massive fields sitting dry due to water shortages. I decided to carry out this project to help solve this global issue.
What it does
Although on a small scale, this project aimed to use color as the primary input to track plant health. Through a series of artificial intelligence and computer vision algorithms, the health of the plant was quantified based on the user's preferences. Based on prior data, the program predicts the ideal water volume to give each plant and the robotic arm irrigates each plant at 10x daily.
How I built it
I used Arduino and Makeblock hardware to build the irrigation mechanism. The mechanism uses a stepper motor to move on top of each plant and a peristaltic pump to give the water. A generic webcam was used to capture the images. I used C++ to code to the hardware. I used visual basic to carry out the AI algorithms, analyze the images, communicate with the database, and provide an interface to the user. A mySQL database was written to store the plant information.
Challenges I ran into
I originally used a color sensor to track the color of the plants, but this was unreliable because it found the color of a very small portion of the plant. By using a camera, I was able to analyze the entire plant and average out the hue of the entire plant on not 1 pixel.
Accomplishments that I'm proud of
I am proud that my project came out better than I expected and is a novel and potentially world-changing idea. I am patent-pending for my use of color input in a micro-irrigation system using artificial intelligence algorithms to make predictive suggestions.
What I learned
I learned how to create a basic database using mySQL. I then learned how to integrate the database with the VB.net code so the interface can access and display the information specific to the selected plant from the database. I also learned how to analyze pictures using the colorsummarizer API, which allowed me to quantify the health of the plants.
What's next for Auto-Irrigation using Color Input
For next year, I plan on scaling the project out using a drone with mulri-spectral imaging technology. The drone will be connected in an IoT system connected to self-sufficient drip irrigation pieces that can implemented with existing infrastructure. I am also in the process of developing a small piece that connects to existing irrigation pipes to give precision capabilities. The mechanism will be designed to be as cheap as possible, using a small turbine to generate power from water current flow, a servo motor to adjust watering volume, and dynamic RFID tags to communicate theMy goal for this is to implement the proven model in an affordable system.