Green Companion

black and white mask CV
light source detection CV

Inspiration

Our project is fueled by a profound desire to share our passion for nature with the global community. We recognize that the art of cultivating and nurturing plants is often undervalued and oversimplified. Many believe that adequate sunlight and water suffice, yet this common misconception leads to either overwatering or undernourishing, resulting in the untimely demise of countless plants. We aim to correct these misunderstandings by introducing innovative, user-friendly technology that ensures plants not only survive but thrive.

What it does

Project GreenCompanion not only simplifies the complexities of plant care but also educates and involves the user, creating a deeper sense of responsibility and connection to the environment. By integrating technology with traditional gardening practices, we offer a solution that enhances living spaces, improves mental well-being, and brings the beauty and benefits of nature into modern homes.

How we built it

We added the moisture sensor with Arduino to detect the moisture levels. We looked at different datasets and statistics to verify the proper moisture value of healthy soil. Then we added a servo to give the plant a dancing effect as the speaker plays rickroll music when the soil is too wet and the doom theme song when the soil is too dry. (We added the illusion of the plant singing to you)

We also used a Raspberry Pie to run a Computer Vision program with a Logitech camera that detects the nearest window or light source and then activates the motor which rotates the plant on its axis facing the direction of the light. (Giving the illusion of a curious pet who always follows the light)

We used Blender and Fusion 360 to design the 3D chassis for the robot, hosting all the electronics inside it neatly while designing it to look like a normal flower pot from the outside.

Challenges we ran into

While using the moisture sensor we found a major flaw with all the other products that use this sensor. When the sensor is being used it starts electrolysis and gets rusted quickly (within 3-5 days)

While 3D designing we had to redesign 3-4 times as we had a lot of moving parts which could in turn loosen the electronic wires.

After making the final design the printing time was coming to 1 day 4 hours which would be unfair to everyone else as we would be using the printers for the majority of the hackathon.

After successfully creating the AI model for light detection and axis rotation, when the time came to implement it, our Raspberry Pi started giving issues and would not connect and take in code. Unfortunately, we could not get it to work even with the faculty leading to us not implementing the AI model which was ready for deployment.

Accomplishments that we're proud of

To avoid rusting, we found out that indoor plants need water for an average of 6-12 hr and we added a logical statement that checks the moisture level in between the above range and checks the moisture level after 10 minutes after watering the plant.

The 3D model was successfully scaled down by 83% reducing the time from 1 day 4 hours to 4 hours. We used different tricks to keep the model working while sticking to the time limit and showing common courtesy.

What we learned

We learned about the limitations of a well-established sensor and how to counter them, we also learned how to pivot in the last moments as things don't always go according to plan as we had to break our final model into smaller pieces while still designing it to be capable to give a demo.

What's next for Green Companion

We will be printing our full-size model and running the AI model with it, testing the project again with its full potential is what we aim for in the future.