My inspiration came from my Mom and her personal garden. She was complaining that she had invasive creepers growing all across her garden but did not know which ones were invasive or not, and so asked me to figure out how to get rid of them.
What it does
It allows you to take photos of the plant you wish to identify and determines if the plant is invasive, and giving you some nice information about it too!
How we built it
It opens as a Photo App, allowing you to take pictures of the plant you wish to determine as invasive or not. Then, it runs the photo you took into a Create ML Image Detector which classifies it as one of many different plants. From there, we take the name given and search our database, which we filled with plants by querying google to determine which plants are invasive as well as giving a short description of google's answer box, a link to website either determining how to get rid of the plant or take care of it (determined by whether the plant is invasive), and the title of that webpage. The app then displays the data from the database using a text format, easy for you to read!
Challenges we ran into
Building a camera for this project (even basic ones) was difficult and time consuming. I ran into a lot of issues with the video display for the camera, as well as the actual taking of the image (pretty much the entire thing).
I also ran into trouble at the end with time, as I was unable to complete the full text and link attachment that I wanted to, but I was able to display the name of the plant. I was able to retrieve the data from the database but had difficulty moving it out of scope.
Accomplishments that we're proud of
I have never done any Swift before! Or any of the technologies used in this app with the exception of Python. I am proud of being able to have learned all of this information on the fly and still come up with a product that I am proud of even despite not having it completed.
What we learned
- Swift (general as well as more advanced functions)
- Create ML
What's next for InvasionDetection
- Fix the scope issue
- add link attachment
- build a better UI
- expand the library