Inspiration
Jack suggested the topic of gardening. Everyone saw the potential and starting investigating the concept. After many hours of throwing around ideas, the team found that many apps exist that do things similar to what was being suggested. There were a few key ideas that inspired further progress: 1) augmented/virtual reality, 2) the challenges faced by inexperienced gardeners, and 3) an interest in engaging the user in an unorthodox way.
What it does
Plant Projector is an app for inexperienced gardeners that predicts a plant's future based on the weather forecast. It uses image classification (photo taken from phone) to identify the plant and determine its ideal temperature range. Afterwards, based on the current weather and forecasted weather, the app gives a prediction on the quality of the plant's health and/or yield in the given days. The initial plant-specific information is provided by farmer's guides online, who have much more knowledge and wisdom in this matter. Obviously, there are many other factors that contribute to a plant's health. Due to time constraints, this prototype only accounts for weather. See the future plans for details on how the app can improve.
How we built it
Python Flask for webserver, tensorflow for image classification, weather api for forecast data
Challenges we ran into
Getting the project done on time. Having enough people
Accomplishments that we're proud of
Executing the idea as much as possible.
What we learned
Learn to ask for help earlier in the process.
What's next for Plant Projector
We plan to include a feature that allows one to see the plant's projection over time period. A time meter that the user can drag would indicate the plant's predicted state at a specific moment assuming things remain as is.
For the level of future accuracy to improve, predictions must be taken more frequently or recently. As mentioned before, the initial plant data is sourced from online farmer's guides. Overtime, data will be collected, aggregated, and stored in a large database for further processing. Many different scenarios will need to be explored and more variables mapped to gain a more accurate idea of what specific factors determine success of a plant more than others.
Log in or sign up for Devpost to join the conversation.