Seedling Sorter is a model that can classify images of 12 types of seedlings (small plants) with 81% accuracy.
My objective was to create a tool that could quickly and accurately classify seedlings (young plants). One possible application of this model is precision weed management, where weeds are identified and removed in real time using sensors, a tractor, weeding tools, and a computer to process images with the model. Such technology can reduce manual labor and herbicide use.
Users can upload their own images to the app, or drag and drop any of the provided example images. The app returns the top three predictions in less than 3 seconds.
The project uses the VS Plant Seedlings Dataset available on Kaggle.