Crop Health

Inspiration

Crop disease can have widespread impact - if the disease spreads throughout a field, it could have significant repercussions. Because of this, it is imperative to detect disease early. Knowing that machine learning could help us achieve this task, we decided to pursue this project.

What it does

We created a machine learning program to scan images of certain crops and determine if they had a specific condition. Specifically, we scanned images of peaches, and determined if they were healthy of if they had a bacterial spot.

How we built it

We utilized Tensorflow and Keras to construct the machine learning model - specifically, we used a convolutional neural network to analyze the photos. The dataset was found on Kaggle, which contained a large set of photos, each showing an image with a crop and a certain condition.

Challenges we ran into

Since there were many project ideas that we were simultaneously considering, it was difficult to narrow our project down to just one topic! In addition, the size of the data set made it difficult to run, since it would frequently run out of RAM (the dataset contained thousands of images). To mitigate this, we reduced the size of the training and testing data by selecting a subset of each.

Accomplishments that we're proud of

We were proud of learning multiple new technologies and how they can apply to not only our current project, but our future projects as well.

What we learned

We gained a deeper understanding of Tensorflow/Keras for machine learning. In addition, we gained a better sense of the format of a Hackathon and the importance of idea formation.