Skip to content

Real-time computer vision classification to showcase the growth stage of a farming crop

Notifications You must be signed in to change notification settings

NoelGore/Growth_Stage_Classifier

Repository files navigation

Crop Growth Stage Classification


This project was completed for the ATB Datathon (Alberta's largest Datathon), with a team of 6. The goal of the competition was to solve problems impacting the lives of everyday Albertans, with potential for global impact. We created a computer vision classification model, using a deep neural network, to predict the growth stage for a given plant. This would help with herbicide application, insect treatment and fungicide decisions, while allowing large scale farms to accelerate their decision making process. The aim was to then integrate this with a crop phenology model to validate the model's predictions.

Full Demo: https://www.youtube.com/watch?v=G6xBIVzubFk

Presentation: Crop growth stage modeling and classification

Project Blog: https://www.linkedin.com/pulse/computer-vision-technology-project-recognized-finalist-kurrant/


Dataset

Download the Aberystwyth Leaf Evaluation Dataset

wget https://zenodo.org/record/168158/files/images_and_annotations.zip?download=1
unzip images_and_annotations.zip

Note: Dataset size is 61GB before extraction.

Preprocess the data and decide the growth stage split:

python preProcess.py

Training

We used tensorflow 1.12. To train run the following code:

python train.py

Trained Model

Trained Model


About

Real-time computer vision classification to showcase the growth stage of a farming crop

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages