A flask web app that automatically counts wheat heads on wheat photos.
This demo app will be live until 30 June 2020.
Demo App: http://wheatcounter.test.woza.work/
My goal for this project was to build and deploy a flask web app that can automatically count wheat heads on wheat photos.
This app could help researchers get quick estimates of wheat density.
The process used to build and train the models is described in this Kaggle notebook:
https://www.kaggle.com/vbookshelf/automated-wheat-counter-pytorch-flask-app
The two models that power this app were trained using data made available during the Kaggle Global Wheat Detection competition. The data was released under an open source MIT license.
https://www.kaggle.com/c/global-wheat-detection
Validation percentage error: 20.5 %
Validation MAE: 5.03
I suggest that you deploy on a Linux server running Ubuntu 16.04. Start with a server that has 4GB of RAM and two CPUs. Once you get the app running you can then test it on smaller and cheaper servers.
The frontend and backend code is available in this repo. The models were too large to be uploaded. Please download them from the Kaggle notebook. There are two models: seg_model.pt and reg_model.pt. Please put both models into the folder called 'flask' before uploading the folder called 'wheat-backend' to your server.
The code is set up to be run as a Docker container. It's based on this video tutorial:
Julian Nash docker and flask video tutorial
https://www.youtube.com/watch?v=dVEjSmKFUVI
The .dockerignore file may not be visible. Please create this file if you don't see it. In this repo I've included a txt file that explains the steps for installing Docker and Docker Compose on a Linux server. There is folder called 'static' containing a predict.html file. This folder is not essential and can be deleted.