This repo provides results presented in this paper on detecting Northern Corn Leaf Blight in images.

1. Darknet
   We train a neural network using the YoloV4 framework, which relies on darknet. Instruction to install darknet's dependencies can be found here

2. Northern Corn Leaf Blight Dataset
   We used a dataset documenting Northern Corn Leaf Blight that was procured by Mohatney et al, which can be found here in order to train our neural network.

   Note: We specifically trained on a subset of the handheld data that contained only 1 instance of Northern Corn Leaf Blight in the image, to lessen the requirements of image augmentation.

Instruction for training a YoloV4 model from scratch can be view on colab (aqcuired from YoloV4 documentation wiki).

All necessary yolov4 dependencies are provided in the yolov4 directory of this repo.

Download and configure darknet:

git clone https://github.com/AlexeyAB/darknet
cd darknet
sed -i 's/OPENCV=0/OPENCV=1/' Makefile
sed -i 's/GPU=0/GPU=1/' Makefile
sed -i 's/CUDNN=0/CUDNN=1/' Makefile
sed -i 's/CUDNN_HALF=0/CUDNN_HALF=1/' Makefile
make

Run model on test data:

./darknet/darknet detector test yolov4/obj.data yolov4/yolov4-nclb.cfg yolov4/yolov4-nclb_best.weights demo/augmentations/00000050.jpg -thresh 0.3