INSTAllation

Requirements

• Linux (Recommend), Windows (not Recommend)
• Python 3.7+
• PyTorch ≥ 1.7
• CUDA 9.0 or higher

I have tested the following versions of OS and softwares：

• OS：Ubuntu 16.04/18.04/20.04
• CUDA: 10.0/10.1/10.2/11.3

Install

CUDA Driver Version ≥ CUDA Toolkit Version(runtime version) = torch.version.cuda

a. Create a conda virtual environment and activate it, e.g.,

conda create -n Py39_Torch1.10_cu11.3 python=3.9 -y 
source activate Py39_Torch1.10_cu11.3

b. Make sure your CUDA runtime api version ≤ CUDA driver version. (for example 11.3 ≤ 11.4)

nvcc -V
nvidia-smi

c. Install PyTorch and torchvision following the official instructions, Make sure cudatoolkit version same as CUDA runtime api version, e.g.,

pip3 install torch==1.10.1+cu113 torchvision==0.11.2+cu113 torchaudio==0.10.1+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
nvcc -V
python
>>> import torch
>>> torch.version.cuda
>>> exit()

d. Clone the swin-roleaf repository.

git clone 
cd swin-roleaf-model

Install swin-roleaf.

pip install -r requirements.txt
cd utils/nms_rotated
python setup.py develop  #or "pip install -v -e ."

Usage

• 1.Put the images into the dataset-corn folder
• 2.go to 'swin-Roleaf-model' directory and run 'python detect.py --imgsz sizeh sizew', then you can get the predicted images and corresponding predictive labels/angles in detect_results folder
• Note: if your planting axis is horizontal, please run 'python detect-horizontal.py --imgsz sizeh sizew' instead of 'python detect.py --imgsz sizeh sizew'. crop_size.py is for croping the original images to the size you want(will be square size), run 'python crop_size.py folder_path_of_image cropize' cropsize can be 640/1280...
• 3.After that, you can use user interface to show the label and correct labels, remember to open detect_results/results folder in user interface, the txt/xml files in this folder are label locations and angles.

Output format

• xml and txt: txt file each line means: difficult(set to 0) x1 y1 x2 y2 x3 y3 x4 y4