Table of Contents
Our method has been tested on python 3.10, CUDA>=11.6. The simplest way to install all dependences is to use anaconda and pip in the following steps:
conda create -n MMUAV python=3.10
conda activate MMUAV
conda install -c "nvidia/label/cuda-11.6.0" cuda-toolkit
conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.6 -c pytorch -c conda-forge
pip install -r requirements.txtInstall Yolov9 Dependency
pip install yolov9/requirements.txt
And download pretrained weight of Yolov9 under ./yolov9.
Then Prepare the dataset utilizing the zero-shot UAV detection by Yolov9
python3 yolov9/detect.py --source $SOURCE --img 640 --device 0 --weights './yolov9-e.pt' --name $OUTPUT --num_kf 5 --kf_int 50 --save-crop
Replace $SOURCE and $OUTPUT to the source and desired output directories of the target trajectory. num_kf is the maximum number of keyframe selected from the trajectory. kf_int is the interval threshold between each keyframe to prevent selecting similar keyframes. save-crop will save the cropped image for detected drones, necessary for the training of efficientnet.