UAV Model and Dataset

The UAV (Project) is a baseline model for CVPR23 Anti_UAV workshop.We encourage participants to develop automated methods that can detect and track UAVs in thermal infrared videos with UAV model and dataset.

Download Dataset

Install modelscope-cv:

pip install "modelscope[cv]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

Run the download code:

from modelscope.msdatasets import MsDataset
from modelscope.utils.constant import DownloadMode

## Download train set
UAV_train = MsDataset.load('Anti_UAV', namespace='damo', split='train', download_mode=DownloadMode.FORCE_REDOWNLOAD)
print(next(iter(UAV_train)))

## Download validation set
UAV_validation = MsDataset.load('Anti_UAV', namespace='damo', split='validation', download_mode=DownloadMode.FORCE_REDOWNLOAD)
print(next(iter(UAV_validation)))

Download Baseline Model

Install modelscope-cv:

pip install "modelscope[cv]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

Run the download code:

from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks
pipeline(Tasks.face_detection, 'damo/cv_manual_uav-detection_uav')

Note: You can find model architecture, weiths, test code in /root/.cache/modelscope/hub/damo/cvmanualuav-detection_uav/.

Evaluation Code

Install CVPRAntiUAV

git clone https://github.com/ly19965/CVPR_Anti_UAV
pip install -r requirements/tests.txt -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html
pip install -r requirements/runtime.txt -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html
pip install -r requirements/cv.txt -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

Evaluation model accuracy

cd CVPR_Anti_UAV
python tests/pipelines/test_uav_detection.py

Source Description

The model and code come from the open source community([Project).Please comply with relevant permits。