Dynamic sensing and correlation loss detector for small object detection in remote sensing images
conda create -n dcdet python=3.7 pytorch==1.7.0 cudatoolkit=10.1 torchvision -c pytorch -y
conda activate dcdet
pip install openmim
mim install mmcv-full
mim install mmdet
git clone https://github.com/CHaunceyshen/DCDet.git
cd DCDet
pip install -r requirements/build.txt
pip install -v -e .Please refer to data preparation for dataset preparation.
- single GPU
- single node multiple GPU
- multiple node
You can use the following commands to infer a dataset.
# single-gpu
python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments]
# multi-gpu
./tools/dist_test.sh ${CONFIG_FILE} ${CHECKPOINT_FILE} ${GPU_NUM} [optional arguments]
# multi-node in slurm environment
python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [optional arguments] --launcher slurmExamples:
Inference RotatedRetinaNet on DOTA-1.0 dataset, which can generate compressed files for online submission. (Please change the data_root firstly.)
python ./tools/test.py \
configs/dcdet/dcdet_sods_corr_ss_r50_fpn_1x_dota_le90.py \
checkpoints/SOME_CHECKPOINT.pth --format-only \
--eval-options submission_dir=work_dirs/Task1_resultspython tools/train.py ${CONFIG_FILE} [optional arguments]./tools/dist_train.sh ${CONFIG_FILE} ${GPU_NUM} [optional arguments]@ARTICLE{10545316,,
title = {Dynamic Sensing and Correlation Loss Detector for Small Object Detection in Remote Sensing Images},
journal = {IEEE Transactions on Geoscience and Remote Sensing},
author = {Shen, Chongchong and Qian, Jiangbo and Wang, Chong and Yan, Diqun and Zhong, Caiming},
year = {2024},
volume = {62},
pages = {1-12},
}