Project Page

Introduction

Oriented object detection in aerial images poses a significant challenge due to their varying sizes and orientations. Current state-of-the-art detectors typically rely on either two-stage or one-stage approaches, often employing Anchor-based strategies, which can result in computationally expensive operations due to the redundant number of generated anchors during training. In contrast, Anchor-free mechanisms offer faster processing but suffer from a reduction in the number of training samples, potentially impacting detection accuracy. To address these limitations, we propose the Hybrid-Anchor Rotation Detector (HA-RDet), which combines the advantages of both anchor-based and anchor-free schemes for oriented object detection. By utilizing only one preset anchor for each location on the feature maps and refining these anchors with our Orientation-Aware Convolution technique, HA-RDet achieves competitive accuracies, including 75.41 mAP on DOTA-v1, 65.3 mAP on DIOR-R, and 90.2 mAP on HRSC2016, against current anchor-based state-of-the-art methods, while significantly reducing computational resources.

Installation

Data preparation and download

• DOTA-v1.0: download
• HRSC2016: download
• DIOR-R: download

HA-RDet
├── mmrotate
├── tools
├── configs
├── data
│   ├── split_ss_dota
│   │   ├── trainval
│   │   │    ├── annfiles
│   │   │    ├── images
│   │   ├── test
│   │   │    ├── annfiles
│   │   │    ├── images
│   ├── DIOR-R
│   │   ├── trainval
│   │   ├── test
│   ├── HRSC
│   │   ├── ImageSets
│   │   ├── FullDataSets

Our experiment relies on the MMRotate framework provided by Open MMLab. MMRotate depends on PyTorch, MMCV and MMDetection. Quick steps for installation follows as:

• Git clone

git clone https://github.com/PhucNDA/HA-RDet

• Environment setup

conda create -n [NAME] python=3.7 pytorch==1.7.0 cudatoolkit=10.1 torchvision -c pytorch -y
conda activate [NAME]
pip install openmim
mim install mmcv-full
mim install mmdet
cd 'Hybrid-Anchor-Rotation-Detector'
pip install -r requirements/build.txt
pip install -v -e .

Training and Inference

• Training command:

python tools/train.py ${CONFIG_FILE} [optional arguments]
# Example:
python tools/train.py configs/ha_rdet/hardet_baseline_r50_fpn_1x_dota_le90.py

• Inference command for online submission:

python ./tools/test.py  \
    configs/ha_rdet/hardet_baseline_r50_fpn_1x_dota_le90.py \
    checkpoints/SOME_CHECKPOINT.pth --format-only \
    --eval-options submission_dir=[SAVE_FOLDER]

• Visualize the results

python ./tools/test.py  \
    configs/ha_rdet/hardet_baseline_r50_fpn_1x_dota_le90.py \
    checkpoints/SOME_CHECKPOINT.pth
    --show-dir [SAVE_FOLDER]

Benchmark and Model Zoo

DOTA-v1.0 dataset

HRSC2016

DIOR-R

Visualization