YOLOX-ROS Object Detection Package

YOLOX-ROS

YOLOX + ROS2 Foxy (cuda 10.2)

NVIDIA Graphics is required

Japanese Reference (Plan to post):Qiita

Requirements (Python)

  • ROS2 Foxy
  • CUDA 10.2
  • OpenCV 4.5.1
  • Python 3.8 (Ubuntu 20.04 Default)
  • Torch '1.9.0+cu102 (Install with pytorch)
  • cuDNN 7.6.5 (Install with pytorch)
  • YOLOX
  • TensorRT : is not supported
  • WebCamera : v4l2_camera

Requirements (C++)

  • C++ is not supported

Installation

Install the dependent packages based on all tutorials.

STEP 1 : CUDA Installation

STEP 2 : YOLOX Quick-start

YOLOX Quick-start (Python)

git clone https://github.com/Megvii-BaseDetection/YOLOX.git
cd YOLOX
pip3 install -U pip && pip3 install -r requirements.txt
pip3 install -v -e .  # or  python3 setup.py develop
pip3 install cython; pip3 install 'git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI'

STEP 3 : Install YOLOX-ROS

source /opt/ros/foxy/setup.bash
sudo apt install ros-foxy-v4l2-camera
git clone --recursive https://github.com/Ar-Ray-code/yolox_ros.git ~/ros2_ws/src/yolox_ros/
cd ~/ros2_ws
colcon build --symlink-install # weights files will be installed automatically.

Demo

Connect your web camera.

source ~/ros2_ws/install/setup.bash
# Example 1 : YOLOX-s demo
ros2 launch yolox_ros_py demo_yolox_s.launch.py
# Example 2 : YOLOX-l demo
ros2 launch yolox_ros_py demo_yolox_l.launch.py

Topic

Subscribe

  • image_raw (sensor_msgs/Image)

Publish

yolox/image_raw : Resized image (sensor_msgs/Image)

yololx/bounding_boxes : Output BoundingBoxes like darknet_ros_msgs (bboxes_ex_msgs/BoundingBoxes)

※ If you want to use darknet_ros_msgs , replace bboxes_ex_msgs with darknet_ros_msgs.

YOLOX + ROS(1, 2) object detection package

Parameters : default

  • image_size/width: 640
  • image_size/height: 480
  • yolo_type : 'yolox-s'
  • fuse : False
  • trt : False
  • rank : 0
  • ckpt_file : /home/ubuntu/ros2_ws/src/yolox_ros/weights/yolox_s.pth.tar
  • conf : 0.3
  • nmsthre : 0.65
  • img_size : 640

Reference

YOLOX + ROS(1, 2) object detection package

@article{yolox2021,
  title={YOLOX: Exceeding YOLO Series in 2021},
  author={Ge, Zheng and Liu, Songtao and Wang, Feng and Li, Zeming and Sun, Jian},
  journal={arXiv preprint arXiv:2107.08430},
  year={2021}
}

About writer

  • Ar-Ray : Japanese student.

Download Details:

Author: Ar-Ray-code

Source Code: https://github.com/Ar-Ray-code/YOLOX-ROS

 

YOLOX-ROS Object Detection Package
1 Likes1.80 GEEK