https://github.com/alexandrefch/ros-yolov7

ROS package that implement a light yolov7 version in ros node style that allow user to run inference easely.
https://github.com/alexandrefch/ros-yolov7

noetic object-detection opencv python python3 ros ros-noetic ros-package torch yolov7

Last synced: about 2 months ago
JSON representation

ROS package that implement a light yolov7 version in ros node style that allow user to run inference easely.

• Host: GitHub
• URL: https://github.com/alexandrefch/ros-yolov7
• Owner: alexandrefch
• License: apache-2.0
• Created: 2023-02-17T17:39:31.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2023-02-23T16:50:13.000Z (almost 2 years ago)
• Last Synced: 2024-01-25T11:56:26.640Z (11 months ago)
• Topics: noetic, object-detection, opencv, python, python3, ros, ros-noetic, ros-package, torch, yolov7
• Language: Python
• Homepage:
• Size: 10 MB
• Stars: 5
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        


# ROS package for official YOLOv7

![Ros](https://img.shields.io/badge/Ros-Noetic-informational?style=for-the-badge&logo=ROS)

![Python](https://img.shields.io/badge/Python-3.8-informational?style=for-the-badge&logo=Python&logoColor=FFFFFF)



> **Note** 


> This ROS noetic package is an implementation of the [official yolov7](https://github.com/WongKinYiu/yolov7), most credit goes to them.



  


  **Fig.1** RViz view of detection projection from a Gazebo Simulation



## :rocket: Installation from source

First you need to download all required source, like the [ros-image-transport-py](https://github.com/alexandrefch/ros-image-transport-py) package, that manage automaticly image stream.

Moreover you might need some python package like [OpenCV](https://github.com/opencv/opencv), you can install them by simply execute `pip install opencv-python` or better compile from source for GPU.

### :gear: Clone and Compile

Clone thoose two packages inside the `src` directory of your catkin workspace.

```shell

git clone https://github.com/alexandrefch/ros-yolov7.git

git clone https://github.com/alexandrefch/ros-image-transport-py.git

```

Now you can compile them using your catkin tool (`catkin_make`, `catkin_build` ...).

## :clipboard: Configuration

### Weight

You can download some of the the YOLOv7 weights from the [official repository](https://github.com/WongKinYiu/yolov7) or use your own.

### Classes

Classes should be stored inside a simple text file where each single line represent a classe, like in the exemple below. 

```

car

bus

person

bike

motorcycle

traffic light

truck

person

```

### Launch file

```xml

    

        

        

        

        

        

        

        

        

        

        

        

        

        

        

        

        

        

        

        

        

    

```

## :black_nib: Citation

If you use this package in a scientific way or public way, don't forget to mention the official Yolov7 paper, and star their work on their [github repo](https://github.com/WongKinYiu/yolov7).

```

@article{wang2022yolov7,

  title={{YOLOv7}: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors},

  author={Wang, Chien-Yao and Bochkovskiy, Alexey and Liao, Hong-Yuan Mark},

  journal={arXiv preprint arXiv:2207.02696},

  year={2022}

}

```