search

Alpaca-zip / ultralytics_ros

ROS/ROS 2 package for Ultralytics YOLOv8 real-time object detection and segmentation. https://github.com/ultralytics/ultralytics
GNU Affero General Public License v3.0
184 stars 35 forks source link

How to detect camera information in real time #45

Closed wyb17js closed 7 months ago

wyb17js commented 7 months ago

作者你好，我使用gazebo搭建了一个环境，并且在我的车体模型上安装了深度相机插件，我想实现识别结果以视频的形式展示出来，随着摄像头检测到画面的变化而不断变化，但是当我运行了readme里面指定的launch文件后，我只得到了一张识别结果图片，并且rviz显示无响应，当我关闭后控制台报错如图所示 image 我想知道我应该怎么配置或者编写launch文件才可以实现我想要的效果？还有我该怎么训练自己的模型？ image

Alpaca-zip commented 7 months ago

Thank you @wyb17js , for asking your question.

我想知道我应该怎么配置或者编写launch文件才可以实现我想要的效果？

As it is difficult to solve this problem with the information you have provided, could you please share the scripts( e.g. gazebo configuration file) you are running?

还有我该怎么训练自己的模型？

I don't provide the functionality to train models directly in ROS/ROS 2. Please check the Ultralytics YOLOv8 Docs link for more information.

Thank you.

Alpaca-zip commented 7 months ago

Dear @wyb17js,

I misunderstood your query earlier. Currently, the humble-devel branch does not support 3D object detection using camera_info and point clouds. Please submit a feature request through this link, and wait for it to be implemented. Alternatively, I recommend using ROS1.

Thank you.

Alpaca-zip commented 7 months ago

Announcement 📢 3D object detection using camera_info is now available in the humble-devel branch. Please check it out!!

Thank you.