双目相关

最新推荐文章于 2024-06-26 10:54:04 发布

原创最新推荐文章于 2024-06-26 10:54:04 发布 · 425 阅读

0 ·

CC 4.0 BY-SA版权

Ros 专栏收录该内容

38 篇文章

订阅专栏

1，安装cheese确认摄像头可用， cheese黑屏：在VM设置（切换USB2.0/3.0 之后重新disconnect/connect） camara。

2，分辨率设置：

<node pkg="usb_cam" type="usb_cam_node" name="camera">
        <param name="video_device" value="/dev/video0" />
        <param name="image_width" value="640" />
        <param name="image_height" value="480" />
        <param name="pixel_format" value="yuyv" />
        <param name="camera_frame_id" value="usb_cam" />
        <param name="io_method" value="mmap"/>
</node>


<group if="$(arg view)">
    
    <node pkg="image_view" type="image_view" name="image_raw">
      <remap from="image" to="camera/image_raw"/>
      <param name="autosize" value="true" />
    </node>
</group>

******************************************** sterie camera to cloud point ******************

3, Depth Estimation From Stereo Video
https://cn.mathworks.com/help/vision/examples/depth-estimation-from-stereo-video.html

4: QUESTION : At this moment, we are working with a standard Videre stereo camera and make use of standard nodes available from ROS repositories to publish a dense point cloud (stereo_image_proc: node). It would be of great help if we could have some suggestions on how to correlate a point on a 2D stereo images to a 3D point in a point cloud? Do we need to make a new node to perform this operation? What would be its feasibility, accuracy?

ANSWER :

1. If you want to associate points in the 3D cloud with real pixels in the image, look at the points2 message (of type PointCloud2) that stereo_image_proc publishes. The points in the points2 message are arranged in an array that corresponds to the original image array, so points[x][y] corresponds to image[x][y].

2. If you want to project a new point in space onto the image, look at the image_geometry package (http://www.ros.org/wiki/image_geometry).

SEE ALSO : http://wiki.ros.org/stereo_image_proc http://wiki.ros.org/image_geometry