完成 双目视差匹配disparity matching, 生成当前相机坐标系下的点云point clouds camera's local reference frame.
a. <stereo>/left/<image> (sensor_msgs/Image) 左图矫正后的图像 Left rectified stereo image.
b. <stereo>/right/<image> (sensor_msgs/Image) 右图矫正后的图像 Right rectified stereo image.
c. <stereo>/left/camera_info (sensor_msgs/CameraInfo) 左相机参数 Camera info for left camera.
4. <stereo>/right/camera_info (sensor_msgs/CameraInfo) 右相机参数 Camera info for right camera.
a. disparity (sensor_msgs/Image) 视差 话题
Image representing the calculated disparity at each point.
Note that the values are scaled to be in the range [0, 255]
b. point_cloud (sensor_msgs/PointCloud2) 左相机当前坐标系下 的点云地图
The point cloud assosciated with the calculated disparity in the left camera's local reference frame.
c. frame_data (elas_ros/ElasFrameData) 结构数据 用于三维重建
Frame data used for the point cloud concatenation process. See node pc_construction.
a. ~queue_size (int, default: 5) 队列大小
Length of the input queues for left and right camera synchronization.
b. ~approximate_sync (bool, default: false) 同步
Whether the node should use approximate synchronization of incoming messages.
Set to true if cameras do not have synchronised timestamps.
This node concatenates the point clouds generated by the elas_ros node.
<frame_data> (elas_ros/ElasFrameData) 结构数据 用于三维重建
The frame data generated by the elas_ros node.
<pose> (geometry_msgs/PoseStamped) 相机姿态数据 来自视觉里程计
The pose transformation from base frame to pose frame.
This accounts for motion of the camera. This can typically be determined by appropriate visual odometry.
See usage notes below.
point_cloud (sensor_msgs/PointCloud2) 世界坐标系下的点云地图
The concatenated point cloud.
~queue_size (int, default: 5) 队列大小
Length of the input queues for frame data and pose synchronization.
~approximate_sync (bool, default: false) 同步 Whether the node should use approximate synchronization of incoming messages. Set to true if frame data and pose do not have synchronised timestamps.
base_frame_id (string) 基坐标系
The name of the base frame that the concatenated point cloud should be cast into.
pose_frame_id (string) 相机位姿坐标系
The name of the frame that the given pose transforms into.
- 文本1
- 文本2
- 文本3
如果你希望有序列表, 也可以在文字前面加上 1. 2. 3. 就可以了,例如:
- 文本1
- 文本2
- 文本3
一盏灯, 一片昏黄; 一简书, 一杯淡茶。 守着那一份淡定, 品读属于自己的寂寞。 保持淡定, 才能欣赏到最美丽的风景! 保持淡定, 人生从此不再寂寞。
EwenWan
妙趣横生
Hello GitHub
第一段
第二段
Tables | Are | Cool |
---|---|---|
col 3 is | right-aligned | $1600 |
col 2 is | centered | $12 |
zebra stripes | are neat | $1 |
![][1] [1]: http://latex.codecogs.com/gif.latex?\prod%20(n_{i})+1
- 基于局部的块匹配 Block Matching(BM) StereoBM
- 半全局块匹配 Semi-Global Block Matching(SGBM) StereoSGBM
第一步对每一个Pixel使用块匹配BM进行匹配,得到了全部Pixel的disparity map。 第二步对Disparity map建立图,用Graph Cut对其进行全局优化。 利用Rectification将二维转化为一维
- 基于全局的图割 Graph Cut(GC)cvStereoGCState()
- 基于全局的 动态规划 Dynamic Programming(DP)cvFindStereoCorrespondence()