基于easy_handeye开源项目，对Franka_panda进行手眼标定（Kinect v2）；里面主要是修改了一些easy_handeye中的一些launch文件，实现手眼转换矩阵的计算与发布。

1. 需要安装easy_handeye包;参看https://github.com/IFL-CAMP/easy_handeye；
2. 安装我们修改的panda_moveit_config，以及panda_server；

1. 自行安装好相机和标定板，标定板使用 ar_track_alvar生成的10cm*10cm的二维码标签，id为7；注意标定板完全处于相机视野范围，并且距离桌面一定距离，防止自动移动时与桌面出现碰撞

   

2. 解锁panda机械臂，并启动Moveit

   cd ~/catkin_ws #cd where you put your "panda_client.sh"
   source panda_client.sh -r
   # ctrl+c
   roslaunch panda_moveit_config zzu_panda_move.launch

3. 另一个命令窗口，打开相机

   cd ~/catkin_ws #cd where you put your "panda_client.sh"
   source panda_client.sh
   roslaunch kinect2_bridge kinect2_bridge.launch publish_tf:=true

4. 打开窗口进行监视

   cd ~/catkin_ws #cd where you put your "panda_client.sh"
   source panda_client.sh 
   rosrun kinect2_viewer kinect2_viewer kinect2 sd cloud

5. 另一个命令窗口，进行标定，按照提示进行标定

   cd ~/catkin_ws #cd where you put your "panda_client.sh"
   source panda_client.sh 
   roslaunch panda_hand_eye_calibrate panda_eob.launch

1. 发布手眼姿态矩阵

   cd ~/catkin_ws #cd where you put your "panda_client.sh"
   source panda_client.sh 
   roslaunch panda_hand_eye_calibrate publish_panda_eob.launch

2. 解锁panda机械臂，并启动Moveit

   cd ~/catkin_ws #cd where you put your "panda_client.sh"
   source panda_client.sh -r
   # ctrl+c
   roslaunch panda_moveit_config zzu_panda_move.launch

如果tf出现问题，就先重新启动Moveit步骤；