Hi Everyone,

I am trying to do Eye-to-hand calibration with ros package (https://github.com/ros-planning/moveit_calibration.git). I have a 6 DOF robot and the logitech C270 camera. After launching the rviz I added hand eye calibration panel,created the marker and in context tab i selected frames such as base0 for base frame, link6 for endeffector, Handeye target for target frame and finally usbcam for sensor frame in the drop down menu. I also set the usbcam initial pose guess with respect to my base0 such as x= 0.01, y= 0.500, z= 0.750 Rx = -1.87,Ry = 2.99, Rz = 0.35 (in metres and radians respectively) which is i have manually measured and given.

Now I have started the calibration process by taking samples. I have fixed marker to my link6 (endeffector). Once I have take 4 samples and when go for the 5th sample my camera is calibrated and giving the transformation matrix from base0 to usbcam frame.

Problem
The resulted transformation matrix is very far away from my physical setup.In clear, the resulted translational values are very different from my physical setup. Can anyone tell me what I have done wrong or any other steps I need to follow to get the correct transformation matrix from base0 to usbcam. I am stucking in this part for many days.If anyone helps me it will be very useful.