ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
1 | initial version |
The kinect sensor, like all other sensors, doesn't provide perfect measure, there's a variation in the depth but it shouldn't be too large.
http://en.wikipedia.org/wiki/Sensor
Regarding the depth of a particular point, you should do some point cloud tutorials so you can understand the meaning of the data you are receiving. For each point you can obtain the x, y and z coordinates in the frame of the camera.