Removed extra -1 y,z multiplier for static links

[IanTheEngineer]

As I integrated the realsense static transforms with our system, I noticed there were a few unnecessary -1 multipliers in this driver for the camera extrinsic frames read from each realsense. When I overlaid them with the URDF, it became more obvious the -1's needed to go. The following are screenshots of the corrected static TF's read and published from a live realsense D415 camera (I also tested with a D435 and D435i and found these changes to work for all three models):

Front View:

Top View with Frame Names:

[IanTheEngineer]

Just a follow-on comment: I suspect a similar issue is occurring in the refactor branch when calculating the camera frame translations -
https://github.com/intel/ros2_intel_realsense/blob/refactor/realsense_ros/src/rs_base.cpp#L283-L285

void RealSenseBase::composeTFMsgAndPublish()
...
  geometry_msgs::msg::TransformStamped msg;
  RCLCPP_DEBUG(node_.get_logger(), "Publish Static TF from %s to %s", from.c_str(), to.c_str());
  msg.header.stamp = t;
  msg.header.frame_id = from;
  msg.child_frame_id = to;
  msg.transform.translation.x = translation.z;
  msg.transform.translation.y = -translation.x;      // <--- here
  msg.transform.translation.z = -translation.y;      // <--- and here
  msg.transform.rotation.x = q.getX();
  msg.transform.rotation.y = q.getY();
  msg.transform.rotation.z = q.getZ();
  msg.transform.rotation.w = q.getW();
  static_tf_broadcaster_-> sendTransform(msg);

The translation frames read from the realsense camera device already have the correct sign.

[ahuizxc]

Just a follow-on comment: I suspect a similar issue is occurring in the refactor branch when calculating the camera frame translations -
https://github.com/intel/ros2_intel_realsense/blob/refactor/realsense_ros/src/rs_base.cpp#L283-L285

void RealSenseBase::composeTFMsgAndPublish()
...
  geometry_msgs::msg::TransformStamped msg;
  RCLCPP_DEBUG(node_.get_logger(), "Publish Static TF from %s to %s", from.c_str(), to.c_str());
  msg.header.stamp = t;
  msg.header.frame_id = from;
  msg.child_frame_id = to;
  msg.transform.translation.x = translation.z;
  msg.transform.translation.y = -translation.x;      // <--- here
  msg.transform.translation.z = -translation.y;      // <--- and here
  msg.transform.rotation.x = q.getX();
  msg.transform.rotation.y = q.getY();
  msg.transform.rotation.z = q.getZ();
  msg.transform.rotation.w = q.getW();
  static_tf_broadcaster_-> sendTransform(msg);

The translation frames read from the realsense camera device already have the correct sign.

Hi Ian,The reason we do this is consistent with the ros version.
See here:
https://github.com/IntelRealSense/realsense-ros/blob/5190897d906a51d9bda437d8afe7d361102e5171/realsense2_camera/src/base_realsense_node.cpp#L1822