Add PCLPointCloud2 Header fields (Stamp, Seq, Frame_id) to PCDWriter/PCDReader

The PCDWriter/PCDReader combination deals with PCLPointClouds2. However it does not include the header fields in the file format(Stamp, Seq, Frame_id).

Your Environment

• Operating System and version: Ubuntu 17.10
• Compiler: GNU GCC
• PCL Version: trunk

Context

The newest trunk allows serialization to an iostream, allowing an easy way to use the PCD format over the network. However the header fields are not stored in the PCD format. It could also be useful for figuring out metadata of the PCD at a later time.

Expected Behavior

Converting a PCLPointCloud2 to PCD Format and back results in the header field in PCLPointCloud2 being identical to the pointcloud before conversion.

Current Behavior

Converting a PCLPointCloud2 to PCD Format and back results in an empty header.

Possible Solution

I propose adding three fields to the header
STAMP (Time since epoch in microseconds, header.stamp)
FRAME (Frame name, header.frame_id)
SEQ (Sequence id, header.seq)

Start at the PCDWriter and PCDReader entry points and trace how PCLPointCloud2 data is serialized and reconstructed. Check how header.stamp, header.frame_id, and header.seq are represented, then verify that a PCLPointCloud2-to-PCD-to-PCLPointCloud2 round trip preserves those fields.