CloudCompare forum

Hello CC community
this is an issue initialy posted here https://github.com/CloudCompare/CloudComPy/issues/161

actually the issue is the following.
I've received a E57 file, containing several TLS stations embeeded in that single E57 file.
please note that each of TLS also contains cubic map imageries. https://en.wikipedia.org/wiki/Cube_mapping
when I load that E57 file into CC, spatial information (or pose) of cubic map image is well being read. No issue;
(The idea is to import photo & external orientation to map a 3D mesh model created on the pointcloud)

now I want to break down those TLS stations and split that single E57 into several E57 files, one per station.
so I pre-process the E57 file in order to break it down into several E57 files per scan station.
I use that script command :
CloudCompare -O <my_single_E57_file>.e57 -C_EXPORT_FMT E57 -SAVE_CLOUDS

but we noticed the resulting E57 files lost spatial information and drop metadata of cubic map image, even though images are well being stored...
when We check the XML structure of the resulting files (using e57xmldump.exe), The images are described by the visualReferenceRepresentation structure. This structure can be used for viewing, but is not suitable for evaluating the position of the scanner in space, so apparently CC skips such images metadata...

does it come from CC library ? Am I understanding well ?
antoine

Yes, sadly there are a lot of specific structures in a E57 file that CC can't load. Therefore it is lost in 'translation'... Sadly CC is generic and can't map its internal structures with a specific format. Not sure how to improve this.

Is there any command line options to load the points and sensors only into CC and exclude the Images? The image loading is where it crashes.

Nope... that would have to be added as an option.

Okay, progress on this mystery. I opened the e57 in recap and saved it as an e57 without the images. Works great in CC. The file was 10GB less. (239 sweeps). The sensors are present along with the individual clouds.

I can unpack the e57 with e57unpack.exe into a directory structure and delete the images.

Anyone know how to repack the e57 file from the directory structure?

I would really like to be able to do all of this without recap.

The reason that I am doing this is I want to filter out points from each sweep that are greater than 10M, hopefully this with get rid of some of the noise.