Image frustrum of FOV isn't calculated when adding via gRPC directly

[Mark-Niemeyer]

The calculation of the bounding box based on the frustrum is called correctly when loading from HDF5:

https://github.com/agri-gaia/seerep/blob/8aa1b30be6f9515fc4b7ec4a77cf9ad7d38cbb69/seerep_hdf5/seerep_hdf5_core/src/hdf5_core_image.cpp#L57C3-L57C24

but it isn't when adding via gRPC directly. It also has to be called here:

seerep/seerep_srv/seerep_core_fb/src/core_fb_image.cpp

Lines 54 to 57 in 8aa1b30

	auto hdf5io = CoreFbGeneral::getHdf5(dataForIndices.header.uuidProject, m_seerepCore, m_hdf5IoMap);
	hdf5io->writeImage(boost::lexical_cast<std::string>(dataForIndices.header.uuidData), img);
	m_seerepCore->addDataset(dataForIndices);

and here:

seerep/seerep_srv/seerep_core_pb/src/core_pb_image.cpp

Lines 54 to 57 in 8aa1b30

	auto hdf5io = getHdf5(dataForIndices.header.uuidProject);
	hdf5io->writeImage(boost::lexical_cast<std::string>(dataForIndices.header.uuidData), img);
	m_seerepCore->addDataset(dataForIndices);

before adding it to the core