DIMAP: for PNEO products, use the new color interpretations for the NIR, RedEdge and DeepBlue/Coastal bands

[rouault]

CC @tbonfort

In b3f3ce7 , you added logic to register metadata on the bands, but it doesn't look to work with the fake dataset we have in autotest/gdrivers/data/dimap2/vhr2020_ms_fs/MS-FS/DIM_MS-FS.XML, nor a PNEO_ORTHO_BUNDLE-FS_REFLECTANCE_JPEG2000_4326_NONE_STD_A one I likely got during the development of the PNEO support.
The reason I ask is that I rather see a Radiometric_Data.Radiometric_Calibration.Band_Spectral_Range (rather than Radiometric_Data.Radiometric_Calibration.Instrument_Calibration.Band_Measurement_List as expected by the driver) element with subnodes like:

                        <Band_Spectral_Range>
                                <BAND_ID>R</BAND_ID>
                                <CALIBRATION_DATE>2019-05-20T09:07:00Z</CALIBRATION_DATE>
                                <MEASURE_DESC>Spectral Range values of raw radiometric Band</MEASURE_DESC>
                                <MEASURE_UNIT>micrometer</MEASURE_UNIT>
                                <MEASURE_UNCERTAINTY>0</MEASURE_UNCERTAINTY>
                                <FWHM>
                                        <MIN>0.619</MIN>
                                        <MAX>0.69</MAX>
                                </FWHM>
                                <GAIN>0</GAIN>
                                <BIAS>0</BIAS>
                        </Band_Spectral_Range>

I guess we could take the mean of FWHM.MIN and .MAX as the central wavelength, and MAX-MIN as the FWHM value to report in the new IMAGERY band metadata domain (cf https://gdal.org/en/latest/user/raster_data_model.html#imagery-domain-remote-sensing)

[coveralls]

coverage: 69.45% (+0.005%) from 69.445%
when pulling 79b446f on rouault:DIMAP_PNEO_GCI
into c044df7 on OSGeo:master.