Both nucleo.fit and orione.fit are .fit files that contain 4 subimages (instead of one image with 3 channels). I think they have not been generated with PI. There are frequent problems .fit files because the FITS standard does not really define how to interpret such files.
Try loading them into PI, and generate a PI conformant .fit files. If the 4 subimages are LRGB, you need to do LrgbCombination before saving them. See output that I get with my fit files on HDRComposition below: Each .fits contains one image with 3 channels.
Georg
HDRComposition: Global context
Opening files:
C:/Users/Public/Pictures/Astro/m42_HDR/registered/M42-Core_ISO1600_40x5s_3m20s_r.fit
Reading FITS: 16-bit integers, 3 channel(s), 4290x2856 pixels: 100%
C:/Users/Public/Pictures/Astro/m42_HDR/registered/M42-Main_ISO1600_15x300s_1hr15m_r.fit
Reading FITS: 16-bit integers, 3 channel(s), 4290x2856 pixels: 100%
Computing relative exposures:
Reading FITS: 16-bit integers, 3 channel(s), 4290x2856 pixels: 100%
Reading FITS: 16-bit integers, 3 channel(s), 4290x2856 pixels: 100%
HDR composition of 2 images:
1 : 1.0000e+000 : C:/Users/Public/Pictures/Astro/m42_HDR/registered/M42-Main_ISO1600_15x300s_1hr15m_r.fit
2 : 5.1778e-003 : C:/Users/Public/Pictures/Astro/m42_HDR/registered/M42-Core_ISO1600_40x5s_3m20s_r.fit
* HDR image component 1 of 2
Reading FITS: 16-bit integers, 3 channel(s), 4290x2856 pixels: 100%
* Initial quantization:
q0 = 1.69940e+004 (14.05 bits)
q1 = 1.35110e+004 (13.72 bits)
q2 = 1.43790e+004 (13.81 bits)