Hello Delphine,
without looking at your vi's:
In 32bit-mode you got a blue picture. That's because of RGB-interpretion of your data. You supplied only the lowest 8 bit, this gives you the blue color. To get "gray" you have to supply R, G and B. All you have to do is to "multiply" your 8bit to 24bit: join 3 U8 values to a U32 for each pixel.
In 8bit-mode you got a multicolored picture. That's probably because you didn't supply a colortable. If you would supply a colortable with all gray values your picture should be fine.
Best regards,
GerdW

using LV2016/2019/2021 on Win10/11+cRIO, TestStand2016/2019