I have not tried this, but on the FME HUB you can find a RasterHSVConverter which looks like does what you need.

Never tried this, but on the FME hub you can find the RasterHSVConverter.

Never tried it, but on the FME hub you can find the RasterHSVConverter which probably does what you need. For some reason safe won’t let me add a link to their own FME hub, but you can google it.

thanks for your hint @nielsgerrits . I tried it and it is generated a hsv color space of the tiff image.

But the values for saturation and value are sometime over 100%. I thing the values should be between 0 and 100%! Maybe the calculation inside the transformer is different?

I’m afraid I can not help you without diving into it. You can right click the transformer and choose edit to see what is inside. The computing is done using Python in a the PythonCaller.

In the writer you can choose YCbCr as the colorspace, but not HSV

https://docs.safe.com/fme/2023.1/html/FME-Form-Documentation/FME-ReadersWriters/tiff/feature-types-w.htm

Could be worth submitting an idea to also specify HSV?

As @nielsgerrits suggests, python is probably going to be your best option here. There’s some helpful snippets on the subject of using PIL and Pillow here: https://stackoverflow.com/questions/22236956/rgb-to-hsv-via-pil-and-colorsys

Of course the option to specify hsv in the writer would be nice!

I think for raster data processing (pixel-based color segmentation) it is helpful to have hsv color space instead for rgb sometime!

I will try it out via python.