convert rgb image to mean power spectral density

Hi,

I like to convert rgb data of a photo in a mean power spectrum density of this photo.

Giusi

Answer 1

[Guillaume]

For that you would first need to know the frequency response of the camera sensor that has been used to acquire that photo. I can guarantee that it's not going to be flat at all (non-scientific sensors tend to respond more to green). In addition, you'll need to know the exact settings under which that photo has been acquired (exposure, aperture, iso settings, any digital correction applied, etc.). Without any of this information it is impossible to get the PSD.

Answer 2

[Giuseppa Buscaino]

thanks,

I have a lot information from the file (see below) I wondering if would be possible however to obtain a not calibrated PSD in first.

Dimensione file:	35.1 MB
Dimens. immagine:	L (4288 x 2848)
Qualità immagine:	TIFF - RGB
Informazioni fotocamera
Dispositivo:	Nikon D300
Obiettivo:	VR 18-200mm f/3.5-5.6G
Lunghezza focale:	200mm
Modo messa a fuoco:	AF-S
Modo area AF:	Area dinamica, 9 punti
VR:	ON
Regolaz. fine AF:	No
Esposizione
Diaframma:	f/11
Tempo di posa:	1/500sec.
Modo di esposizione:	Auto programmato
Comp. esposiz.:	0EV
Taratura fine esposimetro:
Misurazione:	Matrix
Sensibilità ISO:	Auto (ISO 320)
Flash
Dispositivo:
Impostazioni Immagini
Bilanc. del bianco:	Auto, 0, 0
Spazio colore:	sRGB
NR su ISO elevati:	No
NR su esposizioni lunghe:	No
D-Lighting attivo:	Normale
Autenticazione:	Sì
Controllo vignettatura:
Controllo distorsione auto.:
Picture Control
Picture Control:	[SD] Standard
Basato su:	[SD] Standard
Regolazione rapida:	0
Nitidezza:	3
Contrasto:	D-Lighting attivo
Luminosità:	D-Lighting attivo
Saturazione:	0
Tinta:	0
Effetti filtro:
Tono:
Posizione
Latitudine:
Longitudine:
Altitudine:
Rif altitudine:
Orientamento bussola:
UTC:
Datum di mappa:

[Guillaume]

No, you will need the frequency response of the sensor. It's not something that's normally published so that's something you would normally measure under lab conditions. And I forgot that you also need the frequency response of the lens.

That response is also going to depend on the camera settings. Some of them are indeed encoded in your file but it doesn't look that it's enough. In particular, since most settings seem to be on Auto. In particular, auto ISO and auto white balance will affect the frequency response of the sensor/electronics.

In general, unless you're using calibrated equipment (that you've calibrated yourself under the exact same conditions) it's not possible to get information about the light spectrum from a photo.