Jonathan Thomson's web journal

Results September 7, 2012

 

Spectral Calibration
I used the spectral lines from a neon lamp to verify the calibration that resulted from using a CFL’s mercury spectrum for spectral calibration.

 

Radiometric Calibration
This radiometrically calibrated spectrogram of a 60W clear Sylvania bulb is an average of nine spectrographs, has been flat frame corrected, and has the dark frame removed. The percent error is 0.29019% when compared to the total irradiant flux of a mathematical model of the 60W incandescent lamp’s spectrum.

This spectrogram of a 60W clear GE bulb has a higher total irradiant flux percent error of 1.2519%. The shape of the spectrum doesn’t match the mathematical model nearly as well as the Sylvania bulb. This is probably because the ensemble images used to produce the GE bulb’s spectrogram had a lower SNR because less light was radiated toward the slit due to the shape of the bulb’s filament. The distortion might also be a result of the bulb having a very different spectral emissivity from the model. For more on the variability of the spectral emissivity of tungsten filaments see this article by Far Associates farassociates_tungstenfilaments.pdf

 

Impact of Noise Reduction
This radiometrically calibrated spectrogram is computed from only one spectrograph, has been flat frame corrected, and has the dark frame removed. The percent error is 0.52002% when compared to the total irradiant flux of a mathematical model of the 60W incandescent lamp’s spectrum.

This radiometrically calibrated spectrogram is an average of nine spectrographs, has NOT been flat frame corrected, and has the dark frame removed. The percent error is 0.3003% when compared to the total irradiant flux of a mathematical model of the 60W incandescent lamp’s spectrum.

This radiometrically calibrated spectrogram is an average of nine spectrographs, has NOT been flat frame corrected, and has NOT had the dark frame removed (a dark level of 32 was subtracted from the spectrographs). The percent error is -2.8722% when compared to the total irradiant flux of a mathematical model of the 60W incandescent lamp’s spectrum.

 

It appears that flat frame correction barely impacts incandescent spectrograms. Judging by the change in the percent error of a spectrogram without dark frame correction, one might concluded that the impact of dark frame correction is negligible. However if you compare the spectrogram’s plot to the corrected one you can see quite a difference where the signal of each channel is low.

 

JPEG Tone Linearization
JPEG encoding involves a non-linear conversion of the raw data captured by the CCD. I attempted to find an inverse tone curve that would re-linearize the data in a JPEG so that it matched the data in a raw of the same subject, but I was unsucessful. The method I used involved capturing several images at different exposure times of the LCD I used to make flat frames. Exposure time has a linear relationship with the number of photons admitted and therefore counts measured by the CCD. Therefore relating exposure time to JPEG intensity is the same as relating counts to JPEG intensity. Once I found the curve that counts to intensity, I inverted it to find an inverse tone curve. Unfortunately this inverse curve didn’t give accurate results when used to linearize JPEG data.

 

Absorption Spectroscopy
I tried a small absorption spectroscopy experiment by placing a cuvette filled with helium in front of the slit. I wasn’t able to locate any helium absorption spectral lines in the resulting spectrograph.

 

Continue Reading

>>>Next — Code>>>

<<<Previous — Determining Camera’s Gray and Taking RGB Flats<<<

 

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s