Spectrum
Initially the fits file was loaded into python, the image was then projected onto the x-axis to obtain a spectrum
. This was in the units of N_{adu} against position, so from the eqaution N_{pe} = 2.3*N_{adu} the data was plotted in N_{pe} against position.
- Cadmium_spectrum.png:
Background
To calculate the background of the data, the position between 350 and 500 was selected, this have a plot of which had a bright pixel as shown,
- Background with bright pixel:
by ommiting the bright pixel a histogram of the data could be plotted.
- Background histogram:
The mean and the standard deviation were calculated using the numpy package and were equal to:
mean =
149775.0,
stdev = 20571.1,
however for a poisson distribution where,
f(n,l) = (l^n)( e^{-l})/(n!)
l is equal to the expextation value and the variance, therefore for the distribution to be poisson the mean should be equal to the square of the standard deviation.
Spectrum with removed background
The background could then be reomved from the inital spectra of the lamp leaving only the significant peaks. This was done by removing the mean+stdev.
- Cadmium_removedBackground.png:
--
JosephBayley - 13 Oct 2015
Comparison with Unfocused Image
For Comparison, an unfocused image taken from the Cadmium lamp was analysed. As before, the fits file was loaded into python, projected onto the x axis, and then multipled by the gain, 2.3, to gain the number of photoelectrons.
As you can see, the spectral lines have been broadened significantly, reducing the accuracy of locating the peaks. This shows the difference between the quality of the focused image and the unfocused image, highlighting the difference it would make to our data analysis.
Questions to ask at next meetingon 14/10/15:
- How to fit a poisson distribution to our histograms for the background
- How to find the exact position of the peaks.
In comparison, the background for the unfocused region is much more irregular than for the focused image.