Day 19 Blog

Today in the morning I took measurements with the Shack-Hartmann wavefront sensor. The rest of the morning was spent working on the code. For the Shack-Hartmann sensor I need 3 different code parts. One to analyze the the intensity data, the other to measure the position data and the final (and hardest) is one that takes the centroid data, does a bunch of stuff to it to get the displacement data, then uses that to get the wavefront slope data.

At 11 all of the interns took a bike ride to U of R and ate lunch there before coming back to RIT around 1:30. The bike ride was a lot of fun and a good way to interact with the other interns.

After the bike ride I went back to coding and finished the position code and started the slope code.

Outline

• Title Slide
• Summary
• What is going to be discussed in the presentation
• Basics of wavefront sensing
• what a wavefront is,
• Purpose of what I am doing
• What is an optical differentiation sensor
• diagram and explanation
• prototype
• Comparison between OD and Shack-Hartmann sensor
• OD has better dynamic range, can look at polychromatic sources
• Important to make sure quality of image is the same
• Explanation of Signal to Noise Ratio
• Important to compare different sensors
• Shows how much of original image is preserved
• Equation for the SNR of OD and SH sensor
• From paper by Oti
• Explanation of terms
• Derivation
• Numerical calculation of SNR of both sensors
• Estimation for number of photons and ensemble average
• Actual calculation
• Analysis of numerical calculation
• Oti Graph
• Difference in Airy rings
• Atmospheric turbulence term for Shack-Hartmann sensor
• Graph and explanation
• Centroid error term for Shack-Hartmann sensor
• Graph and Explanation
• Larger project and Deformable mirror
• To develop a sensor to compete with the Shack-Hartmann
• SNR of camera for broadband
• graphs and analysis
• SNR of camera for lasers
• graphs and analysis
• Different way of calculating SNR
• graphs and analysis
• Blur and mean functions
• code and results
• Analysis of camera SNRs
• Conclusion    

Day 18

Today I analyzed the SNR graph from yesterday. The second linearity is the reason for that large dip in SNR, which is very surprising since in everything I have read so far there time where more photons means less SNR.  This afternoon we had a very interesting discussion from Bob Cremens, who works on analyzing forest fires. In the afternoon I worked on by outline and updating my abstract for the final presentation.

Tomorrow I will be working on data from the Shack-Hartmann sensor. I started writing code in order to be able to analyze the data tomorrow, but this code seems complicated to it will probably take a few days to completely iron out.

Day 17 Blog

In the morning we had a meeting with Jie before she left about our plans for the future. She told me to see if the weird stuff I saw yesterday was still present with another camera.

Once the meeting was over, we set up the other camera and analyzed the data for the new camera, unfortunately even after setting it up the data we got from it was meaningless. Today I also turned the graphs of variance into graphs of SNR.

Day 16: Some more data

In the beginning of the morning I wrote the sort that I talked about yesterday. The following graph is a comparison of the sorted data(orange) and the data from the sensor(blue), as well as a zoomed in version to show what the sort algorithm(and blur) does to the data

The data is now close enough to a function in order for me to be able to analyze it through matlab. Based on theory as well as our broadband results, we are expecting some sort of linear relationship between the mean and the variance. A pure derivative of the data is meaningless due to the noise that I still can't remove from the data, but matlab has a linear regression which I ran on the two obvious sections of the data. Since there were questions about the validity of the data with the turbulence, we got new (and better data). Which confirmed that there were some major influences to the data at some point in the data. (Graph is data with sort and blur algorithm).

Day 15: (actual data)

Today in the morning I took a bunch of data from the Shack-Hartmann sensor. The first trial I did involved me taking a bunch of data with broadband light shined onto paper over the sensor. I varied both the intensity and the exposure time and found that the variance seemed to be linear with the amount of photons. We tried to do the same experiment with the laser, but was our result due to there being pixelation of the paper. As a result, we tried keeping the intensity the same and measuring how the image varied, and I got

Due to the huge variation , I wrote a blurring program which gave me slightly better results, but not simplified enough to let me analyze it.

At the end of the day I turned the points into dots instead of lines and examined it and found that the problem stems from

so tomorrow I will be writing code that will sort the x vector and removes any values that repeat, as well as go through the y vector and for each x value that is the same, average all of the y values that correspond to that one point.

Day 14 Summary

Today everyone but our PHD student (who has her own project on how heat is transferred with very high powered lasers) is working on a new deformable mirror that the lab has received. Currently my task is to look at the SNR from the results that we get from the mirror, but as I can't do that until we have the mirror set up, I am looking for more equations that could explain the SNR of the Shack-Hartmann sensor. Today ended up being another day looking at papers, but tomorrow I am going to be analyzing some baseline values for the signal to noise ratio. Tomorrow I will be looking at data from the sensors, so at the end of the day I started writing some preliminary code to look at tomorrow's data.