Day 16 - more code

Today i added to my original program to make it so the user can now input the values by clicking on a b-scan. These values are then used to focus the data to the most important part, this includes removing air and other noise and looking at just the echo from the paper. In my down time today i tried to learn more about the Hilbert transform, because i didn't understand Maria's explanation even though it was very well done. It is hard to teach linear algebra concepts to a kid who has not even taken calculus. Monday I will hopefully fix this program and begin focusing all of our data so we can analyze it for the rest of the week and begin sending information to Portugal.

Day 15 - Data Focus

today Kunal helped me to finally get my program working that takes every A-line and crops the useless data out. This will create more precise data for us to analyse when we begin our quantitative analyses. Tomorrow we will probably begin calculating normalized moments out of each A-line.

Day 14 - Hanning

The first hour of my day was spent filing a report with the campus safety office about the recent damage to my car. After that i went to Kelsey's lab (because mine was being cleaned) to make a few more 3D surface plots of the paper and I also read a couple articles of propagated ultrasound detection. The basic idea is that a lower resolution transducer can pick up smaller objects then theoretically possible by receiving data from the horizontally propagated waves after collision with the surface. The rest of my day consisted of understanding the post processing matlab script to make a program that creates a Hanning window over the data to focus on the most important part.

This is a 3D model of the surface of photo paper, as you can see there are multiple very small (900 micron thick) lines that run across the  paper this is the key to what makes photo paper hold so much ink.

Day 13 - ImageJ

Today maria and I spent a lot of time using image J to create stacks of our B and C Scans. Then we made 3D models of the images, but matlab and Image J interpret color scales differently so we went back through and reprocessed the images on a 8 bit color scale (gray scale) to then reload them into Image J. We received some spectacular results that i will show you tomorrow.

Day 10 - Dark Side of The Paper?

after looking at the papers through a microscope we decided to also image the back side of the paper. This led to some very interesting results showing how the paper fibers were aligned through out the entire paper and how the pattern changed. we noticed that with the photo paper that it has vertical fibers on top and horizontal fibers on the bottom, that are all held together by diagonal fibers in the middle. Not something someone would expect from looking at this very reflective flat paper.

Day 12 - Rochester Precision Optics

Today we visited the Rochester Precision Optics factory. They produce lenses for everything from thermal sights on M16s to focal lenses in the Hubble telescope. After the tour Maria and I worked on making stacks of our images to better display the data collected. We also created 3D images out of the C-Scans by using a color analyzer in imageJ. hopefully these will help display the caracteristics of the paper to the Portuguese paper company.

Day 11 - more scans

Today was mostly about using my PSF data to redo my previous scans to make them more accurate. I had a problem with the post processing in our lab view program so Kunal made an addition to the post processing program to allow inputs. This makes it so I can use the program outside of lab-view. I am currently finishing up my second to last scan, tomorrow I will be processing these scans to get images.

Day 9 - Programming

Today i started out by making a program that graphs a best fit line over my data for my PSF in mat-lab. Then i helped Kelsey in the freshmen lab by making a program that selects multiple images to be separated into three color filters (red, blue, and green) to be used for their polynomial texture maps. tomorrow i plan on making some scans in the morning then analyzing some of Maria's code to determine standard deviation and altering it to my purposes.

Day 8 - PSF

Today I came in and Maria wanted me to remap the PSF (point spread function) of my transducer using a very thin wire. using this function we determined the finest step size that we can use to resolve an accurate image. we then graphed this function (distance in microns vs volts) on a matlab figure.

Day 7 - Matlab

Today i spent most of my time going through Matlab tutorials and reading a book on Matlab. I also helped Jessica try and find her PSF for her transducers so she can use that info. Finally i spent some time with Kelsey to see what she wants as a UI for her image software.

Day 6 - Problem Solved

Today one of Maria's grad students came back from vacation to finish up his paper on bio-films. He showed me that the problem with the data sharing between the oscilloscope and the computer was because the ip address was DHCP instead of static as I had assumed(silly me!). once we made the change in the labVIEW program everything forked great.

The rest of the day was spent determining the PSF (point spread function), the pint of this(haha pun) is to find the resolution of the transducers by seeing how well it images a point in space and how much noise around the wire exits. In the image to the right that rounded object is a 1,000 micron thick wire. but this image shows this point as being 6,000 microns in width, this is the point spread function. Now we know that any step size between reading smaller than this is simply a waste of time, because then you are collecting redundant data that will eventually be over lapped.

Day 5 - Epic Fail

First thing in the morning I come in and my computer is quarantined on the RIT network, for "unknown" reasons. So when I come in I spend about 45 minutes talking to the ITS help desk to receive guidance from their infinite knowledge of computing. And as if I was calling a corporate help-desk I was passed along from person to person because no one could fix my problem, then at last they passed me on to Brett (the CIS tech specialist) I   soon remembered that he was only 6 doors down the hall. So I just hung up the phone and walked down to his room he told me to talk to Jordan in the computer lab. After bringing him to my lab he quickly fixed the problem with some more discussion with the help-desk. At this point we are about two hours into my day, and now that I can access all of my data on the network drive called Tank I can start another scan. But the program continued to crash because it was having trouble connecting to the oscilloscope. After lunch I couldn't fix the problem so to prevent from making the entire day a loss, I went upstairs and started working with Kelsey on making another program in Matlab to process multiple images into 3 different spectrum's.

Day 4 - Scan Time

Since we got the program to work yesterday I was able to image all of the pieces of paper today. I took a one centimeter squared cross-sections at 5 depths lower than 17000 microns. The first images were 25 by 25 readings and the rest were 35 by 35, that i tried to get a better resolution out of the images. I ran a total of 8 scans today and the program worked relatively well except for a few crashes here and there. Tomorrow i will post some of my results that i have acquired.

Day 3 - we're going to need a bigger scan

First thing this morning when i came in i was trying to get our C-scan program to produce a result. But after messing around with lab view (the editor/compiler for the program) I noticed that the problem stemmed from an issue with the temp folder not being able to be overwritten. Lucky for me, mike (a MET graduate student) came over to help me out. Mike showed me that when the program fails it doesn't have time to dump the temp folder and leaves garbage files in it. Right before lunch we got the program working and took a small B-scan of the paper. To the right is one of the B-scans we generated,

and shows just noise until about 20,000 microns and then air, that line at about 50,000 microns is the paper. Since this turned out pretty well we decided to do a much larger scan, a scan so large we later found out it has never been attempted before. And yes it was to large we estimated it would have taken 2 hours of just sampling not even including post processing. So after lunch we canceled the scan and went on with smaller scans. also through out the day I went and visited Matt in the space weather lab, Jason in the astronomy lab, and Elaine in the MRI lab to see what everyone was working on.

Day 2 - The Transducer

The beginning of my day started with taking simple pictures of the paper samples to show the texture of the samples before we scan them. The next step was setting up the mount to hold the paper in place as the transducer ran across the top of it. Then we used a substance that we knew was already reflective to calibrate the transducer. just like any other lens the transducer has a focal length, but this one is unknown. Using the oscilloscope we move the transducer until we have maximized the reflection off of the substance. Then we worked on orienting the transducer perfectly vertical, this is done by looking at the echo and trying to make it symmetrical  Think of a cosine wave with a slightly higher peak at the origin, that is our goal, and it takes much longer than you would think.

Day 1 - introduction to project

The beginning of my day mostly consisted of reading the first three chapters of a textbook about diagnostic ultrasound. This laid down the basics so i could understand maria's lesson later. She started out by teaching me how the reflection of sound wave off of an object creates an A-line. then the transducer can be moved to form a vertical 2D slice of the object called a B-scan, after you have acquired multiple B-scans they can be compiled to form a 3D image of the object. The final step is to take a horizontal slice of this object, which is called a C-scan. Next maria started to show me how the equipment was set up. This was fairly straight forward, the transducer was hooked up to a pulser/receiver that would switch between sending the pulse and then reading it. this was hooked up to an oscilloscope which displayed the function as volts on the y-axis and time on the x-axis. then the oscilloscope was hooked back up to the pulser/receiver to stabilize the image in one location. The pulser/receiver also had dials to set the HPF(high pass filter) and LPF(low pass filter), both of these are used to filter out the image to show the most important wave, the first reflection. The rest of the day was spent learning how to use all parts of the oscilloscope.