Acitivity 1: Digital Scanning

As of today, almost everything is done with the use of computers. Libraries filled with the once elusive encyclopedias and journals are now being replaced by the existence of e-books, online journals and even the ultimate weapon of almost all students out there –Wikipedia. Artists such as musicians and dancers who were once having a hard time trying to find their own spotlight can now easily make it with the help of Youtube, Myspace and the hopes of having Oprah Winfrey or Ellen DeGeneres stumble on your videos. For researchers, computers and other digital gadgets play an important role in their own fields. The presence of different programming languages and its features that comes along with it transform the once complicated computational analysis into something hassle free, almost stress-free (well, maybe just a little stress) and also environment friendly! Imagine all the pieces of papers you saved in trying to find the value of x!

No matter how promising the technology of today offers us, there are still things you learned from the past that you need and continue to refer to.  Masterpieces of the past can now be easily transferred into something digital through the use of cameras, recorders, converters, scanners, etc. But what if you need something more than an ordinary scanned image? This is one of the problems researchers may encounter. Plots years ago were done by hand and its corresponding data set is not usually available in paper. 

Take for example the image presented to the right. This plot coming from Mr. Marlon Daza’s thesis entitled “Characterization of a UV Pre-Ionized TEA CO2 Laser” back in 1989 was scanned through the use of a printer available in our laboratory. However, simply scanning the image does not magically give us the data set behind this plot. If such a problem arise, the use of digital scanning can be done as a solution.

What do I mean by digital scanning? Well, this method goes further than simply clicking the scan button on your printer/scanner. It requires the use of a program that can give you the pixel coordinates of a point in your scanned image (Photoshop, GIMP and even Paint will do!), spreadsheets than can aid in keeping your data and doing simple calculations (Microsoft Excel or OpenOffice Spreadsheet), and the basic knowledge of the concept of ratio and proportion with a little shifting.

Step one: Calibration – Pixel to Actual!

 Open the file and properly adjust it in such a way that your x and y-axes are aligned parallel to the x and y-axes in your computer. With the plot slanted a certain degree, that’ll just give us a messed up data! Anyway, after ensuring that your image is properly oriented, we now begin the calibration process. It is important to find the corresponding number of pixels of a certain scale in your plot. And to do this, we zoom in towards the x and y-axis of your plot.

The red line was drawn so that we can easily see the certain scale we want to concentrate one. As for my case I chose the 0.20 to 0.30 scale (0.1) for my x-axis. Placing my cursor on the edges of the red line that I placed, I take note of the pixel coordinates of its ends as seen in the image.

The same method was done for the y-axis wherein I considered the distance between 0.00 and 0.10 yielding a same value of 0.1. After taking note of these coordinates, we now proceed to finding the number of pixels for our chosen scales. This is done by simply subtracting the initial coordinate to the final coordinates. For the x-axis, we get 0.936 pixels for 0.1 while the y-axis gives us 0.847 pixels for 0.1.

If you thought this data is enough to properly calibrate the scale of the plot to the number of pixels in the computer, you are wrong! We should take note that this is only enough when the (0,0) point in the plot scanned corresponds to the (0,0) coordinate of the pixels in the computer. As seen in the images above, the (0,0) coordinate in the plot is actually equal to the (0.869,4.287) coordinates in the computer.  Similarly, it is important to take note that the (0,0) coordinate in the plot is not always equal to 0 for our x-axis and y-axis in its actual measurements. We see that the x-axis starts its scale from 0.20 to 0.8 while the y-axis from 0.0 to 0.5.  These values are important so that we can properly shift our values in order to get the exact same scale to that of original plot. Below is the table the summarizes the values that are essential in properly calibrating our scales:

Step two: Gathering of Data – Collect them all!

Now that we have all the values we need in properly calibrating our plot, we now manually collect some data points coming from the plot. This is done by simply hovering your cursor to over the plot of choice and continuously gathering their corresponding pixels. It is important to, as much as possible, be as accurate as you can so as to have minimal error.

Step three: Conversion – Transform me baby!

Off to the fun part, the calculations proper! Now that we have the values for our calibrations as well as the pixels from our desired plot we hope to digitalize, we now convert these into its actual values. This is done by making use of the equation:

Where:

• Xplot –> corresponding x-coordinate of the specific data in pixels
• X(0,0)pixel –> actual x-coordinate in pixel of point (0,0) of scanned plot
• S Actual –>actual value of chosen scale
• S Pixel –> number of pixels corresponding to the chosen scale
• X(0,0)actual –>initial value of x-axis of the scale.

The similar equation is used for the y-coordinates only that everything must be replaced with the y-axis counterpart of the equation. Finally, we get the values of the converted x and y coordinates of our plot!

Step four: Plot – TADA!!!

With your data now available, you can now recreate the plot you have scanned on your computer! To ensure that the values you have gotten is equal to your plot, we can make use of the original plot as the background of your plot. Take note however that you must crop you plot in such a way that its (0,0) is also equal to the (0,0) in our plot.

With our finished plot overlayed on the original plot scanned looking exactly similar, it is assured that the digital scanning method is an accurate and easy way of retrieving data from a scanned plot.

It is important to take note that the accuracy of the conversion process is highly dependent on the accuracy of the values gathered needed for the calibration process. Ensuring that your original plot is properly aligned is also essential as having it inclined will give us errors which will propagate as the process goes on.

With all honestly, I have always wondered how I could be able to retrieve raw data with only the plot available in my hands. I have always thought that maybe researchers would always contact one another and try to ask for the data from their fellows but learning this method actually answers all the mystery. Proper calibration is the key to scaling things! No wonder whenever I present data from various experiments, my senior labmates would usually ask me if I have ensured the proper calibration of instruments and such before doing the experiment. Similar to engineers and architects, scaling and calibration can bring a lot of possibilities. From their sketches of their design or estimation of forces to handle a certain stress in designing bridges to actual life-sized gigantic megastructures! Even the cast of mythbusters would usual try to do their experiments in smaller scales in order to see what the results would most probably be.

Overall, having fully understood the concept behind the technique and being able to even list down in detail the methods as well as getting a final plot that looks almost exactly similar to the original plot… I would like to give myself a score of 10/10. Plus, I have even managed to come up with an equation that’ll work out in almost all cases just as long as all the values needed are taken note of accurately.