Extract data from a scanned PDF Chart

[IndyTek]

Hi,

I have a scanned PDF chart, which shows linear relationship between two variables. Is there a way to extract data from the scanned PDF using Acrobat?

I want to avoid error in my calculations by eyeballing the data. Using "Measuring tool" may be an option, but wanted to ask whether any forum members have a better and efficient way to extract data, which can later be used in a spreadsheet software.

FYI, I have Acrobat X installed on my Windows computer.

Thanks in advance for your help.

 

[FredWagner]

From Acrobat X, try Save As, to an Excel spreadsheet, and then see if Excel will give you tabular data from the graph.

Fred Wagner

 

[IndyTek]

Fred,

Thanks for your reply. I tried this option, but it does not work.

As an example, please refer to the attached link, I will like to extract data from Figure 6 in this document:

http://www.seas.columbia.edu/earth/wtert/sofos/nawtec/nawtec13/nawtec1

3-3164.pdf

Is there an alternative way to extract data?

 

[FredWagner]

IndyTek - I'm unable to access the links you provided. You may save yourself some time by just printing the graph on paper with a preprinted graphic overlay, and working out the equations based on a couple of the points. Since you're at Columbia University, you might see if you can get some TA's in either the Math or Engineering departments to help you. Have you tried contacting the author of the paper with the graphs, to ask if you can get the data the graphs were derived from ?

Fred Wagner

 

[IndyTek]

Fred,

Try this link:

http://www.seas.columbia.edu/earth/wtert/sofos/nawtec/nawtec13/nawtec13-3164.pdf

If you are unable to access this link, you can google "generator output vs throttle flow" and this is the fifth link in the results.

I am not at Columbia Univ, just found the paper, in order to provide an example.

Regarding your suggestion of preprinted graphic overlay, I will try and see if it works.

Thanks!

 

[FredWagner]

Indy - the link did open, to a 14 page PDF, with some interesting graphs in the 2nd half. Some of the graphs were simple straight lines, but the one with the jagged lines was based on one reading a day, over a month. That would not be appropriate to try and derive specific math functions from, it just shows a trend.
If the graphs with smooth straight lines contain the info you need, just interpolate some readings where the data line crosses one of the calibrated lines - I wouldn't worry about trying to automate the process - the graph is intended to be interpreted visually - like in the old days when we used slide rules!
It's been several decades since my engineering courses, and my statistics courses are only slightly more recent. I'm sure there are terms for parts of the graph - Ordinate and Abscissa come to mind -that I ought to be using. The graphs are 'smoothed' anyway, representing physical measurements.

Fred Wagner