PDF drawings can contain both real text (including actual letters & fonts) and elements that are shaped like text (not including letters or fonts).
The difference between the two is important because if drawings don’t have real text, in certain circumstances, that can increase the chances of false positives.
Below is a symbol which includes real text.
A symbol that includes real text has the following additional metadata that Countfire’s automatic counting algorithm can use:
Actual letters (the letter ‘D’ in the above example).
A font type.
A font size.
Because of this additional metadata information, real text is preferable.
Text like elements
Below is a symbol which includes text like elements (and is not actually proper text).
A symbol that doesn’t include real text is just one or more elements that are shaped to look like text.
This means it doesn’t have additional metadata like letters, font types and font sizes available to Countfire, and automatic counting is done purely on the shape of the elements alone.
How do I know whether my drawings have real text or text-like elements?
The key difference is that real text has crisp edges, whereas text-like elements have a certain jaggedness to them.
Also, letters and words within PDFs that have real text will often be selectable as text when viewed within a PDF viewer such as Adobe Reader.
In the clip above, recorded in Preview for Mac, the text is selectable which confirms this PDF file contains real text as opposed to text like elements.
Text like elements and false positives.
Because text like elements don’t include the additional metadata that real text does, Countfire’s automatic counting algorithm can’t use that additional information to help differentiate between symbols.
In most cases this isn’t an issue, however, with certain symbols, it can increase the chances of running into false positives.
However, it’s important to note that given enough information (selections), Countfire can always tell the difference between two symbols.
In the clip above, an initial selection of a Type D symbol is made. Because the text within the Type D symbol is only a text like element (not real text), this selection causes the Type B and BE symbols to be misidentified (false positives).
These false positives happen for two reasons:
From the software’s perspective, The Type B, BE & D symbols are similar
Countfire hasn’t yet been shown what a Type B and Type BE symbol is
These false positives are easily corrected by making additional selections to give Countfire more information to work with.
Additionally, if you didn’t pick up these false positives while making selections within Countfire, they’re easy to spot when reviewing the check sheets at the end of the project.
Real text and false positives
Because real text has additional metadata information (letters, font types, and font sizes) there’s is a much lower chance of false positives.
Using the Type B, BE & D example above, if those symbols had real text instead, when the initial selection of the Type D symbol was made, because Countfire categorically knows that symbol contains the letter ‘D’, and the other symbols contain the letters ‘B’, there would be no false positives.
I don’t have real text on my drawings, how can I change that?
Whether you have real text in your PDF files or not is decided when the PDF is created and is therefore often out of your control.
There are however a few routes worth trying:
If you have access to the original AutoCAD .DWG files, you can create fresh PDF files from them and try to preserve the real text during that process.
If your PDF drawings have been extracted from a multiple page PDF file, depending on the software used to complete that extraction, it may have caused any real text information to be lost. In this case, it’s worth trying a different software package to complete the extraction.
If you have drawings that have lost their real text information and you have access to their original .DWG files or multiple page PDFs, and aren’t able to preserve that real text information, please contact us for help.
How to: Group items.