Answered

Scanned PDF to OCR: Failed to extract text with OCR ---> System.ArgumentException: Parameter is not valid

(0) Share

Report

Posted on by JotEss

Hello experts,

I'm a beginner at using PAD and I'm trying to get text via OCR out of scanned PDF files. The error message I get - independent of the choice of OCR module is:

Microsoft.Flow.RPA.Desktop.Modules.SDK.ActionException: Failed to extract text with OCR ---> System.ArgumentException: Parameter is not valid. at System.Drawing.Bitmap..ctor(String filename)
at Microsoft.Flow.RPA.Desktop.Modules.OCR.Utilities.Utilities.GetImageForOCR(OCRSource source, SourceScanMode sourceScanMode, Nullable`1 scanRegionX1, Nullable`1 scanRegionY1, Nullable`1 scanRegionX2, Nullable`1 scanRegionY2, IEnumerable`1 imagesToFind, Int32 tolerance, Boolean waitForImage, Boolean timeoutSet, Nullable`1 timeout, Nullable`1 searchRegionImageX1, Nullable`1 searchRegionImageY1, Nullable`1 searchRegionImageX2, Nullable`1 searchRegionImageY2, Action suspendSecureScreen, Action restoreSecureScreen, String imageFilepath, IImageFinder imageFinder)
bei Microsoft.Flow.RPA.Desktop.Modules.OCR.Actions.ExtractTextWithOCRBase.Execute(ActionContext context)
--- Ende der internen Ausnahmestapelüberwachung ---
bei Microsoft.Flow.RPA.Desktop.Modules.OCR.Actions.ExtractTextWithOCRBase.Execute(ActionContext context)
bei Microsoft.Flow.RPA.Desktop.Robin.Engine.Execution.ActionRunner.Run(IActionStatement statement, Dictionary`2 inputArguments, Dictionary`2 outputArguments)

(As I'm working on a German Windows machine I tried to translate at least the beginning of the message back to English but the wording might not match exactly. Unfortunately there seems to be no option to switch languages of PAD.)

As far as I understand the message, PAD fails extracting a getting the image out of the file.

I have tried with a screenshot of the pdf and fed that PNG file to the OCR module, this worked. So could someone tell me what could have gone wrong? Can there be differences in the way a scanned PDF is made up so that PAD might be able or not to extract something?

There is another thread on this forum, "How to extract text from PDF using PAD?", that looks like the problem could be the same but I couldn't figure out what the solution post means.

Best regards, Joachim

Categories:

Power Automate Desktop

I have the same question (0)

All responses (4)

Answers (2)

Sort by

Verified answer

VJR 7,635 on at

Like
a
(0)

Report
Copy link

Link copied!

Hi @JotEss

If I understood your post correctly I hope my response makes sense to you.

I was not able to directly extract the text from the images inside a PDF. Looks like the PAD action does not support this yet.

I had to first use Extract images from PDF into a folder and then while looping through the images of this folder, passed each of the images to Extract text with OCR.

Was this reply helpful? Yes No
JotEss 11 on at

Like
a
(0)

Report
Copy link

Link copied!
Hi @VJR,
yes, I guess you understood. I suspected I would have to do something like this and I have already tried "Extract Images from PDF" (or similar wording) (sorry I forgot to mention), but it also resulted in an error. My try to translate:
The transformation specified is invalid.: Microsoft.Flow.RPA.Desktop.Modules.SDK.ActionException: Error extracting images from „D:\Documents\Scan\Automate\scan.pdf“ ---> System.InvalidCastException: The transformation specified is invalid.
at System.Linq.Enumerable.<CastIterator>d__97`1.MoveNext() ...
Original:
Die angegebene Umwandlung ist ungültig.: Microsoft.Flow.RPA.Desktop.Modules.SDK.ActionException: Fehler beim Extrahieren von Bildern aus „D:\Documents\Scan\Automate\scan.pdf“ ---> System.InvalidCastException: Die angegebene Umwandlung ist ungültig.
bei System.Linq.Enumerable.<CastIterator>d__97`1.MoveNext()
bei System.Linq.Buffer`1..ctor(IEnumerable`1 source)
bei System.Linq.Enumerable.ToArray[TSource](IEnumerable`1 source)
bei Microsoft.Flow.RPA.Desktop.Modules.PDF.Actions.PDFium.PdfBitmap.GuessPallete(Byte[] indices)
bei Microsoft.Flow.RPA.Desktop.Modules.PDF.Actions.PDFium.PdfBitmap.ToBitmap()
bei Microsoft.Flow.RPA.Desktop.Modules.PDF.Actions.ExtractImagesFromPDFAction.Execute(ActionContext context)
--- Ende der internen Ausnahmestapelüberwachung ---
bei Microsoft.Flow.RPA.Desktop.Modules.PDF.Actions.ExtractImagesFromPDFAction.Execute(ActionContext context)
bei Microsoft.Flow.RPA.Desktop.Robin.Engine.Execution.ActionRunner.Run(IActionStatement statement, Dictionary`2 inputArguments, Dictionary`2 outputArguments)
No clue what is wrong. I found "move next" and thought the algorithm is working on several items in the pdf, so I deleted all pages except the first and deleted an annotation that was contained. But the error persists, even with this cleaned pdf.
Best regards
Joachim

Was this reply helpful? Yes No
Verified answer

JotEss 11 on at

Like
a
(0)

Report
Copy link

Link copied!

I believe I found the cause myself. it was greyscale vs. 1-bit b/w.
I thought I have to try with a pdf that I am sure I have not yet tampered with, i.e. processed with some editor, before. To make sure it is not some way of edit that makes the pdfs unprocessable. So I happened to grab a scanned page of handwritten notes and it worked (at least extracting the image). And then I noticed, that this was a greyscale scan. Now I tried systematically: Greyscale works, 1-bit black-an-white doesn't!
This is a pity, because I would prefer 600 dpi b/w to 200 dpi greyscale as it is sharper while only a fraction of the file size.
Why the h*** did they make this action only work on some sorts of pdf?
Now I have to feed the images to OCR one by one. Will take me a while as I'm a near complete novice.

Was this reply helpful? Yes No
sdeokule 31 on at

Like
a
(0)

Report
Copy link

Link copied!

Hello VJR

I also need same thing extract data from pdf but in Power automate desktop this action is not available first we need to extract images from pdf. This is unnecessary extra step to follow.

Was this reply helpful? Yes No