OCR and image automation

uiautomation
ocr
activities

#1

Hi,
I was trying out the example of using OCR and image automation. https://studio.uipath.com/docs/example-of-using-ocr-and-image-automation

I need to scrape the $500 value from the image in the pdf.
I have attached my xaml file. Could someone please point out where I went wrong as I am not getting the required value. I have attached the xaml and pdf along with this. Your help will be much appreciated.
ocr xaml.zip (242.3 KB)

Thanks in advance


#2

I am also looking for the same. I hope someone helps. Also I am looking for solution for scrap data from image. Which OCR engine gibves best accurate result? also how to install ABBY, tesseract ocr engines? @ddpadil
@ovi . Expecting help guys


#3

You should use ‘Read PDF with OCR’. This returns the full text of the PDF file. In this case your desired output is on the last line. I used some simple string manipulation to extract it. If the PDF files are always like this, then this will do. If they are unstructured, you’re gonna need different manipulation techniques to do the job.
Main.xaml (8.2 KB)


#4

Thanks for the reply. I tried using “Read PDF with OCR”. But its not accurate, everytime I am getting different values.


#5

If the required data was in between it would have been difficult. So is there any other way we can scrap the required data other than using string manipulation technique. I mean like using “set clipping region” (the one showed in the example).

Thanks in advance


#6

You could try setting the zoom to a certain percentage to get the $500 inside the screen (instead of using page down) and then set a clipping region, but that’s gonna be pretty unstable I reckon (as the text will always need to be exactly inside your clipping region). I still think that using Read PDF with OCR will get you the most stable results. Did you try both the Microsoft and Google OCR engine? And have you tried different scales? For me the microsoft ocr engine with a scale of 1 already did the trick.