How is OCR used in Pega Robotic Automation?

From PegaWiki
Using OCR capabilities in your automations / This is the approved revision of this page, as well as being the most recent.
Jump to navigation Jump to search

How is OCR used in Pega Robotic Automation?

Description Describes how the OCR and PDF components are used when creating robotic automations.
Version as of 19.1
Application Robotic Process Automation
Capability/Industry Area Robotic Automation



Using OCR capabilities in your automation[edit]

Pega Robotic Automation includes the ABBYY® Fine Reader Engine (FRE) to provide the following optical character recognition (OCR) capabilities:

  • To convert images of application screens into text, so Pega Robot Studio can find key words and you can automate those applications based on the key words. This is known as screen-scraping. For more information, see Creating automations using the Screen OCR feature.
  • To convert image-based PDF files into text. After you use Robot Studio to convert the image-based PDF files into text using the Document OCR component, you can use the Robot Studio PDF component to extract data. You can use the resulting data in your automations. For more information, see Incorporating PDF files into your automation using Pega Robot Studio.

Note: Robot Studio does not provide direct access to the Fine Reader Engine from the automation surface. Instead, communication between Robot Studio and the ABBYY Fine Reader Engine is handled internally.

Screen scraping[edit]

The Screen OCR feature uses an image of the application to find the location of the controls that are used in the automation. The image is processed by the Fine Reader Engine, which is an OCR library that is included in Pega Robotic Automation. This library tells the system where key words are located on the screen, so you can use Robot Studio methods to click, double-click, get, and set text relative to those key words. In most cases, Robot Studio's interrogation capabilities provide a better way to automate applications, but screen-scraping is included to provide you the maximum flexibility when automating applications.

Working with PDF files[edit]

When text-based documents, like Microsoft Word documents, are converted into PDF files, the text is retained and can be directly extracted from the PDF file using Robot Studio's PDF component. This process does not require the use of OCR. You can find the PDF component in the Robot Studio Toolbox.

If the PDF file is actually a form that is filled out using Acrobat Reader or a similar tool, Robot Studio can extract the data from the form fields using the PDF component and produce excellent results. This process does not normally require OCR, but can include some OCR functionality, like checking for a signature. The PDF component uses the Fine Reader Engine to check signatures.

Some PDF files only contain an image of the text. This usually happens when the PDF was created from a fax or a scanned document. To extract the text from these image-based PDF files for use in an automation, use the Document OCR component. This component is also found in Robot Studio's Toolbox. The Document OCR component uses the Fine Reader Engine to process the image to create a text-based PDF file. You then extract the data from the new PDF file using the PDF component.

If the PDF file was printed and data was handwritten onto the form and then scanned or faxed, the results can vary widely. Handwritten text is not supported, but the Fine Reader Engine system can detect that a signature is present and interpret handwritten check marks.

No matter which method is required, incorporating a verification process may be an essential part of your automation strategy. The Reconcile method, available in the PDF component, makes it easy to incorporate the reconciliation process into your automation workflow.