OCR results in dots
Hello;
I am building tests which include taking action and reading information from an application provided via remote desktop. Because it is coming across remote desktop, I have very little visibility into the application from the object browser. Specific to this question, I can't get the text objects appearing in pop-up windows.
As an aside, I have tried to enable TextObjects per these instructions, but the object browser is not listing the objects, nor can I access them through code. Therefore, I am trying to read the text using the OCR module.
I have attached an image of the window (Information) I am trying to OCR. I was previously trying to perform OCR against just the number (4300836 in this case), but that was consistently failing to recognize any text at all.
I have experimented extensively with trying different fonts, font sizes, and changing the active recognition set. My best guess is that it is failing due to the greyscale binarization - but I've experimented with some values between 0-255 and have not found any value that improves OCR performance.
For comparison, I find that when I run OCR on a screen that looks like the blue-grey screen with what appears to be MS Sans Serif text, the results are pretty good using just the default options.
Any ideas on how I can get better insight into improving the OCR performance? I'm also open to alternative solutions - I don't care if OCR factors into the solution or not - I'm just looking to capture that number from the screen.
Thanks in advance for any help.
I wasn't able to get OCR working on the target window. I was able to leverage the fact that the subscription number I was trying to read is generated by a database. So I wrote a function using ADO objects to run a query against the database and then I read the resulting recordset to get the value I needed.