The simplest screen OCR character recognition

The simplest screen OCR character recognition

Using the office2003 included with the Modi can be OCR text recognition, do not need sampling, easy to use, the recognition rate is very good.
The text that is appropriate to identify the comparison specification. The disadvantage is that the office2003 and Microsoft Document Imaging components must be installed.

This method is suitable for ordinary character recognition, and is not suitable for verification code identification.
Verification Code Identification please refer to Tutorial: http://www.yhhe.net/ape/book/fap/f2/ix.html

This program uses the COMX plug-in, please refer to the post: http://www.yhhe.net/bbs/dispbbs.asp?BoardID=4&ID=179&replyID=

Here is the demo source code:

Download Simulation Wizard: Http://www.yhhe.net/Fairy_Ape.exe
Open the Simulation wizard , paste the following code into the source editor, press F5 to run.

img = image.new ();--Create a Picture object
Img:capture (0,100,200,300,400);--grab screen, range x=100,y=200, wide =300, high =400
Img:save (_lasdir ... " Test.bmp "); --Save the picture to the script directory (aka _lasdir)

--Import COMx Plugin
Import ("Std");
Import2 ("COMx", "Http://www.yhhe.net/ape/import/comx/comx.dll");

--Create a Modi object (you must install the Microsoft Document Imaging component in office2003)
Mdoc = COMx. CreateObject ("MODI. Document ");
if (not mdoc) then
Win.messagebox ("Please install office2003 and Microsoft Document Imaging Components", "Screen OCR text recognition")
Return
False
End

--Import Pictures
Mdoc:create (_lasdir ... " Test.bmp ");
--OCR recognition, the parameters are the language ID, whether auto-lure, whether automatic stretching
MDOC:OCR (0x804,_false,_false);

Local mi = mdoc. Images (0);
--Quick access to all text
Win.messagebox (mi. Layout.text, "Mdoc. Images (0). Layout.text ");

--Get character details
Local word = mi. Layout.words (0)
Local str = "Id:". Word. Id.. "/r/n"

str = str.. "Line Id:". Word. LineId. "/r/n";
str = str.. "Region Id:". Word. RegionID. "/r/n";
str = str.. "Font Id:". Word. Fontid. "/r/n";
str = str.. "Recognition confidence:". Word. Recognitionconfidence. "/r/n";
str = str.. "Text:". Word. Text;

Win.messagebox (str, "Mdoc. Images (0). Layout.words (0) ")

The first parameter of the MDOC:OCR function specifies a language ID
The optional language IDs in the Simplified Chinese office are: Auto Select 0x800
English 9
Simplified Chinese 0x804

The optional language ID in Traditional Chinese office is: Auto Select 0x800
English 9
Traditional Chinese 0x404

The correct language ID can improve the recognition rate.
If MDOC:OCR does not find the text in the picture, it will error and terminate the simulation program.