NumCR Super Fast EASY NUMBER OCR

[sandeep]

need help , when I run the script with the example bmp it only return " . " or " .. " instead of numbers.

------------------------------------------

#include "NumCR.au3"

$test = _OCR(0, 0, 300, 300)

MsgBox(0, "Example", $test)

------------------------------------------

Is it working??

[ian.b]

Is it working??

Yes, it works just fine for me - I would like to thank pingpong24 for sharing.

The critical thing is to build the proper definition for used font/size,

some string manipulation may be needed to remove extra dots.

ian

[xlamik]

I little modified the script for NumCR.

In my version, I count the number of pixels in area to determine number (or another character). And to distinct the numbers with the same numbers of pixels (2 and 4, 6 and 9 in my case) I added some multiplicity factor to next row. It means that next row have bigger value than previous and using this method I was able to generate different patters to different number. Before usage you must modify the script to set the number of pixels for every digit you will use. Also different multiply strategies and values can be used to generate unique pattern.

expandcollapse popup

Func _OCR($Ax, $Ay, $Dx, $Dy)
    
    $width = $Dx - $Ax
    $length = $Dy - $Ay
    
    Dim $char = 0;
    Dim $sign = 1 
        
    For $xwidth = 0 To $width - 1
        For $xlength = 0 To $length - 1
            If PixelGetColor($Ax+$xwidth, $Ay+$xlength) = 0 Then
            $char += $sign ; or 1*$sign, 2*$sign. For unique pattern
            EndIf
        Next
        If $char <> 0 Then
            $sign += 1 ; increase the row multiply factor to avoid the same # of black pixels in different numbers
        EndIF
    Next
    
    Select
        Case $char = 152
                $result = 0
        Case $char = 80
                $result = 1
        Case $char = 120
                $result = 2 
        Case $char = 138
                $result = 3
        Case $char = 126
                $result = 4 
        Case $char = 130
                $result = 5
        Case $char = 131
                $result = 6 
        Case $char = 102
                $result = 7
        Case $char = 156
                $result = 8
        Case $char = 149
                $result = 9
    EndSelect
    Return $result 
EndFunc

[Root_]

I know this is an extreme gravedig. But I am working on a project and trying to get some OCR for numbers only -- something like this that I can customize to my own liking/text. I tried using tesseract and it kept messing up changing digits to random characters.

 

From what it looks like in this script i am to specify the text color on line 24, which happens to be green (4dc92b) then count the green pixels of each column from top to bottom -- here is an image of what I envision I am supposed to do, please correct me if im wrong. The numbers in the picture represent each vertical pixel. meaning that my if statement to catch that number should look like the code below, right? 

 

If $row[$xwidth] = 5 And $row[$xwidth+1] = 3 And $row[$xwidth+2] = 3 And $row[$xwidth+3] = 3 And $row[$xwidth+4] = 5 Then
            $result &= "8"
        EndIf

 

I am willing to pay a few bucks to anyone that can get this text to recognize for me using this lightweight OCR.

[JLogan3o13]

What is the app that you're pulling the numbers from (desktop app, web app, etc.)? Are you not able to get any info with the AutoIt Window Info Tool, that you have to resort to OCR?