5 posts in this topic
tesseract doesnt detect the easiest image *image to text*
i have this picture here *attached* and this script here:
$ImageToReadPath = @MyDocumentsDir & "\GDIPlus_Image2.jpg" $ResultTextPath = @MyDocumentsDir & "\Result" $OutPutPath = $ResultTextPath & ".txt" $TesseractExePath = @MyDocumentsDir & "\Tesseract.exe" ShellExecuteWait($TesseractExePath, '"' & $ImageToReadPath & '" "' & $ResultTextPath & '"', "", "", @SW_HIDE) If @error Then Exit MsgBox(0, "Error", @error) EndIf MsgBox(0, "Result", FileRead($OutPutPath)) FileDelete($OutPutPath)
but tesseract doesnt recognized the correct word... and gives me trash back...
this is the image >>
and the result was >> "samm"
the image was an normal jpg and generated with this code here:
_ScreenCapture_Capture(@MyDocumentsDir & "\GDIPlus_Image2.jpg", 712,268,853,284)
Could anybody give me a hint what i can do better to get this easy image to text?
thank u very much!!!
Edit: i also tried to capture the screen as bmp with a higher resolution... nothing changed...
_ScreenCapture_SetBMPFormat(4) _ScreenCapture_Capture(@MyDocumentsDir & "\GDIPlus_Image.bmp", 712,279,853,295)
Image quality is bad compared to paint
#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_Icon=car.ico #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** #include <GDIPlus.au3> #include <File.au3> #include <Array.au3> #include <ColorConstants.au3> #include <GUIConstantsEx.au3> #include <WindowsConstants.au3> #include <ProgressConstants.au3> ; Declare array Dim $Images ; Gets all JPG files in the current directory (@ScriptDir). Local $search = FileFindFirstFile("*.jpg") ; Check if the search was successful If $search = -1 Then MsgBox(0, "Error", "No JPG files could be found.") Exit EndIf ; Resize array While 1 If IsArray($Images) Then Local $Bound = UBound($Images) ReDim $Images[$Bound+1] EndIf $Images[$Bound] = FileFindNextFile($search) If @error Then ExitLoop WEnd ; Close the search handle FileClose($search) ; Create directory "resized" if not there yet $nymappe = InputBox("Mappe / Bil Navn", "Mappe / Bil Navn") If NOT FileExists(@ScriptDir & "\" & $nymappe & "\") Then DirCreate(@ScriptDir & "\" & $nymappe & "\") EndIf ; Loop for JPGs - gets dimension of JPG and calls resize function to resize to 50% width and 50% height For $i = 1 to Ubound($Images)-1 If $Images[$i] <> "" AND FileExists(@ScriptDir & "\" & $Images[$i]) Then Local $ImagePath = @ScriptDir & "\" & $Images[$i] _GDIPlus_Startup() Local $hImage = _GDIPlus_ImageLoadFromFile($ImagePath) Local $ImageWidth = _GDIPlus_ImageGetWidth($hImage) Local $ImageHeight = _GDIPlus_ImageGetHeight($hImage) _GDIPlus_ImageDispose($hImage) _GDIPlus_Shutdown() ;MsgBox(0,"DEBUG", $ImageWidth & " x " & $ImageHeight) Local $NewImageWidth = ($ImageWidth / 100) * 15 Local $NewImageHeight = ($ImageHeight / 100) * 15 ;MsgBox(0,"DEBUG: " & $i,$Images[$i]) _ImageResize(@ScriptDir & "\" & $Images[$i], @ScriptDir & "\" & $nymappe & "\" & $Images[$i], $NewImageWidth, $NewImageHeight) EndIf Next ; Resize function Func _ImageResize($sInImage, $sOutImage, $iW, $iH) Local $hWnd, $hDC, $hBMP, $hImage1, $hImage2, $hGraphic, $CLSID, $i = 0 ;OutFile path, to use later on. Local $sOP = StringLeft($sOutImage, StringInStr($sOutImage, "\", 0, -1)) ;OutFile name, to use later on. Local $sOF = StringMid($sOutImage, StringInStr($sOutImage, "\", 0, -1) + 1) ;OutFile extension , to use for the encoder later on. Local $Ext = StringUpper(StringMid($sOutImage, StringInStr($sOutImage, ".", 0, -1) + 1)) ; Win api to create blank bitmap at the width and height to put your resized image on. $hWnd = _WinAPI_GetDesktopWindow() $hDC = _WinAPI_GetDC($hWnd) $hBMP = _WinAPI_CreateCompatibleBitmap($hDC, $iW, $iH) _WinAPI_ReleaseDC($hWnd, $hDC) ;Start GDIPlus _GDIPlus_Startup() ;Get the handle of blank bitmap you created above as an image $hImage1 = _GDIPlus_BitmapCreateFromHBITMAP ($hBMP) ;Load the image you want to resize. $hImage2 = _GDIPlus_ImageLoadFromFile($sInImage) ;Get the graphic context of the blank bitmap $hGraphic = _GDIPlus_ImageGetGraphicsContext ($hImage1) ;Draw the loaded image onto the blank bitmap at the size you want _GDIPLus_GraphicsDrawImageRect($hGraphic, $hImage2, 0, 0, $iW, $iH) ;Get the encoder of to save the resized image in the format you want. $CLSID = _GDIPlus_EncodersGetCLSID($Ext) ;Generate a number for out file that doesn't already exist, so you don't overwrite an existing image. Do $i += 1 Until (Not FileExists($sOP & $i & "_" & $sOF)) ;Prefix the number to the begining of the output filename $sOutImage = $sOP & $i & "_" & $sOF ;Save the new resized image. _GDIPlus_ImageSaveToFileEx($hImage1, $sOutImage, $CLSID) ;Clean up and shutdown GDIPlus. _GDIPlus_ImageDispose($hImage1) _GDIPlus_ImageDispose($hImage2) _GDIPlus_GraphicsDispose ($hGraphic) _WinAPI_DeleteObject($hBMP) _GDIPlus_Shutdown() EndFunc Quality gets quite bad compared to using Paint / Photoshop when resizing with GDIPlus
Any idea how to make the quality better?
Thanks in advance
yFilter UDF beta
A collection of image filter effects usable with AutoIt!
IMPORTANT: You are not allowed to sell this code or just parts of it in a commercial project or modify it and distribute it with a different name!
Distributing copies of this UDF incl. _GDIPlus_BitmapApplyFilter.dll in compiled format (exe) must be free of any fee!
More information can be found in the forum thread!
Getting image dimensions
Hi there. Sorry to bother again. I don't get the tutorials and examples...
Let's say I want to display an image "E:\img\123.jpg" in a GUI. Usually I'd do that by using GUICtrlCreatePic($path, $x, $y, $width, $height).
Unfortunately, the width and height are changing with every picture and they are unknown as well...
First of all, how I can get the width and height of a given picture?
And is there any way that doesn't require me to constantly delete and create that window from scratch?
A new image search solution
Im a new AutoIt user, im from Spain and i would like to share here my own image search detection method.
What's about it?
It's not perfect, but with it you don't need to use any DLL file. You don't need to distribute the image file with your final AU3 Script, or Exe if compiled. The image size does not affect to the detection speed. It's bug free (i believe).
You can find here as an attach file, a distro package with all you need, and it includes a simple how-to guide.
Im going to paste here the how-to guide:
How to use this AutoIt image detector:
Install JRE (Java Runtime Environment), without this you can't execute Java Applications (.jar files). Open the Java tool named AISD with double click on it, or "java -jar aisd.jar" in the cmd. In AISD, open the image you want to detect in your script. Remember: the upper left pixel of the image, MUST be a not common color in the background if you want a fast detection. Then you will get a script file if you want, or only a clipboard copied version, full ready to paste into your main script because this one comes with extra help codes. Anyway, the generated code is the $sign definition of the image (calm down, point 3 explains what is the $sign and it's easy).
Paste the mentioned code in your main script, and try to follow the sample_usage script. If you know how to write au3 codes, it should be very easy. In general terms, all you need is to include image_search.au3, and then you can use the func image_search($sign, $width, $height). You need also the get_image_sign() function, that calculates the first parameter for image_search(). The get_image_sign() function is the one you have generated at point 1.
image_search() returns the center point of the target image. The func needs the $sign from the image, width and height. This $sign is the "footprint" of the image, in the form of an array of 9 integers. In most of the cases the $sign should be enough to identify the image, this is useful because using the $sign instead of the real image let you get a small script and a fast detection. You can also delete the image once you get the $sign. This $sign has always the same size (those 9 integers, it does not matter if the original image is small or big). How can you define the $sign for a certain image?, you don't need to generate the $sign, AISD does (point 1). Follow the sample_usage script and you will get it.
Files in this distro package:
* AISD.jar: The already explained Java app.
* image_search.au3: The library or module with the resources that lets you detect images.
* sample_usage.au3: The mentioned sample_usage script. It contains the smallest example, but i think it should be ok for learning purposes.
* sample_usage.exe: The sample_usage.au3 script compiled (au3 -> .exe). You can execute this to see the image_search() func working.
* sample_target.png: The image to be detected in the sample_usage. You can delete this file and the script will be fine.
* OffsetCalc.jar: It's only a Java tool to calculate the offset to transform an XY point into another one. It's useful if you want to use apply_offset function (from image_search.au3). Both things are only necessary when it's hard to find a version of the target image with upper left pixel in the way mentioned. As you can imagine, the idea is to detect another image and then apply the offset to get the real coordinates of the original target image.