How to Use Tesseract OCR and ImageMagic (both open source) to read values from a
n X-Window
screen such as ICC or SMDH:
Installation of tools:
Both programs have minimal impact on the system. Tesseract only updates PATH, T
EMP, and TMP environment variables and creates TESSDATA_PREFIX. It creates two
registry entries under HKEY_CURENT_USER/Software/Tesseract-OCR; Version (3.00),
and Install Dir (C:\Program Files\Tesseract-OCR).
Configuration folders:
doc
tessdata
configs
tessconfigs
training
Files in top level folder:
[Link]
[Link]
[Link]
[Link]
ImageMagic can be installed anywhere. The relocateable Windows version doesn't
rely on any
dll's or registry settings. Just use the full path when invoking [Link]
Windows Vista, XP, and NT Install:
In order to keep full path names reasonably short, I have installed it at C:\Ima
geMagick-6.6.7-4.
To keep it even shorter, and independant of release number, should probably just
use C:\ImageMagick\.
Most of what gets installed (i.e. copied) is not needed. In fact we could proba
bly survive with just [Link] copied into an existing folder in the Windows
PATH.
Download [Link]/pub/ImageMagick/binaries/ImageMagick-6.6.7-0-Q16-wi
[Link]
Execute it (or "open" it from your browser) to start the installation.
Once ImageMagick is installed, Select Start->Programs->Command Prompt.
In the Command Prompt window type
convert logo: [Link]
imdisplay [Link]
// Use ImageMagic to translate blue and dark blue to black
C:\Program Files\Tesseract-OCR>\ImageMagick-6.6.7-4\convert IOM\[Link] -f
ill black -opaque blue -opaque "rgb(0,0,153)" IOM\COMPOUNDS_BWY.tif {ENTER}
// if necessary, use ImageMagic to translate yellow to white, but so far recogni
tion
// has been 100% with yellow or white
C:\Program Files\Tesseract-OCR>\ImageMagick-6.6.7-4\convert IOM\COMPOUNDS_BWY.ti
f -f
ill white -opaque "rgb(255,252,0)" IOM\COMPOUNDS_BW.tif {ENTER}
// Use Tesseract to find text in the enhanced image. Tess adds the .txt suffix
// You have to run in the Tesseract directory in order for the Leptonic (?) dll
to work.
// But it seems to handle data in subfolders so...
C:\Program Files\Tesseract-OCR>tesseract IOM/COMPOUNDS_BW.tif IOM/COMPOUNDS {ENT
ER}
Tesseract Open Source OCR Engine with Leptonica
C:\Program Files\Tesseract-OCR>type IOM\[Link] {ENTER}
H21FCP_STA
H21FCP_ECB
H21_FBM232
C:\Program Files\Tesseract-OCR>
-----------------------------------------------------------
Convert any box with Cyan text to white on black:
// convert cyan to white
convert [Link] -fill white -opaque "rgb(0,160,158)" FOO_W.tif
//convert menu gray and dark blue to black
convert FOO_W.tif -fill black -opaque "rgb(0,0,153)" -opaque "rgb(114,114,114)"
FOO_BW.tif
Screen capture using AutoIT _ScreenCapture_Capture function requires GDI+: GDI+
requires a redistributable for applications that run on the Microsoft Windows NT
4.0 SP6, Windows 2000, Windows 98, and Windows Me operating systems.
