我的ocr训练记录

最新推荐文章于 2023-11-20 20:28:53 发布

阳十三

最新推荐文章于 2023-11-20 20:28:53 发布

阅读量529

点赞数

分类专栏： ocr

本文为wdful原创文章，未经wdful允许不得转载。

本文链接：https://blog.csdn.net/qq_28114645/article/details/81557178

版权

ocr 专栏收录该内容

1 篇文章 0 订阅

订阅专栏

ubuntu开发训练ocr记录

install tesseract

install dependency libs


    sudo apt-get install g++ # or clang++ (presumably)
    sudo apt-get install autoconf automake libtool
    sudo apt-get install autoconf-archive
    sudo apt-get install pkg-config
    sudo apt-get install libpng12-dev
    sudo apt-get install libjpeg8-dev
    sudo apt-get install libtiff5-dev
    sudo apt-get install zlib1g-dev

    sudo apt-get install libicu-dev
    sudo apt-get install libpango1.0-dev
    sudo apt-get install libcairo2-dev

install tesseract and training tools

    wget https://codeload.github.com/tesseract-ocr/tesseract/zip/3.04
    ./configure
    make && make install 
    make training && make training-install

如果报错checking for leptonica... configure: error: leptonica not found:
wget http://www.leptonica.org/source/leptonica-1.72.tar.gz tar xvzf leptonica-1.72.tar.gz cd leptonica-1.72/ ./configure make && make install
It is also useful, but not required, to build ScrollView.jar

    make ScrollView.jar
    export SCROLLVIEW_PATH=$PWD/java

测试识别:tesseract test.png output_1 –l eng 语言包放置地址:/usr/local/share/tessdata

training

merge tif

jTessBoxEditor -> tools->merge tiff….

makebox

tesseract test.exp0.tif test.exp0 -l eng -psm 7 batch.nochop makebox
我使用 tesseract lang.font.exp0.tif lang.font.exp0 -l chi_sim -psm 3 batch.nochop makebox

change box

use jTessBoxEditor

make font_properties

新建一个名为font_properties的文本文件（注意该文件没有扩展名），内容为字体名fontyp，后面带5个0，分别代表字体的粗体、斜体等属性，这里全部是0
命令建立font_properties:echo font 0 0 0 0 0 >font_properties

make train file

tesseract lang.font.exp0.tif lang.font.exp0 -l chi_sim -psm 3 nobatch box.train
unicharset_extractor lang.font.exp0.box
shapeclustering -F font_properties -U unicharset -O lang.unicharset lang.font.exp0.tr
mftraining -F font_properties -U unicharset -O lang.unicharset lang.font.exp0.tr
cntraining lang.font.exp0.tr

重命名

mv normproto font.normproto
mv inttemp font.inttemp
mv pffmtable font.pffmtable
mv unicharset font.unicharset
mv shapetable font.shapetable
combine_tessdata font.