cubieboard ubuntu 下 PocketSphinx 实时语音识别

gamexg · 发表于 2013-12-4 10:39:32

本帖最后由 gamexg 于 2013-12-4 10:49 编辑

我是在 arm cubieboard 下做的。

对了，在实际做之前先看一下 http://www.chenwang.net/2013/11/ ... ef%bc%881%ef%bc%89/ 文章里面有中文语音识别模型，下面的那个例子把中文模型注释掉了，如果需要的话可以下载下来和 03.py 放到一个目录。

sudo apt-get install gstreamer0.10-alsa gstreamer0.10-tools libgstreamer0.10-0 libgstreamer0.10-dev python-gst0.10 python-gst0.10-dev libgstreamermm-0.10-2 libgstreamermm-0.10-2 make gcc g++ bison libbison-dev libasound2-dev libasound2-python python-alsaaudio python-pyalsa python-gtk2 soundconverter vim
sudo apt-get install gstreamer0.10-alsa gstreamer0.10-tools libgstreamer0.10-0 libgstreamer0.10-dev python-gst0.10 python-gst0.10-dev libgstreamermm-0.10-2 libgstreamermm-0.10-2
sudo apt-get install make gcc g++
sudo apt-get install bison
sudo apt-get install libbison-dev
sudo apt-get install libasound2-dev libasound2-python python-alsaaudio python-pyalsa
sudo apt-get install python-gtk2
sudo apt-get install soundconverter
wget http://downloads.sourceforge.net/project/cmusphinx/pocketsphinx/0.8/pocketsphinx-0.8.tar.gz?r=http%3A%2F%2Fsourceforge.net%2Fprojects%2Fcmusphinx%2Ffiles%2Fpocketsphinx%2F0.8%2F&ts=1385189168&use_mirror=softlayer-dal
mv pocketsphinx-0.8.tar.gz\?r\=http\:%2F%2Fsourceforge.net%2Fprojects%2Fcmusphinx%2Ffiles%2Fpocketsphinx%2F0.8%2F pocketsphinx-0.8.tar.gz
wget http://downloads.sourceforge.net/project/cmusphinx/sphinxbase/0.8/sphinxbase-0.8.tar.gz?r=http%3A%2F%2Fsourceforge.net%2Fprojects%2Fcmusphinx%2Ffiles%2Fsphinxbase%2F0.8%2F&ts=1385189298&use_mirror=softlayer-dal
mv sphinxbase-0.8.tar.gz\?r\=http\:%2F%2Fsourceforge.net%2Fprojects%2Fcmusphinx%2Ffiles%2Fsphinxbase%2F0.8%2F sphinxbase-0.8.tar.gz
tar xzf sphinxbase-0.8.tar.gz
cd sphinxbase-0.8
./configure
make
sudo make install
cd ..
export LD_LIBRARY_PATH=/usr/local/lib
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig
tar xzf pocketsphinx-0.8.tar.gz
cd pocketsphinx-0.8
./configure
make
sudo make install
pocketsphinx_continuous
echo "export GST_PLUGIN_PATH=/usr/local/lib/gstreamer-0.10/" >> ~/.bashrc
echo "export LD_LIBRARY_PATH=/usr/local/lib/" >> ~/.bashrc
export GST_PLUGIN_PATH=/usr/local/lib/gstreamer-0.10/
linaro@cubieboard:~/yysb$ cat 03.py
import pygtk,gtk,gobject
gobject.threads_init()
import gst
def result(asr, text, uttid): print text
pipeline=gst.parse_launch('alsasrc device=1 ! audioconvert ! audioresample ! vader name=vad auto-threshold=true ! pocketsphinx name=asr ! appsink sync=false name=appsink')
#pipeline=gst.parse_launch('alsasrc ! audioconvert ! audioresample ! vader name=vad auto-threshold=true ! pocketsphinx name=asr ! appsink sync=false name=appsink')
#pipeline=gst.parse_launch('pulsesrc ! audioconvert ! audioresample ! vader name=vad auto-threshold=true ! pocketsphinx name=asr ! appsink sync=false name=appsink')
#pipeline=gst.parse_launch('pulsesrc ! tee name=t ! queue ! audioconvert ! audioresample ! vader name=vad auto-threshold=true ! pocketsphinx name=asr ! appsink sync=false name=appsink t. queue ! audioconvert ! audioresample ! wavenc ! filesink location=o.wav')
asr=pipeline.get_by_name('asr')
asr.connect('result', result)
#asr.set_property('hmm', 'zh_broadcastnews_ptm256_8000')
#asr.set_property('lm', '1.lm')
#asr.set_property('dict', '1.dic')
asr.set_property('configured', True)
pipeline.set_state(gst.STATE_PLAYING)
gtk.main()
linaro@cubieboard:~/yysb$ python 03.py

复制代码

好了，现在可以测试了。

识别准确率固定几条命令的话识别率还行，但是会把其他的不相干的话识别成命令。现在正在尝试本地+在线语音识别，但是板子的音源输入只允许单个程序独占很烦人，现在正在考虑解决办法。

配置声卡：http://www.chenwang.net/2013/11/ ... %e9%85%8d%e7%bd%ae/

原文地址：http://www.chenwang.net/2013/11/23/ubuntu-%E4%B8%8B%E5%AE%89%E8%A3%85-pocketsphinx-with-gstreamer-and-python/

		自动登录	找回密码
密码			立即注册