seglink 环境配置

代码地址：https://github.com/dengdan/seglink 
1、安装opencv3.1.0 
参考http://blog.csdn.net/u011440558/article/details/78358447

2、创建虚拟环境seg-env，安装tensorflow

sudo apt-get install python-pip python-dev python-virtualenv
virtualenv --no-site-packages seg-env     #纯净版本

E: Could not get lock /var/lib/dpkg/lock - open (11: Resource temporarily unavailable)
E: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it?

sudo rm /var/lib/apt/lists/lock
sudo rm /var/cache/apt/archives/lock
sudo rm /var/lib/dpkg/lock

tensorflow安装http://blog.csdn.net/isuker/article/details/52459992

#打开虚拟环境
source seg-env/bin/activate
pip install -i https://pypi.mirrors.ustc.edu.cn/simple/ tensorflow_gpu-1.0.0-cp27-cp27mu-manylinux1_x86_64.whl

报错：

(seg-env)**@gpuserver1:~$ pip install -i https://pypi.mirrors.ustc.edu.cn/simple/ tensorflow_gpu-1.0.0-cp27-cp27mu-manylinux1_x86_64.whl
tensorflow_gpu-1.0.0-cp27-cp27mu-manylinux1_x86_64.whl is not a supported wheel on this platform.
Storing debug log for failure in /home/**.pip/pip.log

python -m pip install --upgrade pip

再次pip install，OK..

3、运行 test_seglink.py 
3.1、下载并解压模型（seglink-384）和测试数据（ICDAR2015），代码地址中有提供。 
3.2、修改 test_seglink.py中的图像文件路径，模型路径

tf.app.flags.DEFINE_string('checkpoint_path', '/home/ly/seglink-master/seglink-384/model.ckpt-136750', 

tf.app.flags.DEFINE_string('dataset_dir', 
           util.io.get_absolute_path('/home/ly/seglink-master/datasets/ICDAR2015/Challenge4/ch4_test_images'), 
           'The directory where the dataset files are stored.')

修改./tf_extended/seglink.py 808行(否则会报错，opencv3没有cv.BoxPoints()函数):

# points = cv2.cv.BoxPoints(bbox)   #opencv2.4.9
points = cv2.boxPoints(bbox)       #opencv3.1.0

3.3、运行

#需要在代码中填写好参数
source seg-env/bin/activate
python test_seglink.py

或者：

 ./scripts/test.sh 0 /home/**/seglink-master/models/seglink/seglink-384/model.ckpt-136750  /home/**/seglink-master/datasets/ICDAR2015/ch4_test_images

报错： 
import util 
ImportError: No module named util

download the project pylib and add the src folder to your PYTHONPATH 
在linux virtualenv下添加PYTHONPATH 
http://blog.csdn.net/u011440558/article/details/78611829

ImportError: No module named matplotlib.pyplot
pip install -i https://pypi.mirrors.ustc.edu.cn/simple/ matplotlib
ImportError: No module named _tkinter, please install the python-tk package
sudo apt-get install python-tk

报错： 

pip install -i https://pypi.mirrors.ustc.edu.cn/simple/ enum

pip install --upgrade pip enum34

pip install -i https://pypi.mirrors.ustc.edu.cn/simple/ setproctitle

再运行

python test_seglink.py

OK…

注：将test_seglink.py输出.txt路径改下，否则会出现http://blog.csdn.net/u011440558/article/details/78606662所示报错

# logdir = util.io.join_path(FLAGS.checkpoint_path, 'test', FLAGS.dataset_name + '_' +FLAGS.dataset_split_name)
logdir = util.io.join_path('out-txt', FLAGS.dataset_name + '_' +FLAGS.dataset_split_name)

4、可视化检测结果 
输入：

python visualize_detection_result.py 
--image=/home/ly/seglink-master/datasets/ICDAR2015/Challenge4/ch4_test_images/ \
--det=/home/ly/seglink-master/seglink-384/model.ckpt-136750/test/icdar2015_test/model.ckpt-136750/seg_link_conf_th_0.900000_0.700000/txt \
--output=/home/ly/seglink-master/out

结果：