Tclip 用于图片裁剪，有以下特点：

1. 能进行人脸识别。图片中有人脸，将自动视为人脸区域为重要区域，将不会被裁剪掉。
2. 自动识别其它重要区域。如果图片中未识别出人脸，则会根据特征分布计算出重区域。
   总而言之，自动识别图片中的重要区域，并且在图片裁剪时保留重要区域。

源码地址：https://github.com/exinnet/tclip

安装opencv

根据github上的说明，在CentOS上安装没有问题，但在我的MacOS上就挂了。

首先遇到的问题就是，opencv安装不过去。还好，上github下载最新的opencv-2.4.11 安装成功。

下载地址：https://github.com/Itseez/opencv/releases

使用当前最新版本 OpenCV 2.4.11

安装依赖

在安装opencv之前，先安装一些依赖包：

brew install gtk+  pkgconfig libpng zlib libjpeg libtiff cmake

提示： brew的安装使用请参考 http://brew.sh/

安装opencv

开始安装 opencv：

tar zxf opencv-2.4.11.tar.gz
cd opencv-2.4.11
cmake CMakeLists.txt 
make && make install 

安装php tclip

先下载：https://github.com/exinnet/tclip/archive/master.zip

然后继续：

unzip tclip-master.zip 
cd tclip-master/php_ext
phpize
./configure

如果不出意外，到这一步，应该就挂掉了，提示：

checking for opencv.pc file in default path... found in /usr/lib/pkgconfig
found in /usr/local/lib/pkgconfig
configure: error: no result from pkg-config opencv --libs --cflags opencv

在 Tclip 作者的页面上 http://www.bo56.com/tclip%E4%BA%BA%E8%84%B8%E8%AF%86%E5%88%AB%E5%9B%BE%E7%89%87%E8%A3%81%E5%89%AA/#download

评论中，也有网友遇到了类似的问题，提出修改办法为：

将 config.m4 中的 判断语句test ${i:${#i}-3} = “.so”，改为 test ${i:${#i}-6} = ".dylib”，并尝试将46行的.so 也改称 .dylib 

依然提示no result的错误～

再翻翻config.m4的代码和执行 pkg-config opencv –libs –cflags opencv的输出结果：

-I/usr/local/include/opencv -I/usr/local/include -L/usr/local/lib -lopencv_calib3d -lopencv_contrib -lopencv_core -lopencv_features2d -lopencv_flann -lopencv_gpu -lopencv_highgui -lopencv_imgproc -lopencv_legacy -lopencv_ml -lopencv_nonfree -lopencv_objdetect -lopencv_ocl -lopencv_photo -lopencv_stitching -lopencv_superres -lopencv_ts -lopencv_video -lopencv_videostab

感觉怎么对不上，于是跑到Linux安装成功的服务器执行了一下，输出结果如下：

-I/usr/local/include/opencv -I/usr/local/include /usr/local/lib/libopencv_calib3d.so /usr/local/lib/libopencv_contrib.so /usr/local/lib/libopencv_core.so /usr/local/lib/libopencv_features2d.so /usr/local/lib/libopencv_flann.so /usr/local/lib/libopencv_gpu.so /usr/local/lib/libopencv_highgui.so /usr/local/lib/libopencv_imgproc.so /usr/local/lib/libopencv_legacy.so /usr/local/lib/libopencv_ml.so /usr/local/lib/libopencv_nonfree.so /usr/local/lib/libopencv_objdetect.so /usr/local/lib/libopencv_photo.so /usr/local/lib/libopencv_stitching.so /usr/local/lib/libopencv_ts.so /usr/local/lib/libopencv_video.so /usr/local/lib/libopencv_videostab.so

再看他的判断代码：

OPENCV_FLAGS="`pkg-config opencv --libs --cflags opencv`"
for i in $OPENCV_FLAGS;do
if test ${i:0:2} = "-I" ;then
PHP_ADD_INCLUDE(${i:2})
elif test ${i:${#i}-3} = ".so" ;then
dir_name=`dirname $i`
file_name=${i/$dir_name/}
file_name=${file_name/\/lib/}
file_name=${file_name/.so/}
PHP_ADD_LIBRARY_WITH_PATH($file_name,$dir_name,TCLIP_SHARED_LIBADD)
else
AC_MSG_ERROR([no result from pkg-config opencv --libs --cflags opencv])
fi
done

我立刻就明白了，Linux上输出的都是具体的.so路径，MacOS上都是相对的路径，而config.m4中是根据具体路径和扩展名判断的，明白了问题，解决就简单了。

将 pkg-config opencv –libs –cflags opencv 的执行结果修改为具体的路径，并替换到config.m4中：

OPENCV_FLAGS="-I/usr/local/include/opencv -I/usr/local/include /usr/local/lib/libopencv_calib3d.dylib /usr/local/lib/libopencv_contrib.dylib /usr/local/lib/libopencv_core.dylib /usr/local/lib/libopencv_features2d.dylib /usr/local/lib/libopencv_flann.dylib /usr/local/lib/libopencv_gpu.dylib /usr/local/lib/libopencv_highgui.dylib /usr/local/lib/libopencv_imgproc.dylib /usr/local/lib/libopencv_legacy.dylib /usr/local/lib/libopencv_ml.dylib /usr/local/lib/libopencv_nonfree.dylib /usr/local/lib/libopencv_objdetect.dylib /usr/local/lib/libopencv_photo.dylib /usr/local/lib/libopencv_stitching.dylib /usr/local/lib/libopencv_ts.dylib /usr/local/lib/libopencv_video.dylib /usr/local/lib/libopencv_videostab.dylib"

再继续执行：

phpize
./configure
make
make install

顺利安装完成。