install opencv on ubuntu 12.10

apt-get  install on shiyanlou 的速度？
apt-source ?

After this operation, 100 MB of additional disk space will be used.

step 1

setp 1.1 get opencv source

準備原始碼 opencv
Opencv的原始碼,2.4.9,90M左右
百度網盤連結：

或者類似方法

wget http://downloads.sourceforge.net/project/opencvlibrary/opencv-unix/2.4
 
.2/OpenCV-2.4.2.tar.bz2

tar xvf OpenCV-2.4.2.tar.bz2

step 1.2 install build-essential

build-essential 軟體包，涵蓋gcc/g++/gdb/make

    sudo apt-get install build-essential  

step 1.3 install pakcage about multi-media

支援讀寫圖片以及視訊等,about 100MB

    sudo apt-get install  libgtk2.0-dev libavcodec-dev libavformat-dev
 
  libtiff4-dev  libswscale-dev libjasper-dev  

step 1.4 install cmake

編譯，大小10M左右

sudo apt-get install cmake

step 1.5 install pkg-config

pkg-config，它是一個提供從原始碼中編譯軟體時查詢已安裝的庫時使用的統一介面的計算機軟體。

sudo apt-get install pkg-config  

二、編譯安裝過程

2.1 環境監測：

使用cmake來檢測編譯環境以及生成makefile以及其他project資訊，在使用cmake過程可以設定一些引數。

在本目錄下(opencv原始碼目錄)使用cmake，注意後面的點表示在當前目錄下尋找CMakeLists.txt。

    cmake .  

結果簡略如下：

   ...
-- Configuring done
-- Generating done
-- Build files have been written to: /media/sda_m/opencv-2.4.9

2.2 使用make編譯

耗時間20分鐘，下一步試試有沒有已經編譯好的bin檔案

make  

正確編譯結果是：

[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/traincascade.cpp.o
[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/cascadeclassifier.cpp.o
[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/boost.cpp.o
[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/features.cpp.o
[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/haarfeatures.cpp.o
[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/lbpfeatures.cpp.o
[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/HOGfeatures.cpp.o
[100%] Building CXX object apps/traincascade/CMakeFiles/opencv_traincascade.dir/imagestorage.cpp.o
Linking CXX executable ../../bin/opencv_traincascade

2.3 使用make install 安裝

sudo make install  

正確的結果提示是：

...
-- Set runtime path of "/usr/local/bin/opencv_haartraining" to "/usr/local/lib"
-- Installing: /usr/local/bin/opencv_createsamples
-- Set runtime path of "/usr/local/bin/opencv_createsamples" to "/usr/local/lib"
-- Installing: /usr/local/bin/opencv_performance
-- Set runtime path of "/usr/local/bin/opencv_performance" to "/usr/local/lib"
-- Installing: /usr/local/bin/opencv_traincascade
-- Set runtime path of "/usr/local/bin/opencv_traincascade" to "/usr/local/lib"

2.4 新增動態庫資訊

在檔案/etc/ld.so.conf中新增 /usr/local/lib（這個跟安裝目錄有關， {CMAKE_INSTALL_PREFIX}/lib），
也可以在/etc/ld.so.conf.d 目錄下增加一個conf檔案（可以命名為 opencv.conf），同樣新增 /usr/local/lib

檢視已經生成的動態庫
[email protected]:/usr/local/lib# ls
libopencv_calib3d.so           libopencv_gpu.so            libopencv_nonfree.so          libopencv_superres.so
libopencv_calib3d.so.2.4       libopencv_gpu.so.2.4        libopencv_nonfree.so.2.4      libopencv_superres.so.2.4
libopencv_calib3d.so.2.4.9     libopencv_gpu.so.2.4.9      libopencv_nonfree.so.2.4.9    libopencv_superres.so.2.4.9
libopencv_contrib.so           libopencv_highgui.so        libopencv_objdetect.so        libopencv_ts.a
libopencv_contrib.so.2.4       libopencv_highgui.so.2.4    libopencv_objdetect.so.2.4    libopencv_video.so
libopencv_contrib.so.2.4.9     libopencv_highgui.so.2.4.9  libopencv_objdetect.so.2.4.9  libopencv_video.so.2.4
libopencv_core.so              libopencv_imgproc.so        libopencv_ocl.so              libopencv_video.so.2.4.9
libopencv_core.so.2.4          libopencv_imgproc.so.2.4    libopencv_ocl.so.2.4          libopencv_videostab.so
libopencv_core.so.2.4.9        libopencv_imgproc.so.2.4.9  libopencv_ocl.so.2.4.9        libopencv_videostab.so.2.4
libopencv_features2d.so        libopencv_legacy.so         libopencv_photo.so            libopencv_videostab.so.2.4.9
libopencv_features2d.so.2.4    libopencv_legacy.so.2.4     libopencv_photo.so.2.4        pkgconfig
libopencv_features2d.so.2.4.9  libopencv_legacy.so.2.4.9   libopencv_photo.so.2.4.9      python2.7
libopencv_flann.so             libopencv_ml.so             libopencv_stitching.so
libopencv_flann.so.2.4         libopencv_ml.so.2.4         libopencv_stitching.so.2.4
libopencv_flann.so.2.4.9       libopencv_ml.so.2.4.9       libopencv_stitching.so.2.4.9

使用下面的動態庫管理命令ldconfig，讓opencv的相關連結庫被系統共享

    sudo ldconfig -v  

如果成功的話，我的檢測方法如下，正確的結果將會發現前面動態庫的名字

sudo ldconfig -v >temp.so.txt && cat temp.so.txt|grep opencv

2.5 指定標頭檔案位置

完成了動態庫的共享，還不能使用opencv來程式設計，還需要為程式指定openvc的標頭檔案位置。這裡使用pkg-config命令來完成。首先在 /etc/profile 中新增

export  PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig  

pkg-config 維護著一個關於opencv的配置檔案，這個檔案（opencv.pc）在目錄 /usr/local/lib/pkgconfig，這個檔案記錄著opencv的一些動態庫資訊，還有標頭檔案資訊

輸入pkg-config –cflags opencv 和pkg-config –libs opencv ,可以發現以下資訊。

正確的結果是如下：

[email protected]:/usr/local/lib# pkg-config --cflags opencv
-I/usr/local/include/opencv -I/usr/local/include 


[email protected]:/usr/local/lib# pkg-config --libs opencv


/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_ocl.so /usr/local/lib/libopencv_photo.so /usr/local/lib/libopencv_stitching.so /usr/local/lib/libopencv_superres.so /usr/local/lib/libopencv_ts.a /usr/local/lib/libopencv_video.so /usr/local/lib/libopencv_videostab.so -lrt -lpthread -lm -ldl 

3 測試

編譯sample

在解壓的opencv目錄下找到 samples/c/build_all.sh，執行該檔案

[email protected]:/media/sda_m/opencv-2.4.9/samples/c$ ./build_all.sh 
compiling contours.c
compiling convert_cascade.c
compiling delaunay.c
compiling fback_c.c
compiling morphology.c
compiling motempl.c
compiling polar_transforms.c
compiling pyramid_segmentation.c
compiling adaptiveskindetector.cpp
compiling bgfg_codebook.cpp
compiling blobtrack_sample.cpp
compiling facedetect.cpp
compiling find_obj.cpp
compiling find_obj_calonder.cpp
compiling find_obj_ferns.cpp
compiling latentsvmdetect.cpp
compiling mser_sample.cpp
compiling mushroom.cpp
compiling one_way_sample.cpp
compiling smiledetect.cpp
compiling tree_engine.cpp

list all the executive

find . -perm /u=x,g=x,o=x

結果是

.
./find_obj_calonder  #?
./adaptiveskindetector #?
./delaunay #ok

./facedetect #ok
./mser_sample #？
./blobtrack_sample #？
./one_way_sample #?
./latentsvmdetect  #ok
./tree_engine #?
./bgfg_codebook #?
./motempl #?
./fback_c #?
./polar_transforms #?
./contours #ok
./find_obj #Ok 
./smiledetect #ok
./mushroom #?
./convert_cascade  #?
./morphology #ok
./pyramid_segmentation #ok
./find_obj_ferns   #?

期中末尾表示為 #?是當前不知道作用的，其他的ok都有結果的

example find object

smile detect

face detect

動態的 delaunay 三角+ voronoi分割

This program demostrates iterative construction of
delaunay triangulation and voronoi tesselation.
It draws a random set of points in an image and then delaunay triangulates them.

有趣 morphology

fail example

find_obj_ferns

./find_obj_ferns box.png box_in_scene.png 

This program shows the use of the "fern" plannar PlanarObjectDetector point
descriptor classifier
Usage:
./find_obj_ferns <object_filename> <scene_filename>, default: box.png and box_in_scene.png

Trying to load box.png_model.xml.gz ...
The file not found and can not be read. Let's train the model.
Step 1. Finding the robust keypoints ...

adaptiveskindetector

需要一個照片序列

./adaptiveskindetector

This program demonstrates the contributed flesh detector CvAdaptiveSkinDetector which can be found in contrib.cpp
Usage: 
./adaptiveskindetector fileMask firstFrame lastFrame

Example: 
./adaptiveskindetector C:\VideoSequences\sample1\right_view\temp_%05d.jpg  0  1000
   iterates through temp_00000.jpg  to  temp_01000.jpg

find_obj_calonder:Calonder point descriptor classifierSURF

./find_obj_calonder
This program shows the use of the Calonder point descriptor classifierSURF is used to detect interest points, Calonder is used to describe/match these points
Format:
   classifier_file(to write) test_image file_with_train_images_filenames(txt)   or
   classifier_file(to read) test_image

others

• blobtrack_sample 輸入未知

• one_way_sample 2 photo need

• tree_engine

• /bgfg_codebook

Learn background and find foreground using simple average and average difference learning method:
Originally from the book: Learning OpenCV by O'Reilly press

檢視example下面的圖片

 ll *.jpg
eog 檢視下面的圖片

小結

搭建環境
玩玩example
找到興趣方向

• 動態的 delaunay 三角+ voronoi分割

參考

windows install

參考 淺墨

get sample

form git or pan

error 1:    error C1083: 無法開啟包括檔案:“opencv2/opencv.hpp”: 

vc++目錄
- include
- include
- opencv2
- opencv
- lib
- linker input
- opencv_core248d.lib
- opencv_imgproc248d.lib
- opencv_highgui248d.lib

error 2: 丟失opencv_core*.dll

至此，環境變數在裡面，但是竟然一直找不到。只有使用蠻力法，全部拷貝