vigra1.8.0的使用

VIGRA stands for "Vision with Generic Algorithms". It's a novel computer vision library that puts its main emphasis on customizablealgorithms and data structures.

1、首先，从http://hci.iwr.uni-heidelberg.de/vigra/下载最新源代码，并解压缩；

2、 从http://www.cmake.org/cmake/resources/software.html 下载cmake，并安装；

3、 打开cmake，在where is the source code中添加vigra源代码所在的位置，如：D:/soft/vigra/vigra-1.8.0-src/vigra-1.8.0；在where to build the binaries中添加编译后所存放的位置，如：D:/Program Files/vigra/vs2008。

4、 点击Configure，在弹出的对话框中选择Visual Studio 9 2008；如果有红色框显示，可根据实际情况进行修改，继续点击Configure;

点击Generate;会出现一个错误, CMake Error: Unknown Target referenced : doc_python，修改方法：打开D:/soft/vigra/vigra-1.8.0-src/vigra-1.8.0/ CMakeLists.txt，将其中的

IF(WITH_VIGRANUMPY)

ADD_DEPENDENCIES(PACKAGE_SRC_TAR doc_python)

ENDIF()

替换为：

IF(WITH_VIGRANUMPY AND PYTHON_SPHINX)

ADD_DEPENDENCIES(PACKAGE_SRC_TAR doc_python)

ENDIF()

然后重新打开cmake，再重复上面步骤即可。

5、上述步骤完成之后，在vs2008文件夹中，就会看到有vigra.sln工程文件，打开此工程，分别在Debug和Release下编译整个工程，将会在vs2008/src/impex文件夹下生成Debug和Release的vigraimpex动态和静态库。

默认支持的图像格式包括：BMP、GIF、HDR、PNM、SUN、VIFF，若想支持更多种格式的图像，需要：源代码解压后会有一个” vigra-dependencies-win32-vs8.zip”压缩文件，将此文件解压到当前目录，会产生一个” dependencies”文件夹。在CMake中，在DDEPENDENCY_SEARCH_PREFIX中，将其value设为dependencies所在的目录即可。

此图像库的缺点是速度慢。

例子：

// TestVigra.cpp : Defines the entry point for the console application.

//

#include "stdafx.h"

#include <iostream>

#include <string>

#include "vigra/stdimage.hxx"

#include "vigra/impex.hxx"//Image import and export functions

#include "vigra/edgedetection.hxx"

using namespace vigra;

using namespace std;

int _tmain(int argc, _TCHAR* argv[])

{

try

{

cout<<"supported formats: "<<impexListFormats()<<endl;

string strImageName = "E:\\vigra\\Image\\base.jpg";

//string strImageName = "E:\\vigra\\Image\\basegray.jpg";

string strOutImage = "E:\\vigra\\Image\\new.jpg";

ImageImportInfo info(strImageName.c_str(), 0);//read image

//vigra_precondition(info.isGrayscale(), "Sorry, cannot operate on color images");

double threshold=200, scale=0.5;

if (info.isGrayscale())

{

BImage out(info.width(), info.height()); // create a gray scale image of appropriate size

BImage in(info.width(), info.height());

importImage(info, destImage(in));

out = 255;// paint output image white

importImage(info, destImage(out));// import the image just read

//differenceOfExponentialEdgeImage(srcImageRange(in), destImage(out), scale, threshold, 0);

//cannyEdgeImage(srcImageRange(in), destImage(out), scale, threshold, 0);// call edge detection algorithm

transformImage(srcImageRange(in), destImage(out), linearIntensityTransform(-1, -255));//invert image

exportImage(srcImageRange(out), ImageExportInfo(strOutImage.c_str()));// write the image to the file

}

else

{

BRGBImage out(info.width(), info.height());// create a RGB image of appropriate size

BRGBImage in(info.width(), info.height());

importImage(info, destImage(out));

importImage(info, destImage(in));

//RGBValue<int> offset(-255, -255, -255);

//transformImage(srcImageRange(in), destImage(out), linearIntensityTransform(-1, offset));

double sizefactor = 1.2;

int nw = (int)(sizefactor*(info.width()-1) + 1.5); // calculate new image size

int nh = (int)(sizefactor*(info.height()-1) + 1.5);

BRGBImage out1(nw, nh);

resizeImageSplineInterpolation(srcImageRange(in), destImageRange(out1));// resize the image, using a bi-cubic spline algorithms

exportImage(srcImageRange(out1), ImageExportInfo(strOutImage.c_str()));

}

}

catch (StdException &e)

{

cout<<e.what()<<endl;// catch any errors that might have occurred and print their reason

return 1;

}

return 0;

}