开发环境:Ubuntu 18.04 LTS + ROS Melodic + ViSP 3.3.1
文章内容主要参考ViSP官方教学文档:https://visp-doc.inria.fr/doxygen/visp-daily/tutorial_mainpage.html

本文主要介绍了如何使用ViSP自动设定阈值对图像进行二值化处理,主要涉及Huang 、Intermodes、Isodata、Mean、Otsu、Triangle等自动阈值划分算法。本文主要参考了imgproc中的 tutorial-autothreshold.cpp 例程。首先要获取这个例程文件并编译它
svn export https://github.com/lagadic/visp.git/trunk/tutorial/imgproc
cd imgproc/autothreshold
mkdir build
cd build 
cmake .. -DCMAKE_BUILD_TYPE=Release -DVISP_DIR=$VISP_WS/visp-build
make 

执行例程,查看效果

./tutorial-autothreshold

在这里插入图片描述

 下面介绍一下代码实现过程

#include <cstdlib>
#include <iostream>
#include <visp3/core/vpImage.h>
#include <visp3/gui/vpDisplayGDI.h>
#include <visp3/gui/vpDisplayOpenCV.h>
#include <visp3/gui/vpDisplayX.h>
#include <visp3/io/vpImageIo.h>

#if defined(VISP_HAVE_MODULE_IMGPROC)
//! [Include]
#include <visp3/imgproc/vpImgproc.h>
//! [Include]
#endif

int main(int argc, const char **argv)
{
//! [Macro defined]
#if defined(VISP_HAVE_MODULE_IMGPROC) && (defined(VISP_HAVE_X11) || defined(VISP_HAVE_GDI) || defined(VISP_HAVE_OPENCV))
  //! [Macro defined]
  //!
  std::string input_filename = "grid36-03.pgm";

  for (int i = 1; i < argc; i++) {
    if (std::string(argv[i]) == "--input" && i + 1 < argc) {
      input_filename = std::string(argv[i + 1]);
    } else if (std::string(argv[i]) == "--help" || std::string(argv[i]) == "-h") {
      std::cout << "Usage: " << argv[0] << " [--input <input image>] [--help]" << std::endl;
      return EXIT_SUCCESS;
    }
  }

  vpImage<unsigned char> I;
  vpImageIo::read(I, input_filename);

  vpImage<unsigned char> I_res(3 * I.getHeight(), 3 * I.getWidth()); //新建图像容器用于储存处理后的图像
  I_res.insert(I, vpImagePoint(I.getHeight(), I.getWidth())); //把输入图像放入图像容器中间位置

#ifdef VISP_HAVE_X11
  vpDisplayX d;
#elif defined(VISP_HAVE_GDI)
  vpDisplayGDI d;
#elif defined(VISP_HAVE_OPENCV)
  vpDisplayOpenCV d;
#endif
  d.setDownScalingFactor(vpDisplay::SCALE_2);
  d.init(I_res);

  //! [Huang]
  vpImage<unsigned char> I_huang = I;
  vp::autoThreshold(I_huang, vp::AUTO_THRESHOLD_HUANG); //huang算法处理
  //! [Huang]
  I_res.insert(I_huang, vpImagePoint());

  //! [Intermodes]
  vpImage<unsigned char> I_intermodes = I;
  vp::autoThreshold(I_intermodes, vp::AUTO_THRESHOLD_INTERMODES); //intermodes算法处理
  //! [Intermodes]
  I_res.insert(I_intermodes, vpImagePoint(0, I.getWidth()));

  //! [IsoData]
  vpImage<unsigned char> I_isodata = I;
  vp::autoThreshold(I_isodata, vp::AUTO_THRESHOLD_ISODATA);//isodata算法处理
  //! [IsoData]
  I_res.insert(I_isodata, vpImagePoint(0, 2 * I.getWidth()));

  //! [Mean]
  vpImage<unsigned char> I_mean = I;
  vp::autoThreshold(I_mean, vp::AUTO_THRESHOLD_MEAN); // mean算法处理
  //! [Mean]
  I_res.insert(I_mean, vpImagePoint(I.getHeight(), 0));

  //! [Otsu]
  vpImage<unsigned char> I_otsu = I;
  vp::autoThreshold(I_otsu, vp::AUTO_THRESHOLD_OTSU); //otsu算法处理
  //! [Otsu]
  I_res.insert(I_otsu, vpImagePoint(I.getHeight(), 2 * I.getWidth()));

  //! [Triangle]
  vpImage<unsigned char> I_triangle = I;
  vp::autoThreshold(I_triangle, vp::AUTO_THRESHOLD_TRIANGLE); //triangle算法处理
  //! [Triangle]
  I_res.insert(I_triangle, vpImagePoint(2 * I.getHeight(), 0));

  vpDisplay::display(I_res);

  vpDisplay::displayText(I_res, 30, 20, "Huang", vpColor::red);
  vpDisplay::displayText(I_res, 30, 20 + I.getWidth(), "Intermodes", vpColor::red);
  vpDisplay::displayText(I_res, 30, 20 + 2 * I.getWidth(), "IsoData", vpColor::red);
  vpDisplay::displayText(I_res, 30 + I.getHeight(), 20, "Mean", vpColor::red);
  vpDisplay::displayText(I_res, 30 + I.getHeight(), 20 + I.getWidth(), "Original", vpColor::red);
  vpDisplay::displayText(I_res, 30 + I.getHeight(), 20 + 2 * I.getWidth(), "Otsu", vpColor::red);
  vpDisplay::displayText(I_res, 30 + 2 * I.getHeight(), 20, "Triangle", vpColor::red);

  vpDisplay::flush(I_res);
  vpDisplay::getClick(I_res);

  return EXIT_SUCCESS;
#else
  (void)argc;
  (void)argv;
  return 0;
#endif
}

如果大家对于深度学习与计算机视觉领域感兴趣,希望获得更多的知识分享与最新的论文解读,欢迎关注我的个人公众号“深视”。

在这里插入图片描述