TVD_mm

CGAL，Computational Geometry Algorithms Library，计算几何算法库，设计目标是，以C++库的形式，提供方便，高效，可靠的几何算法。CGAL可用于各种需要几何计算的领域，如计算机图形学，科学可视化，计算机辅助设计和建模，地理信息系统，分子生物学，医学成像，机器人运动规划，网格生成，数值方法等等。 　　 计算几何算法库（CGAL），提供计算几何相关的数据结构和算法，诸如三角剖分（2D约束三角剖分及二维和三维Delaunay三角剖分），Voronoi图（二维和三维的点，2D加权Voronoi图，分割Voronoi图等），多边形（布尔操作，偏置），多面体（布尔运算），曲线整理及其应用，网格生成（二维Delaunay网格生成和三维表面和体积网格生成等），几何处理（表面网格简化，细分和参数化等），凸壳算法（2D，3D和dD），搜索结构（近邻搜索，kd树等），插值，形状分析，拟合，距离等。(The goal of the CGAL Open Source Project is to provide easy access to efficient and reliable geometric algorithms in the form of a C++ library. CGAL is used in various areas needing geometric computation, such as: computer graphics, scientific visualization, computer aided design and modeling, geographic information systems, molecular biology, medical imaging, robotics and motion planning, mesh generation, numerical methods... More on the projects using CGAL web page. The Computational Geometry Algorithms Library (CGAL), offers data structures and algorithms like triangulations (2D constrained triangulations, and Delaunay triangulations and periodic triangulations in 2D and 3D), Voronoi diagrams (for 2D and 3D points, 2D additively weighted Voronoi diagrams, and segment Voronoi diagrams), polygons (Boolean operations, offsets, straight skeleton), polyhedra (Boolean operations), arrangements of curves and their applications (2D and 3D envelopes, Minkowski sums), mesh generation (2D Del)

说明：  国外一个博士写的图像复原的源码 ，基于自适应的正则化的全变分图像复原，源码内有作者的主页，相应文章可以下载。代码测试，可用，效果不错 ，做这方面的同志可以下来看看 (Dr. written by a foreign source image restoration based on adaptive regularization total variation image restoration, source inside of the home page, the corresponding articles can be downloaded. Code test, available, good results, so this could be down to see Comrade)

matlab实现的直线检测方法，主要为hough变换，包括图片及实验结果(Matlab implementation of the line detection method, mainly for the Hough transform, including pictures and experimental results)

本文针对多光谱图像配准中不同光谱图像间存在较大的灰度差异、同一景物在不同波段表现的特征也各不相同等问题，提出了基于互信息相似性测度的多光谱图像配准算法。(In this paper, multi-spectral image registration between different spectral image there is a big difference in gray scale, the same scene in different bands performance characteristics are different questions, we present a similarity measure based on mutual information of multi-spectral image registration algorithm .)

说明：  基于距离变换的分水岭分割算法，该算法可以实现对粘连物体的分割。(Watershed algorithm)

图像纹理特征提取，包括分形维数，灰度游程长度，灰度共生矩阵等五种图像纹理特征。(To extract image texture features, including fractal dimension, length of gray-level run-length, gray level co-occurrence matrix and other five kinds of image texture features。)

sift和surf特征点提取，然后进行配准，通过ransac进行提纯(Sift and surf feature points are extracted, then registered and purified by RANSAC)

基于MATLAB平台下，可获取图片任意点的像素坐标。(Based on MATLAB platform, the pixel coordinates of any point in the picture can be obtained.)

关于图像配准中的交叉相关的实现，可以直接调用。(About Image Registration in the implementation of cross-correlation can be directly called.)

说明：  性能上比经典分水岭有很大的提高。特别对于自然图像的分割，效果特别好。(Performance than the classical watershed have greatly improved. Especially for natural image segmentation, the effect particularly good.)