wigner_ville

基于otsu的自适应阀值canny边缘检测(adaptive canny edge detection based on otsu)

EM算法是最为经典的统计图像分割方法，代码基于matlab，使用EM算法实现图像分割,源代码加实例图像，非常好用(By using EM algrithm to realize the segementation of image)

实现图像配准:线性保角变换，投影变换及仿射变换。(Achieve image registration: linear conformal transformation, projection transformation and affine transformation.)

孔径雷达图像固有的相干斑噪声严重降低了图像的可解译程度，影响了后续目标检测、分类和识别 等应用。因此，SAR图像的相干斑抑制问题一直是SAR图像应用的重要课题之一。一个理想的去斑算法应该在平 滑的同时保持图像的边缘等细节不受损失，目前存在各种各样的算法，但没有一种方法能够完美的满足这一要求。 本文对SAR图像的相干斑抑制问题进行了全面系统的研究，分析了相干斑的形成原因，总结了目前存在的相干斑 抑制算法的主要思路，介绍了具有代表性的算法，并对各种算法的性能进行了定性分析和比较，给出了去斑算法定 量评估方法(The inherent speckle noise of the aperture radar image seriously reduces the degree of interpretation of the image, and affects the detection, classification and recognition of the subsequent targets. And other applications. Therefore, the problem of speckle suppression in SAR images has been one of the most important topics in the application of SAR images. An ideal speckle algorithm should be flat At the same time, the details such as edges of images are not lost. There are various algorithms at present, but none of them can satisfy this requirement perfectly.)

说明：  从单张图片中去除相机抖动造成的模糊的matlab程序，算法见Removing Camera Shake from a Single Image(err)

说明：  多目标跟踪，动态背景下实现多目标跟踪算法，算法优良，实时性和鲁棒性都比较好(Tracking, dynamic context to achieve multi-target tracking algorithm, algorithm quality, real-time and better robustness than)

说明：  很不错的SVM的MATLAB源码，不用设置路径，比libsvm还好用！(SVM is a good source of MATLAB, do not set up the path, than to use libsvm better!)

说明：  pca-svm的人脸识别，利用pca进行人脸特征提取，再利用svm进行聚类分析(Face recognition based on PCA SVM)

下面给出了48项泽尼克多项式，外加一项常数项。需要注意的是，读者并不需要严格按照下文所示的顺序排列这些泽尼克项，实际上在不同的应用和机构会采用不同的排列顺序。 表中的#0项是个常数或者说是平移项（piston term），这一项的系数也代表了平均光程差；而#1和#2项分别是x和y方向的倾斜项（tilt terms），#3代表了聚焦，因此，#1到#3项代表了波前的高斯或者近轴特性；#4和#5项代表了像散和离焦，#6和#7项代表彗差和倾斜，而#8项代表了3级像差和离焦，也就是说#4到#8项为3级相差项；同样地，#9到#15项代表了5级像差，而#16到#24项代表了7级像差，#25到#35项代表了9级像差，#36到#48项代表了11级像差。(Here are 48 Zernike polynomials, plus a constant term. It is important to note that the reader does not need to strictly arrange the Zernike entries in the order shown below, but in fact different orders of arrangement are applied to different applications and institutions. Table #0 is a constant or a shift (piston term), a coefficient of this term also represents the average optical path difference; while #1 and #2 are inclined X and Y direction (tilt terms), #3 represents the focus, therefore, #1 to #3 on behalf of the wavefront. Gauss or paraxial properties; #4 and #5 represent the astigmatism and defocus, #6 and #7 represent coma and tilt, and #8 represents the 3 order aberration and defocus, that is to say #4 to #8 is 3 level difference; similarly, #9 to #15 on behalf of the 5 order aberration #16 and #24, to represent the 7 level #25 to #35 aberration, on behalf of the 9 #36 to #48 aberration, on behalf of the 11 order aberration.)

经典的ROF模型，可以用于图像去噪，图像分解；TV图像修复(The ROF model can be used for image denoising)