ImageDivision

一种基于半监督的svm的图像分类方法。该方法通过聚类核的方法利用无标记样本局部正则化训练核的表达式。这种方法通过图像直接学习一个自适应的核。该程序仿真的是文章：Semi-supervised Remote Sensing Image Classification with Cluster Kernels。大家可以参考下。(A semi-supervised SVM is presented for the classification of remote sensing images. The method exploits the wealth of unlabeled samples for regularizing the training kernel representation locally by means of cluster kernels. The method learns a suitable kernel directly from the image, and thus avoids assuming a priori signal relations by using a predefined kernel structure. Good results are obtained in image classification examples when few labeled samples are available. The method scales almost linearly with the number of unlabeled samples and provides out-of-sample predictionsds)

通过邻域搜索法对一个相机拍摄的序列图像中圆形非编码标志点进行匹配(Neighborhood search method by camera on a circular non-coding sequence of images to match landmarks )

使用 RANSAC 算法计算仿射变换矩阵(RANSAC algorithm using an affine transformation matrix)

从键盘读取输入信息，并绘制最终的高斯函数。(Read from the keyboard to enter information, and rendering the final Gaussian function.)

给予matlab的JPEG压缩程序，包含原始压缩图片，程序对灰度bmp图片进行JEPG编码，压缩成以0101形式的编码。JPEG编码先把图象色彩RBG变成亮度Y和色度Cr、Cb，它利用人的视觉对色度不敏感的特点，减少一部分色度数据，以达到压缩。 JPEG采取多种编码方式，包含有行程编码(Run Length Coding)和哈夫曼(Huffman)编码，有很高的压缩比。在编码前，先对数据进行分块，离散余弦变换（DCT）及量化，保留能量大的低频信号，丢弃高频信号以达到压缩。(Given to the JPEG compression matlab program, contains the original compressed image, the program of the grayscale bmp picture JEPG encoded, compressed into a 0101 form of encoding. RBG color coded JPEG image into a first luminance Y and chrominance Cr, Cb, which uses the human visual characteristics is not sensitive to the color, reduce the number of color data to achieve compression. Take a variety of JPEG encoding, includes length coding (Run Length Coding) and Huffman (Huffman) coding, a high compression ratio. In the code before the data block, discrete cosine transform (DCT) and quantization, to retain a large low-frequency signal energy, discarding high-frequency signals in order to achieve compression.)

这个实例就像一个魔镜一样，能把原始图像照出各种变换效果,是基于MFC的多文档应用程序(This example is like a mirror,which can convert original image according to a variety of effects)

极化SAR图像处理中，将协方差矩阵转换为STOKES矩阵或Mueller矩阵的程序(Polarization SAR image processing, the covariance matrix will be converted to matrix or Mueller matrix STOKES procedures)

我写的贝叶斯fusion的Visual C+++源码，简明易懂，提供参数说明 可直接使用。 (The Bayesian fusion the Visual C+++ of what I wrote source, easy-to-understand parameters can be used directly.)

1.从图像中识别几何形状的基本方法；基本原理是在于利用点与线的对偶性，将原始图像空间给定的曲线通过曲线表达形式变为参数空间的一个点，这样就把原始图像中给定的检测问题转化为寻找参数空间的峰值问题。 2.基于Gabor变换的人眼检测MATLAB实现。 3.基于Radon的车牌矫正方法。 4.基于模型函数headata产生不同大小的头模型数据。 5.基于函数wdcbm2()设置图像分层阈值压缩参数，实现图像压缩(The basic method of identifying geometry from the image; the basic principle is to use point and line duality, the original image space curve given expression through the curve into a parameter space of a point, so the peak detection problem is given in the original image is transformed into searching the parameter space.)

地震剖面代码,横坐标是时间，纵坐标是道数.(Seismic code, the abscissa is the time, the ordinate is the number.)