objection_tracking

说明：  Zigzag和iZigzag的程序及改进的Zigzag和iZigzag的程序。后者可以大幅提高运算效率(Zigzag and iZigzag procedures and improved Zigzag and iZigzag procedures. The latter can substantially increase the computing efficiency)

仿射变换,2D图形编程,图形图象,有matlab操作界面，可实现任意二维变换(Affine transformation, 2D graphics programming, graphics, images, has matlab interface can realize any two-dimensional transform)

基于偏微分方程的图像,利用扩散方程进行图像修补，同时还有使用CDD (curvature driven diffusions)模型进行修补的程序(Image based on partial differential equations, using diffusion equation for image repair, as well as use CDD (curvature driven diffusions) the procedure to repair the model)

说明：  对偶四元数应用于空间坐标变换，和对偶四元数和齐次矩阵的转换(Dual quaternion coordinate transformation applied to space, and dual quaternion and homogeneous matrix conversion)

基于谱间预测的高光谱图像压缩。单向预测与双向预测，以及参考波段的最优选取(The prediction based on spectral hyperspectral image compression. One-way bi-directional prediction and prediction, as well as reference to select the optimal band)

根据离散数据点拟合圆，得到圆心坐标和拟合半径(Fitting the circle according to the discrete data points, the coordinates of the center of the circle and the fitting radius are obtained.)

Matlab里面没有提供直接以半径和圆心位置画圆的程序，该程序可以实现此功能，命令格式为：circle(radius,x0,y0, color )(Matlab which does not provide the location directly in the center circle radius and the program, the program can achieve this functionality, the command format is: circle (radius, x0, y0, color ))

说明：  在MTLAB环境运行main_msbcsspl.m。该程序的功能是实现图像的稀疏分解并重建图像，采用压缩感知理论，利用小波分析，在小波域分块观测，可以用很少的观测值就能重建原图像。程序包中已经整合了小波包和测试图像，没有多余文件，程序经过亲测并修改，能在MATLAB中直接运行得到很好的结果，速度非常快！(Run main_msbcsspl.m in the MTLAB environment. The function of the program is to decompose and reconstruct the image sparsely. Using compressed sensing theory and wavelet analysis, the original image can be reconstructed with very few observations. Wavelet package and test image have been integrated in the package. There are no redundant files. After testing and modifying the program, it can run directly in MATLAB and get good results. The speed is very fast.)

主要是实现图像的融合,是在梯度域里面进行一个图像的融合，(shi xian tu xiang de rong he，in tidu yulimian rong he ercheng ,matlab huan jing,xia ,keyi yunxing qieyou ceshi shuju d e.)

用于求取三维点云数据的法向量，我的人脸建模就是用的这个，效果很好，但必须保证点云没有重叠的，否则会出现严重错误。(Strike a 3D point cloud data for normal vectors, and my face is to use this model, the effect is very good, but must ensure that there is no overlap of the point cloud, there would be a serious mistake.)