一套图像处理程序(vc++)，支持JPEG、GIF

在计算机视觉中首先一步是计算每个摄像机的参数(外参和内参)，这是重建物体所必须的，这一个过程称为摄像机定标(camera calibration)。这是我自己写的一个简单的采用2维模板(模板是由黑白相间的格子组成，格子大小60cm×60cm)定标的程序，交互性比较好，每个角点通过鼠标点击自动获取。-Computer Vision in the first step is to calculate the parameters of each camera (outside the Senate and internal), which is necessary for the reconstruction of objects, which a process known as camera calibration (camera calibration). This is what I wrote it myself using a simple 2-D template (template is composed of black-and-white lattice, lattice size of 60cm × 60cm) calibration procedures, better interactivity, each corner through the automatic acquisition of mouse clicks.

基于视频运动对象的快速提取 采用了差分法,光流法以及贝叶斯估计-Moving Object Based Video Fast extracted using a finite difference method, optical flow estimation method and the Bayesian

应用于计算机视觉领域，拍摄等像素大小的两张图片，其中一张是细节图，我们可以将我们关注的这张细节图（拍摄得清楚）‘粘贴’到大图中，起到细化局部的作用-Applied to computer vision field, shooting and other pixel size of two images, one of which details of the map, we can be of concern to us the details of this map (taken a very clear) " paste" into the big picture, the play of local refinement role

实现在YUV图像上面按照算法叠加OSD信息-Achieve the YUV overlay the image above, OSD information in accordance with Algorithm

2005/12/07 此程序主要使用mallet算法实现的是97小波变换程序 -2005/12/07 The procedure used mallet algorithm is 97 Wavelet Transform Procedure

读取位图的通用代码 可以在此基础上添加自己所需的图像处理功能-Read bitmap generic code on this basis can add your own image processing functions required for

基于Matlab编写的语音端点检测程序，希望对大家有帮助-prepared based on Matlab voice endpoint detection procedures, we hope to help

基于NL_Means的均值平移图像分割算法-NL_Means based on the mean shift image segmentation algorithm

Hough变换对圆的检测 Hough变换的基本原理在于，利用点与线的对偶性，将图像空间的线条变为参数空间的聚集点，从而检测给定图像是否存在给定性质的曲线。 -Hough transform for circle detection Hough transform basic principle is that the use of point and line duality, the image space into a parameter space of lines gather, thus detecting the existence of a given image given the nature of the curve.

sift特征匹配技术讲义。详细介绍了sift算法的实现过程。-sift matching technology overhead. Sift the details of the algorithm implementation process.