等高线绘制

传播神经网络算法的源代码 这是监督的神经网络，需要学习率和势头的两个学习参数

opencv用多种方法提取特征，分别有颜色提取即：颜色直方图提取；形状提取有Hough直线提取、Hough矩形提取以及边缘直方图提取

应用背景粒子滤波目标跟踪源码，通过粒子滤波技术，画一个待选框，然后在视屏中通过待选框中的运动物体，然后对其进行实时跟踪。本人已经调试。关键技术粒子滤波技术结合目标检测复选框，选择运动目标，然后通过视频序列来检测运动目标并跟踪运动目标，从而达到跟踪目标的效果。

camshift 源码 采用opencv实现在视频中可以跟踪想要跟踪的目标，还可以调节阈值等等，是一个很好地学习样本。

基于PCA的图像融合算法，首先求出源图像的协方差矩阵，然后根据该矩阵的最大特征值所对应的特征向量确定各源图像的权值。

以像素方式读取位图像并在windows中显示

三帧差分算法是相邻两帧差分算法的一种改进方法，它选取连续三帧视频图像进行差分运算，消除由于运动而显露背景影响，从而提取精确的运动目标轮廓信息。该算法的基本原理是是先选取视频图像序列中连续三帧图像并分别计算相邻两帧的差分图像，然后将差分图像通过选取适当的阈值进行二值化处理，得到二值化图像，最后在每一个像素点得到的二值图像进行逻辑与运算，获取共同部分，从而获得运动目标的轮廓信息。

StereoLithography(StL) is something that is widely used in CAD/CAM, RapidPrototyping etc. The concept is that any surface or solid is exported to StL format by data exchange packages after they are modeled in commercial CAD packages. This data consists of triangulated facets that approximate the surface of the solid. No topological or geometric information is exported. The surface is subdivided into small triangles. The approximation uses chordal deflection for curved surfaces to smoothen the surface. More the smoothening required, more number of smaller triangles are generated by subdivision resulting in larger size of data file. Now, along with triangles, their facet normals are also generated. The data is written both in ASCII and binary formats. The data in ASCII is written like this:

图像检索，是颜色直方图，可以用VC6.0运行，注释不多，希望能够帮助需要的人来参考