intrans

RANSAC算法是一种经典的去外点方法，它可以利用特征点集的内在约束关系来去除错误的匹配。(RANSAC algorithm is a classic way to go outside the points, it can take advantage of the inherent characteristics of point sets constraints to remove false matches. )

图像处理程序，基于K均值聚类的图像分割——灰度图像，基于中位数算法的运动目标检测(Image processing program, image segmentation based on k-means clustering -- gray image, moving target detection based on median algorithm)

类似于photoshop软件处理，功能复杂简单耐用，适合于初学者学习参考(Similar to the photoshop software processing, functional complexity of simple and durable, suitable for beginners to learn information)

c#编写的图像平滑处理，源码比较全，可以成功编译，希望对大家有帮助。(c# prepared by the image smoothing, source code comparison, can be successfully compiled, we want to help.)

说明：  这是一个在vc6.0基于OpenCV的控制台程序，本例给出一个椭圆曲线拟合的例子，程序首先发现图像的轮廓，然后用椭圆逼近它。(This is a vc6.0 console based on the OpenCV procedures, in this case gives an example of elliptic curve fitting, the program first to discover the outline of the image, and then close it with an oval.)

说明：  在计算机视觉中，图像分割是将图像分成多个片段的过程。这使我们能够通过以比原始图像更简单的表示方式来更好地分析图像。一种用于分割图像的方法是阈值分割；也就是说，将强度值低于阈值的所有像素设置为背景值，而将高于阈值的其他像素设置为前景值。最简单的阈值处理形式是对所有像素使用全局阈值，而自适应阈值处理根据像素及其周围环境动态改变阈值。典型地，阈值处理从灰度图像开始，并输出二进制图像以清晰地描绘图像中的片段。在本实验中，我们开发了一种自适应阈值算法，并将其与更简单的算法进行了比较。 此外，我们还开发了逆运动学证明来将机械臂移动到给定的x，y，z坐标。这一证明涉及三种命名身份，并获得了对机械臂设计的更好理解。(Image threshold segmentation)

运动目标检测与跟踪，包括目标定位，匹配与跟踪!(moving target detection and tracking, including Target, matching and tracking!)

由图像的灰度，恢复三维物体的形状，包括源程序和图片(By the gray scale image to restore the shape of three-dimensional objects, including source code and pictures)

根据paper思想，自己编写的SPIHT编解码。可以处理不同图像大小的小波系数，严格按照paper中的算法编写。细节见说明文档(under the paper thinking, their preparation SPIHT codec. Image can handle different sizes of the wavelet coefficients and strictly according to the paper prepared algorithm. For details, see the documentation)

对玻璃表面缺陷图像进行处理，进行分类识别(The glass surface defect image processing, classification and identification)