搜索资源列表
OpenCV-tuxiangpeizhu
- 实现相位相关图像配准,编译产生库文件,用于openCV平台。实现了相位相关图像的快速搜索和两幅图像之间的匹配-Implementation phase correlation image registration, create library files compiled for openCV platform. Achieve the rapid phase correlation image search and matching between two images
epipolar-rectification
- 用OpenCV编的对图像进行极线校正的程序,先对图像进行匹配,然后求出基础矩阵,最后对图像进行极线校正-matching, fundametal matric and epipolar recitification
SIFT-algorithm
- D.G lowe 的尺度不变特征转换算法(SIFT)在基于opencv和OpenGL的源码基础上增加了图像的保存功能。用于对图像的尺度不变特征的提取以及图像特征点的匹配-D.G lowe s scale invariant feature transform (SIFT) algorithm based on opencv and on the basis of OpenGL source to increase the function of image preservation. Used
SIFT
- OPENCV,对输入的两幅图像进行特征值提取并且匹配,连线,命令行下以project3.exe 图像1 图像2 方式运行,代码为main.c-SIFT
ORB
- 该算法基于opencv库,能够实现图像ORB角点检测,并实现图像匹配(ORB image key point detection algorithm, fast and efficient implementation of image feature point detection and matching)
Source code
- 在opencv上实现双目测距主要步骤是: 1.双目校正和标定,获得摄像头的参数矩阵: 进行标定得出俩摄像头的参数矩阵 cvStereoRectify 执行双目校正 initUndistortRectifyMap 分别生成两个图像校正所需的像素映射矩阵 cvremap 分别对两个图像进行校正 2.立体匹配,获得视差图: stereoBM生成视差图 预处理: 图像归一化,减少亮度差别,增强纹理 匹配过程: 滑动sad窗口,沿着水平线进行匹配搜索,由于校正后左右图片平行,左图
simulate
- 形状是由图像的轮廓形成的,所以理论上形状识别是通常在边缘或轮廓检测后的步骤。(edge_based_matching The shape is formed by the outline of the image, so theoretically the shape recognition is usually the step after the edge or contour detection)