搜索资源列表
photomachinenewalgorithm
- 针对摄像机固定和共面约束条件下进行目标定位,提出了一种简单可行的全新算法。该算法避开了烦琐的摄像机标定这一过程,也不用精确地求解单应性矩阵。只是运用射影几何的性质,通过简单的几何关系,对非标定单摄像机拍摄的单幅图像,求解其从2D图像平面到真实的3D平面的非线性单映射关系。该算法为解决计算机视觉的测量、定位、三维重建等问题,提供一种全新的思路。 -against cameras fixed and coplanar binding conditions for target location,
Levenberg-Marquardt
- Levenberg-Marquardt优化单应性矩阵,也可经过修改用于相机标定参数的优化
homography
- 自己编写的单应性矩阵的求法,按照张正友的论文进行编写的。希望有用。
homework.rar
- 应用于计算机视觉领域,运用单应性矩阵信息将两幅图片合成为一张,Used in the field of computer vision, using homography information will be combined into a two picture
bird_eye
- 利用已有的摄像头内外参数,将侧视图通过透视变换转换为俯视图,并保存所得到的单应性矩阵H,内含运行实例-Use of existing internal and external camera parameters, the side view through the perspective transformation into top view, and save the resulting homography matrix H, containing a running instance
perspective
- 透视变换,根据已得到的单应性矩阵H,将摄像头所拍摄的侧视图转换为俯视图,内含运行实例-Perspective transformation, according to what has already been homography H, the camera is filming side view into vertical view, containing a running instance
The_Levenberg-Marquardt_Algorithm
- LM算法 老外写的The Levenberg-Marquardt (LM) algorithm is the most widely used optimization algorithm. It outperforms simple gradient descent and other conjugate gradient methods in a wide variety of problems. This document aims to provide an intuitiv
Fns
- 一些图像处理常用的函数,包括图像之间的点匹配、鲁棒性估计、图像旋转、基本矩阵的求解、单应矩阵求解等,可用于摄像机标定-Commonly used in a number of image processing functions, including point match between the image and robustness of the estimates, image rotation, solve the fundamental matrix, homography solv
JHSDB
- :针对常用坐标转换定位算法精确度和效率较低的问题,提出一种基于透视变换原理的定位算法。根据4 个初始定位点的坐标信息, 确定反映点与点之间空间映射关系的单应性矩阵。定位完成后,利用该矩阵进行简单运算即可完成坐标转换。实验结果证明该方法稳定性 较高。-Positioning algorithm for common coordinate conversion accuracy and low efficiency problem, a perspective transformation
image-mosaic
- 图像拼接,利用SIFT算法提取点特征,并进行点匹配,解算单应性矩阵,根据结算出的单应性矩阵对图像进行仿射变换,最后进行拼接,程序中的备注很详细,各位感兴趣的同志可以下载看看。-Image mosaic using SIFT feature point extraction algorithm, and point matching, homography matrix solver, based on the settlement of the homography image affine t
ImageMatch
- 实现两幅图像的拼接,计算单应性矩阵,通过随机抽样一致性算法-Achieve two image stitching, calculate homography, by random sampling consensus algorithm
homography-matrix
- 通过一张图片基于张正友标定的2D视觉标定,其结果为一个单应性矩阵,基于Opencv的角点提取,标定的计算是自己编写,内有注释和原理,已在工程实践中得到应用。-Zhang Zhengyou based 2D calibration for camera calibration by a picture, the result is a homography, extraction based on corner Opencv, the calibration is calculated to wr
birds-eye
- 鸟瞰图变换实例,通过读取摄像机内外参数矩阵和放仿射变换的到的单应性矩阵,获得平面视图的“俯视图”。-Aerial view transform instance, affine transformation to homography by reading the internal and external camera parameters matrix and put, get a plan view of the " plan view."
用RANSAC方法计算两幅图像间的单应性矩阵
- 用RANSAC方法计算两幅图像间的单应性矩阵,包含了点集的选取,ransac去除误匹配,和单应性矩阵计算程序3部分
cuphigh
- 利用摄像机的单应性矩阵测量计算水杯的高度(Measuring the height of a cup by using a homography of a camera)
归一化
- 这个程序运用归一化的方法进行单应性矩阵求解。(The homography matrix is solved using a normalized method.)
compute_homography
- 用于求解张正友相机标定过程中的单应性矩阵(Homography matrix for camera calibration)
图像配准算法
- 1.SIFT得到两幅图像的匹配点对 2.通过RANSAC剔除外点,得到N对内点 3.利用DLT和SVD计算全局单应性 4.将源图划分网格,取网格中心点,计算每个中心点和源图上内点之间的欧式距离和权重 5.将权重放到DLT算法的A矩阵中,构建成新的W*A矩阵,重新SVD分解,自然就得到了当前网格的局部单应性矩阵 6.遍历每个网格,利用局部单应性矩阵映射到全景画布上,就得到了APAP变换后的源图 7.最后就是进行拼接线的加权融合