转载地址:https://en.wikipedia.org/wiki/Perspective-n-Point
PnP
Definition[edit]
Given a set of n 3D points in a world reference frame and their corresponding 2D image projections as well as the calibrated intrinsic camera parameters, determine the 6 DOF pose of the camera in the form of its rotation and translation with respect to the world. This follows the perspective project model for cameras:
- .
where is the homogeneous world point, is the corresponding homogeneous image point, is the matrix of intrinsic camera parameters, (where and are the scaled focal lengths, is the skew(偏移)parameter which is sometimes assumed to be 0, and is the principal point), is a scale factor for the image point, and and are the desired 3D rotation and 3D translation of the camera (extrinsic parameters) that are being calculated. This leads to the following equation for the model:
- .
-
-
Assumptions and Data Characteristics
PnP问题的对象是:相机已校正,求相机外参,即R-t;
UPnP 和 Direct Linear Transform (DLT)它们是针对projection model,即不仅估计相机外参,而且估计相机内参。
-
For each solution to PnP, the chosen point correspondences cannot be coplanar.( 一个3D点和其对应的 2D点构成一条直线,PnP中将来会有n条直线。如果要成为PnP的解,这n条直线不能共面,这为选择3D特征点提供了依据,这句话应该是这个意思。)
In addition, PnP can have multiple solutions, and choosing a particular solution would require post-processing of the solution set. Furthermore, using more point correspondences can reduce the impact of noisy data when solving PnP. RANSAC is also commonly used with a PnP method to make the solution robust to outliers in the set of point correspondences.
一篇介绍P3P很好的博客:http://iplimage.com/blog/p3p-perspective-point-overview/
EPnP 结合 http://www.cnblogs.com/jian-li/p/5689122.html 和EPnP原始的论文 Accurate Non-Iterative O( n) Solution to the PnP Problem 理解 EPnP的基本思想,很多细节还不是清楚,自己阅读论文时,在论文上做的一些疑问