Jinge Wang
Jinge Wang
According to this [paper](https://www.cs.auckland.ac.nz/courses/compsci773s1c/resources/IJCV-Review.pdf), we should "reverse the roles of F1 and F2" instead of reversing the roles of I1 and I2. In this circumstance, this [line](https://github.com/JiawangBian/FM-Bench/blob/master/Evaluation/ComputeNormlizedSGD.m#L21) should be ```...
Hope so!
1. `triangulation`函数的前提是已知相机在两个时刻的相对位姿,也就是参数`R`、`t`。`T1`是单位矩阵,代表相机在第一个时刻的位姿。`T2`是由`R`、`t`构造的矩阵,代表相机在第二个时刻的位姿。`K`代表相机的参数矩阵,也就是投影矩阵。 2. 文件中的所有`K`都代表相机的投影矩阵。该示例中,只有一个单目相机。 3. point in the first camera frame: 第一张图片上特征点的归一化相机坐标 point projected from 3D 3D点在第一张图上的归一化相机坐标, d=3D点相对于第一张图的深度 point in the second camera frame: 第二张图片上特征点的归一化相机坐标 point reprojected from second frame:...
首先,K不是点,K是相机内参,每个相机的内参都是固定的,这与相机观测到的点没有关系。 其次,通过Camera frame计算3D坐标,你指的是计算相机的位姿还是计算特征点的3D位置呢? 如果是计算相机位姿,请看这篇文章http://www.jianshu.com/p/fbf56587a268 如果是计算特征点的3D位置,请看这篇文章http://www.jianshu.com/p/96d3b832330e
你说的是哪个程序?这里面有好几个程序的。
每帧图像的参数是指什么? 最后输出的T是相邻两帧图像之间的转换矩阵。
我是在Ubuntu下跑的。我不知道你的g2o是怎么安装的,我是下载的g2o源码编译安装的,按照这里面的要求安装试一下https://github.com/RainerKuemmerle/g2o
I also encountered the same error.
But I can't reproduce this error. Maybe due to some incomplete installation?
@mikeferguson I think the code works, but one of the checks still failed. Looks like some process out of time. I was confused. Can you help me check the reason...