url-kaist / uv-slam Goto Github PK
View Code? Open in Web Editor NEWUV-SLAM: Unconstrained Line-based SLAM Using Vanishing Points for Structural Mapping
License: GNU General Public License v3.0
UV-SLAM: Unconstrained Line-based SLAM Using Vanishing Points for Structural Mapping
License: GNU General Public License v3.0
hi~ what is the version of EIGEN
I find that you use auto differentiation provided by ceres to realize line factor. But you don't consider the correctness of pose local Parameterization. Because in imu and projection cost function, VINS-Mono directly solves error jacobian w.r.t pose local par parameterization. However, the jacobian of auto differentiation is w.r.t pose full parameterization.
res sir,
will the source code be released.
thank you
Hi! My friend and I are very new to ROS and are trying to follow the instruction of launching vins_estimator by
roslaunch uv_slam euroc.launch
However, the progress is stuck at the point of "waiting for img and imu..."
Do I need to install VINS-Mono to have a successful roslaunch on vins_estimator? Or what could I do to solve this issue?
Thank you!
Hi! I have to say that this work is amazing and I am pretty impressed :)
However, I have two questions:
Hi,
I wonder which CMakeLists.txt file I can edit to set the version as C++17 because the ceres solver I am using requires C++17 for a successful build of your repo.
Thank you very much for your help!
Thanks for your great work! But when I test your algorithm with default settings on Euroc datasets, the results is largely worse than the paper shows. Also when I test it on my self-collected datasets where Vins-mono runs well, I found UV-SLAM do not perform well.
Thank you for opening up your work! But when I tested on the TUM-VI dataset, I found that there was no way to extract line features. I can confirm that there is no problem with the launch parameter settings. What is the reason for this?
Hello,
Thanks for your contribution!
The Jacobian of the reprojection error of the line w.r.t. pose is be automatically derived. But in my opinion, If we use the automatic derivation of ceres, the corresponding
line reprojection Jacobian matrix with respect to the pose is obtained in terms of imu pose should be the global parameters space(4+3, quaternion + trans). In other words, the global parameter dimension of the quaternion should be 4, while the actual optimization space should be 3. Therefore, if the quaternion is automatically derived, then in the pose_local_parameterzation.cpp
, the jacobian w.r.t. quaternion in terms of theta should not be the identity matrix.
bool PoseLocalParameterization::ComputeJacobian(const double *x, double *jacobian) const
{
Eigen::Map<Eigen::Matrix<double, 7, 6, Eigen::RowMajor>> j(jacobian);
j.topRows<6>().setIdentity();
j.bottomRows<1>().setZero();
return true;
}
Best wishes.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.