https://github.com/lukasvst/dm-viohttps://github.com/lukasvst/dm-vio-ros 2、编译 编译主要有两点需要注意: 1)pangolin的版本需要是0.6; 由于我系统里面的版本不是0.6,为了避免冲突,我编译到指定目录下,然后在CMakeLists.txt中添加位置; a、cmake -DCMAKE_INSTALL_PREFIX=/xxx/x ..指定安装位置; b、...
For more information seehttps://vision.in.tum.de/dm-vioandhttps://github.com/lukasvst/dm-vio. This is a ROS-Wrapper for DM-VIO, inspired by theROS-Wrapper for DSO. It interfaces with ROS topics to input images and IMU data and to output poses. Alternatively, it can read images and ...
git clone https://github.com/lukasvst/dm-vio.git The following instructions have been tested with Ubuntu 20.04. The system is also known to work well on Ubuntu 16.04, 18.04 and MacOS Big Sur (only Intel Macs have been tested so far)....
https://github.com/lukasvst/dm-vio https://github.com/lukasvst/dm-vio-ros 1. 2. 2、编译 编译主要有两点需要注意: 1)pangolin的版本需要是0.6; 由于我系统里面的版本不是0.6,为了避免冲突,我编译到指定目录下,然后在CMakeLists.txt中添加位置; a、cmake -DCMAKE_INSTALL_PREFIX=/xxx/x ..指定安装...
要了解DM-VIO,必须得了解DSO和工大TUM的历史,这个Github和公开资料上很多不赘述了。重点讲Lukas组。 DM-VIO是VI-DSO的延续,VI-DSO因为实际使用效果较差,国内各组在跟的应该比较少。 但实际动态边缘化是一种非常精妙的思路,我也不知道这几年发生什么了(疫情?),最终L组并没有在这条路上一路走到黑,而是出了DM...
开源地址: https://github.com/lukasvst/dm-vio 录用:2022 Robotics and Automation Letters (RA-L) DM-VIO是一种基于延迟边缘化和位姿图BA的单目视觉惯性里程计。该系统单目VIO超过了sota 方法,甚至在某些情况下超过了性能最好的双目VIO。 延迟边缘化是啥?先说说边缘化,它常用于保证系统更新时间,但边缘化后很...
项目主页: http://vision.in.tum.de/dm-vio 即将开源: https://github.com/lukasvst/dm-vio 论文名称:DM-VIO: Delayed Marginalization Visual-Inertial Odometry 一种基于延迟边缘化和位姿图束调整两种新技术的单目视觉惯性里程计系统。在包括无人机、大型手持设备和汽车场景数据集上评估该系统。由于提出的IMU初始...
git clone https://github.com/lukasvst/dm-vio.git The following instructions have been tested with Ubuntu 20.04. The system is also known to work well on Ubuntu 16.04, 18.04 and MacOS Big Sur (only Intel Macs have been tested so far). ...
master Breadcrumbs dm-vio / Latest commit History History File metadata and controls 207 lines (176 loc) · 7.01 KB Raw 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 ...
GitHub is where people build software. More than 100 million people use GitHub to discover, fork, and contribute to over 420 million projects.