VINS-Mono result display

The previous article was simply writtenVINS-MonoThe technical framework on which this framework depends,This article briefly shows its results,And several ways of system installation.

在BStation I found a demo work for this video

Used a camera and a cheap oneIMUA synthetic navigation system

这个是可视化IMUThe curve changes of several axes

这个也是,红色的线是IMU,The green line is the camera

This is the display of a trajectory in the office of the Hong Kong University of Science and Technology

A navigation within the library

This is an estimate of the feature points

This is the result of the mapping,看起来效果还是可以的

This is a navigation result outdoors

The perimeter of the drawn line is 642M,See for yourself the range of heights,And the top speed is8.9

The use of this hardware isA3的飞控+An ordinary camera

This is an outdoor oneARThe result of a wraparound mapping in the case of image creation

This is the directory of the code

There is a detailed dissertation in it,虽然50Lots of pages but great

Next, I will write about the role of some folders in this code directory：

ar_demo

benchmark_publisher 发布数据集中参考值

config 硬件、系统配置文件

camera_model

calib

camera_models 各类相机模型

chessboard 用于检测棋盘格特征点

gpl

sparse_graph

intrinsic_calib.cc 相机矫正模块主函数

feature_trackers

feature_tracker_node(main()函数,ROS接受图像的回调函数)

feature_tracker.c 特征点跟踪的具体实现

support_files

vins_estimator

src

factor 实现IMU、camera等残差模型,涉及了ceres优化库,Jacobian矩阵.

initial 系统初始化,外参标定,SFM

loop_closure 闭环检测,这里主要是利用DBOW2作者的一个demo程序

utility 相机显示,四元数等数据转换

estimator_node(main()函数)

多线程 measurement_process、loop_detection、 pose_graph

feature_manager.cpp 特征点管理,三角化,关键帧操作

parameters.cpp 外部系统设置参数输入

Check the directory yourself,If you are worried, you can download the source code of this HKUST

This is a few different ways to install the system

make a wish here,I hope to study at HKUST~