ROS学习笔记------ROS机器人系统设计-----基础编程 day 9 2019/3/14 帅某（slam更多概念了解）

版权声明：转载请注释出处，帅某不胜感激。严禁利用本博客进行商业或从事利益方面工作！！ https://blog.csdn.net/weixin_43262513/article/details/88554457

SLAM算法分为三类：Kalman滤波、概率滤波、图优化

博主地址：
https://www.geek-share.com/detail/2653403000.html

Kalman滤波方法包括EKF、EIF；概率滤波包括RBPF，FastSLAM是RBPF滤波器最为成功的实例, 也是应用最为广泛的SLAM方法；

SLAM分为Full SLAM和Online SLAM

常见的二维激光SLAM算法

1、GMapping is a highly efficient Rao-Blackwellized particle filer to learn grid maps from laser range data.

2、Tinyslam又称CoreSLAM
The advantage of DP-SLAM over CoreSLAM is thus the thoretical ability not to be lost in long corridors, and this
is the goal indeed of the map-per-particle concept - not the loop closing which can’t be achieved in DP-SLAM without an external process. As a matter of fact, we decided that this advantage didn’t worth the complexity - especially as we could rely on a good odometry on our platform and given that our goal was to close rather small loops (exploring laboratories instead of corridors…).
As the idea of CoreSLAM was to integrate laser information in our localization subsystem based on particle filter.

3、DPSLAM works by maintaining a joint distribution over robot poses and maps via a particle filter. The algorithm associates a map to each particle, and focuses on the problem of sharing parts of maps among particles in order to minimize memory (and time through map copy). The problem with DP-SLAM is that it is rather complex to integrate into an existing particle filter based localization susbystem

4、Hector-SLAM

5、Karto-SLAM
http://www.zhihu.com/question/35116055/answer/62001013

https://www.geek-share.com/detail/2651906421.html

视觉SLAM算法

1、orbslam 是14-15年被一个西班牙博士做的，目前还在做，最近又发了新文章：Probabilistic Semi-Dense Mapping from Highly Accurate Feature-Based Monocular SLAM。
orbslam可以看做是PTAM的一个延伸。ptam想必做visual slam 的都知道，它是第一个将tracking和mapping分成两个线程实现的实时slam系统，07年出来的时候很惊艳。几乎成立后来feature-based slam方法的标准。orbslam 算是这个思路的延伸，于ptam相比它又加了一个loopclosing的线程。这个系统基于ptam，个人感觉效果也更好一些（毕竟ptam相对较老），整合了covisible graph，基于bagofwords 做relocalization等技术。

常见的一些开源代码（高博整理）：[转载]

• rtabslam https://github.com/introlab/rtabmap_ros#rtabmap_ros
• ORB-slam https://github.com/raulmur/ORB_SLAM
• rgbd-slam-v2 https://github.com/felixendres/rgbdslam_v2
• lsd-slam https://github.com/tum-vision/lsd_slam
• dvo-slam https://github.com/tum-vision/dvo_slam
• hector-slam https://github.com/tu-darmstadt-ros-pkg/hector_slam
• svo https://github.com/uzh-rpg/rpg_svo

SLAM研究体系分类：

SIFT，即尺度不变特征变换（Scale-invariant feature transform，SIFT）

SIFT是用于图像处理领域的一种描述子。这种描述具有尺度不变性，可在图像中检测出关键点，是一种局部特征描述子。
[1] 该方法于1999年由David Lowe
[2] 首先发表于计算机视觉国际会议（International Conference on Computer Vision，ICCV），2004年再次经David Lowe整理完善后发表于International journal of computer vision（IJCV）
[3] 。截止2014年8月，该论文单篇被引次数达25000余次。

点云

Point Cloud Library (PCL) 是一个独立的大型的处理二维/三维图像和点云数据的开源工程，由Willow Garage公司开发，起初只是以PR2机器人在3D数据感知算法处理上速度更快为目标，后来才渐渐发展为一个独立的函数库。PCL包含了许多先进算法，比如滤波，特征估计，表面重建，模型拟合和分割等等。
http://pointclouds.org/

CARMEN - The Carnegie Mellon Robot Navigation Toolkit

CARMEN is an open-source collection of software for mobile robot control.

CARMEN is an open-source collection of software for mobile robot control. CARMEN is modular software designed to provide basic navigation primatives including: module communication infrastructure, base and sensor control, obstacle avoidance, localization, path planning, and mapping.
　　http://carmen.sourceforge.net/

随机采样一致性 （ Random Sample Consensus， RANSAC）

ICP算法（Iterative Closest Point迭代最近点）

ICP（Iterative Closest Point迭代最近点）算法是一种点集对点集配准方法

而ICP是最广泛应用的配准方法，也就是KinectFusion论文中所提到的ICP（ Iterative Closest Point ）， 最近邻迭代算法。icp利用迭代一步步地算出正确对应关系。

Structure from Motion (SfM) 运动恢复结构