Skip to content
forked from TriLoo/HR-SLAM

Heterogeneous computing based real-time SLAM.

Notifications You must be signed in to change notification settings

Allenwutao/HR-SLAM

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

My Flying HR-SLAM

Description:

This folder includes the implementation HR-SLAM (Heterogeneous Computing based Real-time SLAM), with the purpose of involving the start-of-the-art developments in Computer Vision.

License:

GLP-v2

Folders

  • PreFont: 简易前端
  • BackEnd: 仅包含位姿图
  • LearningApproaches: 包含一些深度学习相关的方法
  • Tools: 包含一些零碎的、自己实现的代码

Current Work :octocat:

  • /LearningApproaches/InstanceSegmentationRGBD

    • Read Dataset

      使用CSVIter来产生DataIter.

    • RGBD Data Fusion

      Multi-layer, Multi-Scale Attention Based middle + late + early fusions.

    • 不知道怎么处理Instance Level ?

    • 该使用什么损失函数 ? 还是交叉熵 ?

Refereneces:

  • 视觉SLAM十四讲
  • Deep Learning
  • 统计学习方法
  • Multiple View Geometry in Computer Vision (Reading)
  • 计算机视觉 模型、学习和推理
  • 计算机视觉----一种现代方法
  • 数字图像处理与机器视觉
  • CUDA by Example
  • Programming Massively Parallel Processors
  • CUDA Best C Practice
  • CUDA C Programming
  • C++ Primer
  • Effective C++
  • C++编程规范
  • 并行程序设计导论
  • Algorithm
  • Mxnet
  • ORB-SLAM2
  • Papers
  • DOing
    • TensorFlow实战
  • TODO
    • 深度探索C++对象模型
    • 深入理解计算机系统
    • State Estimation for Robotics

Contact Me:

Any tutorials or suggestions can be sent to me by e-mail: HR-SLAM

About

Heterogeneous computing based real-time SLAM.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Makefile 35.8%
  • C++ 19.9%
  • CMake 17.8%
  • Python 16.6%
  • C 8.1%
  • Cuda 1.7%
  • Shell 0.1%