Visual Odometry / SLAM Implementations by MODULabs studies

Authors:

• Hyunggi Chang, @changh95 (study leader)
• Sangjin Lee, @LeeSSangJin
• Eunsung Yang, @EunSungYang
• Sunho Kim, @Sunho-Kim
• Jisu Han, @lazyshuu
• Hyejin Hwang, @brillianthhj
• Sangjun Lee, @eowjd0512
• Minwoo Shin, @minooisbusy
• Sukmin Ha, @hasukmin12
• Joo Yong Sim, @jooyongsim
• Jongsik Moon, @jongsik-moon

Background

This repository contains links to several repositories, which each repo contains individual efforts of implementing Visual Odometry (VO) or Visual-SLAM programs. The authors of these repositories are the participants of the '슬기로운 SLAM 스터디' ('Wise SLAM study' in Korean) led by Hyunggi Chang with collaboration of MODU Labs. Each participants aims to build computer vision applications to the best quality as they can, based on their skills and confidence.

The suggested levels of implementation are as below:

• Newbie: Build a marker-based localization system
• Easy: Build a simple visual odometry system
• Medium: Build a visual odometry system, with bundle adjustment or factor graph representations
• Hard: Build a complex visual odometry or SLAM syste of your choice (recommended to implement loop closure, or deep-learning based frontend, or integrate IMU data)

The participant may freely choose the hardware or the dataset to be used for development and evaluation.

List of SLAM Implementation projects

• Hyunggi Chang - AirSim Dataset Generator
• Hyunggi Chang - BArUco: Marker based camera localization, but better than ArUco
• Sangjin Lee - Project name undetermined
• Sunho Kim - SH-SLAM
• Hyejin Hwang - Project name undetermined
• Sangjun Lee - VITAMIN-E re-implementation
• Minwoo Shin - GOSLAM
• Sukmin Ha - Has-SLAM
• Joo Yong Sim - Project name undetermined
• Jongsik Moon - DVL-SLAM re-implementation