Junyi Hou

About

Junyi Hou is a robotics researcher whose work centers on simultaneous localization and mapping (SLAM) and real-time 3D reconstruction, with a particular focus on enabling intelligent robots to operate reliably in dynamic and challenging environments. Hou’s most impactful contribution is the development of a robust dynamic-feature-segmentation SLAM algorithm for RGB-D cameras, published in 2023 and already garnering 36 citations. This work directly addresses the critical problem of mapping in environments with moving objects, a key hurdle for autonomous navigation. Complementing this, Hou has advanced large-scale scene reconstruction through a branch-and-bound optimization method (14 citations) and pioneered high-precision localization using ground texture and binary descriptors (7 citations). Further innovations include a real-time reconstruction system leveraging a multi-task feature extraction network and surfel representation, as well as a self-supervised stereo inertial odometry system. Collectively, Hou’s research bridges the gap between robust feature extraction and dense, high-quality 3D modeling, pushing the boundaries of what autonomous systems can perceive and map in real time.

Research Focus

Key Achievements

Top Papers

1. 1
   Optimization RGB-D 3-D Reconstruction Algorithm Based on Dynamic SLAM

   36 citations · 2023
2. 2
   A high-quality voxel 3D reconstruction system for large scenes based on the branch and bound method

   14 citations · 2022
3. 3
   High-precision offline mapping and localization system based on ground texture with binary descriptors

   7 citations · 2023
4. 4
   Realtime 3D reconstruction algorithm based on improved AKAZE and multi-layer feature detection network

   5 citations · 2024
5. 5
   Real-time 3D reconstruction system using multi-task feature extraction network and surfel

   2 citations · 2021
6. 6
   Robust stereo inertial odometry based on self-supervised feature points

   2 citations · 2022

Key Collaborators

Contact & Links