Skip to content
View malhaged's full-sized avatar

Block or report malhaged

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
malhaged/README.md

Hi, I'm Malte 👋

Robotics & AI Engineer | M.Sc. Mechatronic Systems Engineering
Intelligent Agricultural Systems Research Group · Prof. Dr. Stefan Stiene · Hochschule Osnabrück
Osnabrück, Germany


What I'm working on

Building a Real-to-Sim pipeline for autonomous agricultural robotics — from raw LiDAR/Camera/IMU sensor data to simulation-ready Digital Twins in NVIDIA Isaac Sim from a single sensor pass.

FAST-LIVO2 SLAM3D Gaussian Splatting (3DGRUT)USDZ / NuRecNVIDIA Isaac Sim + PhysX

Research topic overview


Featured research

Synthetic Data & Real-to-Sim Pipeline — AgrifoodTEF (2023–present)

Automated generation of high-fidelity Digital Twin simulation environments for autonomous agricultural robots.

  • Multimodal sensor calibration: LiDAR/Camera/IMU, targetless, PTP/PPS synchronisation (Ouster, Allied Vision, Microstrain)
  • FAST-LIVO2 LiDAR-Inertial-Visual SLAM for accurate pose estimation and point cloud generation
  • Neural rendering via 3DGRUT with USDZ export and PhysX integration into NVIDIA Isaac Sim
  • PSNR ≥ 27 dB reconstruction quality, further optimised via LVBA-based framework
  • Dual simulator integration: ROS2-native Gazebo sensor plugin + NVIDIA Isaac Sim via NuRec

Hyperspectral + LiDAR Fusion — Master's Thesis (2022–2023)

Open-source benchtop visual inspection system fusing laser line profiling and hyperspectral imaging to generate synthetic 3D plant training models for ML applications in precision agriculture.

  • Dual Gocator 2350A laser line profilers + Resonon Pika L hyperspectral camera (281 channels, 400–1000 nm)
  • 2.1 nm spectral resolution and 0.019 mm Z-resolution — exceeds state of the art at the time.
  • PSO-based pixel-accurate sensor alignment; Poisson surface reconstruction; .ply / .obj export
  • Implemented in Python with ROS sensor drivers and Open3D

Short thesis documentation


Tech stack

Robotics & Perception
SLAM Sensor Fusion Point Cloud Processing LiDAR / Camera / IMU ROS2 Open3D

3D Reconstruction & Simulation
3D Gaussian Splatting NeRF Novel View Synthesis NVIDIA Isaac Sim Omniverse / USD Gazebo

ML & Computer Vision
PyTorch OpenCV YOLO Python NumPy Pandas C++

AI & Agentic Workflows
Claude API Gemini CLI


Background

M.Sc. Mechatronic Systems Engineering  —  Hochschule Osnabrück
B.Sc. Mechanical Engineering           —  Hochschule Osnabrück
Research: AgrifoodTEF (EU) · Agri-Gaia · DFKI Nature Robots

Contact


Open to roles in Robotics Engineering · AI Engineering · Technical Program Management · DACH region

Pinned Loading

  1. FAST-LIVO2 FAST-LIVO2 Public

    Forked from integralrobotics/FAST-LIVO2

    FAST-LIVO2: Fast, Direct LiDAR-Inertial-Visual Odometry

    C++