Martin Oelsch

Background

About Martin Oelsch

In August 2016 I initially started working on the "VaMEx-Vipe" project at the Chair of Media Technology, Technical University of Munich.The project dealt with the development of robotic platforms for the autonomous exploration of the Valles Marineris on Mars. The focus of my work for the project's duration until 2018, was the evaluation of suitable visual feature extraction algorithms from 2D images for visual SLAM and content-based image retrieval.2018 to 2022 I was working on the project "AI inspection drone". It is a joint project headed by Lufthansa Technik to develop an autonomous UAV for visual defect inspection of airplanes. My major contributions to the project are software developments on LiDAR data processing, LiDAR-based localization and mapping (LiDAR-SLAM), 3D reconstruction, and trajectory evaluation with ground truth. Sensor data is generated from either a Gazebo simulation environment or real LiDAR sensors, such as Velodyne or Ouster.I have extensive experience with the Robot Operating System (ROS) and C++ libraries such as OpenCV, Point Cloud Library (PCL) or Eigen, and created libraries and applications for LiDAR-SLAM during my time as PhD candidate.I created Python scripts for quick data evaluations and performed most of the graph visualizations in Matlab.