Vision and robotics consultant
self employed
Skills
Computer vision, python, C ++, pybind11, point clouds, 3D model, algorithms, target platform, static analysis, 3D reconstruction, acquisition of images, CAN-bus, sensor, open source, machine learning, image processing, 3D vision, Software Development, C ++, C, Matlab, Shell script, C #, OpenCV, ROS, PCL, QT, UI, Linux, Windows, Microsoft, Visual Studio, VS Code, Unity 3D, CMake, GIT, SVN, Subversion, Jira, Bugzilla, Doxygen, Valgrind, Jenkins, Continuous Integration, Version Control
Projekthistorie
01/2017
-
bis jetzt
- Create python wrapper for library to process raw data from Time of Flight cameras using modern C++ and pybind11.
- Develop software for stitching point clouds to create a highly accurate 3D model in few seconds.
- Implement new algorithms for egomotion estimation, self localization and 3D environment reconstruction.
- Support porting to target platform with static analysis, profiling and pipeline optimization by 3X.
- Develop new features to improve robustness, code maintenance and hardware-portability of a camera based auto parking system.
01/2016
-
12/2016
Computer vision engineer
Hyundai Mobis Parts Europe
* Design and prototype optical flow, visual odometry, moving objects segmentation and 3D reconstruction algorithms.
* Integrate with other modules to boost the efficiency and robustness of pedestrian detector, vehicle detector and tracking of objects.
* Create a low cost data-logger proof of concept for synchronized acquisition of images and CAN-bus information.
* Responsible for collaboration projects between research institutions as well as supervision of master thesis.
* Integrate with other modules to boost the efficiency and robustness of pedestrian detector, vehicle detector and tracking of objects.
* Create a low cost data-logger proof of concept for synchronized acquisition of images and CAN-bus information.
* Responsible for collaboration projects between research institutions as well as supervision of master thesis.
08/2014
-
12/2015
Research assistant
Computer Vision
Linköping University + Scania (Sweden)
* Research and develop software modules for egomotion estimation using direct image alignment methods.
* Install, interface, integrate hardware and software for a prototype truck for camera based surround sensing.
* Author of two scientific papers: Estimation of Automotive Pitch, Yaw, and Roll using Enhanced Phase Correlation on Multiple Farfield
Windows and Motion priors estimation for robust matching initialization in automotive applications.
* Research and develop software modules for egomotion estimation using direct image alignment methods.
* Install, interface, integrate hardware and software for a prototype truck for camera based surround sensing.
* Author of two scientific papers: Estimation of Automotive Pitch, Yaw, and Roll using Enhanced Phase Correlation on Multiple Farfield
Windows and Motion priors estimation for robust matching initialization in automotive applications.
03/2014
-
07/2014
Intern as Software Developer
iRobot
* Develop a method to detect and warn the user when a sensor is occluded or is malfunctioning [co-inventor of 2 patents].
* Create and evaluate system to calibrate the sensors of the Roomba 900 series vacuum cleaning robot based on visual SLAM.
* Extend an open source project for evaluating the path followed by the robot with multiple external cameras.
* Create and evaluate system to calibrate the sensors of the Roomba 900 series vacuum cleaning robot based on visual SLAM.
* Extend an open source project for evaluating the path followed by the robot with multiple external cameras.
03/2013
-
05/2013
Intern as Computer Vision Engineer
Tomra Sorting Solutions
* Analyse 3D sensors for new sorting applications in recycling and food processing industry.
* Design machine learning based segmentation and classification algorithms using 2D and 3D information.
* Acquire and label a dataset for training and evaluating classification algorithms.
* Design machine learning based segmentation and classification algorithms using 2D and 3D information.
* Acquire and label a dataset for training and evaluating classification algorithms.
07/2012
-
09/2012
Intern as Computer Vision Engineer
Freie Universität
* Develop a real-time application to detect and map traffic lights using a camera for an autonomous vehicle.
Reisebereitschaft
Verfügbar in den Ländern
Deutschland
