Code: 2019034
Internship Overview
3D depth cameras are calibrated by the manufacturers. Often, calibration decays with time (e.g., due to usage or physical damage), or they are badly calibrated from the factory. In addition, many manufacturers do not provide software tools to re-calibrate the sensors. Due to all of these problems, miscalibrated sensors can have a negative impact on the result of computer vision algorithms. The goal of this project is to implement a calibration tool based on the OpenNI standard Application Programming Interface (API). The calibration result should be stored in a database. This project can be decomposed into diverse and challenging tasks: ● Devise an algorithm to calibrate a 3D camera depth sensor, based on the RGB color sensor (e.g., using a checkerboard with the size of squares known in advance) ● Analyze the error function of one 3D camera sensor (e.g., Kinect v2), given different distances ● Propose a solution for the RGB-to-depth calibration problem ● Implement the solution using existing libraries (e.g., OpenCV)
Requirements
- Studying towards a degree in Mathematics
- Knowledge of computer processing and image processing
Languages required
EnglishLocation
Martigny, SwitzerlandStarting date
Start during Summer 2019Duration
6+ MonthsType
Full time 40 Hours / weekSalary / Stipend
Minimum 400€/ monthApply now
