“When the staff used to evaluate the footage of a deployed head airbag with partial automation, they still had to catch the right moment exactly,” explained Misheel, information technology trainee at PSW. “When the airbag is deployed, the aim of the analysis is to pinpoint the time at which the positioning lines overlap.” The semi-automatic method used until now has been costly and inefficient, with the team taking an inordinate amount of time to look through all the test videos and note down the necessary information.
Misheel has now developed a fully automated Python-based solution for colour recognition and object identification in videos. “I used the tool to select the relevant section in the film footage and I arrived at an exact count of the positioning line pixels by means of masking and morphological processes,” she said, outlining her project work. The staff from the Vehicle Safety section now use the solution for all the head airbag analyses. “The automation of the video analysis process marks an important step on the way to increasing the quality of our data and to working even more efficiently. Our team thought it was great that Misheel took the initiative to optimise the process,” said Nijaz Dizdarevic, Head of Full-Width Frontal Crash & Side Impact Protection Development, also happy with the outcome. “The work done by Misheel shows very clearly how important IT specialists are to us in the vehicle development process. Their knowledge is like gold in the move towards process automation and, with it, the achievement of the digital transformation in product development. We are therefore very proud of our home-grown talent and our ability to develop our own IT specialists in the PSW training scheme,” said Markus Dick, Head of Training and Professional Development at PSW.
Misheel is now working on a GUI application for video analysis. This will allow the teams the flexibility to adapt the requirements for various test videos. “It was very exciting to take charge of my own project and also to get a more in-depth knowledge of Python in the process,” added Misheel. “The colleagues from the Vehicle Safety section were always there to help me and explained the various requirements step by step. We had regular meetings with Franziska Thalmair and Jürgen Brantsch to discuss the progress and potential of the solution. So I was encouraged to press on and to contribute many ideas.”
