CSE498, Collaborative Design, Fall 2020
Computer Science and Engineering
Michigan State University

Bosch is a global engineering and technology company with products sold in 150 countries worldwide. Founded in Germany in 1886, Bosch is the world’s leading supplier of automotive components.

In the testing and development of advanced driver assistance systems, Bosch collects thousands of hours of video footage and radar data from their vehicles. This data is later used to simulate tests on newly developed software without the need to send a vehicle out into the field.

Certain vehicle software is affected differently by specific environments, so for any new software, it is important for Bosch to conduct a range of tests on these various environments. However, Bosch employees must first manually search through their immense volume of data to locate the proper footage, which can be time- consuming and difficult.

Our Automated Retrieval of ADAS Driving Environments alleviates this issue by automating the task of processing and searching through the video data.

Using computer vision and machine learning, footage is tagged and categorized based on features in the environment (such as rain, night, bridge, highway, etc). Once the videos have been processed, the tags and associated videos are saved for later searching.

Bosch employees can then search for videos containing specific conditions simply by selecting the desired tags and clicking ‘search.’

Our software saves Bosch employees time by automatically tagging video environments and provides easy searching through large quantities of video.

Our front-end web application is built with React.js and Material-UI. The back end is built with Flask and utilizes the OpenCV library and YOLOv3 algorithm. The SQLite database holding the processed video and tags is built with SQLAlchemy.