Dashcams are small, dashboard mounted camera systems that continuously monitor the area around a vehicle and record video images. During the recent years they also became popular in Europe. While dashcams bring many advantages, such as providing evidence in the event of a traffic accident or as a protection against insurance fraud, they also raise important privacy issues.
Because dashcam users operate surveillance equipment in public places, several data protection authorities raised concerns about those systems and marked them to be violating fundamental personal rights. Vendors have already tried to mitigate these concerns by introducing so-called crashcams, which only store video images upon being triggered by an integrated acceleration sensor. However, in Austria the Federal Administrative Court still found that crashcams illegitimately collect personal data, in particular pedestrians and license plates of other cars.
Approaches from the research community aim to detect such privacy-sensitive image areas based on video analytics and to store them in an anonymized fashion, for example by means of blurring detected pedestrians and license plates. However, even in the mid-term computation power of dashcams will not be sufficient to execute state-of-the-art computer vision algorithms. Fraunhofer IOSB therefore introduced a solution, which securely delegates the task of video anonymization to a separate system with more computational resources. This solution requires a storage medium (SD card), which immediately encrypts video data while writing to the storage medium. The storage medium can only be accessed from within a protected system, which ensures that data originating from this medium is anonymized before being accessed by a user.
By this means, dashcams become more privacy-friendly - they can be used in public traffic while at the same time ensuring the protection of personal data of any road user.