Vision-based Automatic Trunk Opener for Automotive Electronics

摘要

This paper applies a 180° fisheye automotive camera to design an automatic trunk opener. Comparing to the commonly used infrared sensor, we adopt the vision based image sensor to share the automotive tail camera for reducing the cost of the system. We detect the kicking to the trunk behavior of drivers or passengers by temporal differencing method and Harris edge detection algorithm. If the behavior is detected, the system will automatically open the trunk and flash the taillights for indication. The system was integrated with cameras, software programs and embedded processors on an embedded system and realized for all-weather test. The detection rates was 98.70%, 97.39%, and 98.52% in the day, rain day and night respectively.