ON-LINE VIDEO PROCESSING EMBEDDED SYSTEM FOR AUTOMOTIVE

Abstract

Real-time processing for images is increasingly used in automotive industry. Video signals are nowadays used for different applications on traffic, such as measuring average speed, monitoring traffic congestion, detect red light crossing, but also for sign recognition, lane assist etc. The paper proposes an embedded system for image processing used for controlling climatization in cars. This system is also useful in case of an accident to activate the airbag if a person is on the seat or not to do it if a heavier object is placed on the seat, for example a bag. For this aim, a Raspberry Pi3 based system is proposed for real-time detecting of the number of passengers existing in a vehicle. The Raspberry Pi3 is used for acquiring frames from a video camera connected via USB interface and for on-line processing of images. The system uses an Haar cascade classifier algorithm (HCCA) for human face detection.