The method presented here is able recognise a person while he gets into the car. But the recognition results are still far from 100 %. One limitation is the currently used CCD sensor. In scenes containing strong light and deep shadows, the camera is either saturated in bright areas of the scene, or loses detail in the dark areas. CCD sensors have disadvantages like limited dynamic range (< 70 dB) and blooming and smearing effects. Better results can be obtained with CMOS image sensors. In contrast to the CCD technology they provide features like high dynamic range (up to 120 dB), random pixel access and no blooming and smearing effects. An other extension from which a performance increase is expected, is the exploitation of several frames of the image sequence for recognition.
Video based driver and passenger monitoring
1999
13 Seiten, 5 Bilder, 6 Quellen
Conference paper
English
Bridging the communication gap: A driver-passenger video link
Tema Archive | 2009
|Automobile driver and passenger sitting posture monitoring system
European Patent Office | 2021
|ONBOARD ACOUSTIC MONITORING SYSTEM FOR DRIVER AND PASSENGER
European Patent Office | 2022
|Visual Monitoring of Driver and Passenger Control Panel Interactions
Online Contents | 2016
|Visual Monitoring of Driver and Passenger Control Panel Interactions
Online Contents | 2017
|