Title :
Detecting moving objects from slowly moving viewpoint for automotive rear view camera systems
Author :
Kwon, Joonsoo ; Kim, Do-Sung
Author_Institution :
Korea Electron. Technol. Inst., Seongnam, South Korea
Abstract :
Traditionally, moving object can be detected by using either background subtraction or optical flow methods. In most cases of moving camera, those methods introduce a lot of false-positives due to the stationary viewpoint assumption. In this paper, we suggest a method for moving object detection working with slowly moving background, specifically for the case of a vehicle´s backing up scenario. We implement the method and the detection result from a sequence is compared with an optical flow method.
Keywords :
automobiles; cameras; image motion analysis; image sequences; object detection; traffic engineering computing; automotive rear view camera systems; background subtraction; moving object detection; optical flow methods; slowly moving viewpoint; stationary viewpoint assumption; Adaptive optics; Computer vision; Image motion analysis; Object detection; Optical filters; Optical imaging; Optical saturation;
Conference_Titel :
Consumer Electronics (ICCE), 2014 IEEE International Conference on
Conference_Location :
Las Vegas, NV
Print_ISBN :
978-1-4799-1290-2
DOI :
10.1109/ICCE.2014.6775903