Front moving object detection for car collision avoidance applications

  • Yeong Kang Lai
  • , Yao Hsien Huang
  • , Chih Ming Hwang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

6 Scopus citations

Abstract

This paper proposed a technique to avoid collisions in the image sequences captured by stereo camera. It has to filter out the independently moving objects by performing ego-motion computation of the vehicle, so that the front moving objects and possible dangerous conditions may be detected using this collision avoidance system. The stereo computation generated the sparse disparity map of the object, and the feature tracker generated the optical flow. Then, it can obtain the three-dimensional position and three-dimensional motion of independently moving objects. After finding out the independently moving objects, it can estimate the positions and motion directions of objects in next frame. This chip is implemented using 90 nm CMOS process with power consumption 7.85mW@346 MHz. Moreover, the chip can process the 1080P resolution, and its gate count is 12.1k only.

Original languageEnglish
Title of host publication2016 IEEE International Conference on Consumer Electronics, ICCE 2016
EditorsFrancisco J. Bellido, Nicholas C. H. Vun, Carsten Dolar, Daniel Diaz-Sanchez, Wing-Kuen Ling
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages367-368
Number of pages2
ISBN (Electronic)9781467383646
DOIs
StatePublished - 10 03 2016
Externally publishedYes
Event2016 IEEE International Conference on Consumer Electronics, ICCE 2016 - Las Vegas, United States
Duration: 07 01 201611 01 2016

Publication series

Name2016 IEEE International Conference on Consumer Electronics, ICCE 2016

Conference

Conference2016 IEEE International Conference on Consumer Electronics, ICCE 2016
Country/TerritoryUnited States
CityLas Vegas
Period07/01/1611/01/16

Bibliographical note

Publisher Copyright:
© 2016 IEEE.

Fingerprint

Dive into the research topics of 'Front moving object detection for car collision avoidance applications'. Together they form a unique fingerprint.

Cite this