The real-time target tracking algorithm based on improved template matching and its hardware implementation

Daqun Li, Jie Guo, Tingfa Xu*

*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

In order to solve the long-time tracking problem in the video stream, this paper proposes a real-time target tracking algorithm based on improved template matching. On the basis of the traditional template matching algorithm, we apply the random forests to generate and train those characteristics. During the process of template matching, we use the improved normalized correlation coefficient to evaluate the similarity. It can also update the real-time template library and ensure the tracking is not lost. Algorithm can achieve a long-term tracking, and able to deal with the complex background, part of the covering, and so on. The algorithm has been applied in the hardware processing platform which uses the FPGA and DSP as the core processing. The result is satisfactory.

Original languageEnglish
Title of host publicationProceedings of the 2015 International Conference on Communications, Signal Processing, and Systems
EditorsJiasong Mu, Wei Wang, Baoju Zhang, Qilian Liang
PublisherSpringer Verlag
Pages531-539
Number of pages9
ISBN (Print)9783662498293
DOIs
Publication statusPublished - 2016
Event4th International Conference on Communications, Signal Processing, and Systems, CSPS 2015 - Chengdu, Sichuan, China
Duration: 23 Oct 201524 Oct 2015

Publication series

NameLecture Notes in Electrical Engineering
Volume386
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

Conference4th International Conference on Communications, Signal Processing, and Systems, CSPS 2015
Country/TerritoryChina
CityChengdu, Sichuan
Period23/10/1524/10/15

Keywords

  • Improved normalized correlation coefficient
  • Random forests
  • Target tracking

Fingerprint

Dive into the research topics of 'The real-time target tracking algorithm based on improved template matching and its hardware implementation'. Together they form a unique fingerprint.

Cite this