A Study on Software-based Sensing Technology for Multiple Object Control in AR Video

Sensors

Sungmo Jung²  Jae-gu Song²  Dae-Joon Hwang¹  Jae Young Ahn³ 

[1] Sungkyunkwan University of Information & Communication Engineering, Suwon, Korea; E-Mail:;Hannam University of Department of Multimedia, Daejeon, Korea; E-Mails:;ETRI Standards Research Center, Daejeon, Korea; E-Mail:

关键词: software-based sensing;    multiple object detection;    multiple object loading;    multiple object control;    computer vision;    augmented reality;   

DOI  :  10.3390/s101109857

来源: mdpi

【 摘 要 】

Researches on Augmented Reality (AR) have recently received attention. With these, the Machine-to-Machine (M2M) market has started to be active and there are numerous efforts to apply this to real life in all sectors of society. To date, the M2M market has applied the existing marker-based AR technology in entertainment, business and other industries. With the existing marker-based AR technology, a designated object can only be loaded on the screen from one marker and a marker has to be added to load on the screen the same object again. This situation creates a problem where the relevant marker’should be extracted and printed in screen so that loading of the multiple objects is enabled. However, since the distance between markers will not be measured in the process of detecting and copying markers, the markers can be overlapped and thus the objects would not be augmented. To solve this problem, a circle having the longest radius needs to be created from a focal point of a marker to be copied, so that no object is copied within the confines of the circle. In this paper, software-based sensing technology for multiple object detection and loading using PPHT has been developed and overlapping marker control according to multiple object control has been studied using the Bresenham and Mean Shift algorithms.

【 授权许可】

CC BY   
© 2010 by the authors; licensee MDPI, Basel, Switzerland.