Doctoral Dissertations
Keywords and Phrases
Activity Recognition; Human Tracking; Information Fusion
Abstract
"Human tracking is an extensively researched yet still challenging area in the Computer Vision field, with a wide range of applications such as surveillance and healthcare. People may not be successfully tracked with merely the visual information in challenging cases such as long-term occlusion. Thus, we propose to combine information from other sensors with the surveillance cameras to persistently localize and track humans, which is becoming more promising with the pervasiveness of mobile devices such as cellphones, smart watches and smart glasses embedded with all kinds of sensors including accelerometers, gyroscopes, magnetometers, GPS, WiFi modules and so on. In this thesis, we firstly investigate the application of Inertial Measurement Unit (IMU) from mobile devices to human activity recognition and human tracking, we then develop novel persistent human tracking and indoor localization algorithms by the fusion of non-visual sensors and visual sensors, which not only overcomes the occlusion challenge in visual tracking, but also alleviates the calibration and drift problems in IMU tracking"--Abstract, page iii.
Advisor(s)
Yin, Zhaozheng
Committee Member(s)
Jiang, Wei
Lin, Dan
Cheng, Maggie Xiaoyan
Qin, Ruwen
Department(s)
Computer Science
Degree Name
Ph. D. in Computer Science
Publisher
Missouri University of Science and Technology
Publication Date
Spring 2017
Pagination
xi, 81 pages
Note about bibliography
Includes bibliographic references (pages 74-80).
Rights
© 2017 Wenchao Jiang, All rights reserved.
Document Type
Dissertation - Open Access
File Type
text
Language
English
Thesis Number
T 11097
Electronic OCLC #
992174339
Recommended Citation
Jiang, Wenchao, "Fusion of non-visual and visual sensors for human tracking" (2017). Doctoral Dissertations. 2562.
https://scholarsmine.mst.edu/doctoral_dissertations/2562