Visible Light Dynamic Positioning Method Using Improved Camshift-Kalman Algorithm

• Published in: IEEE photonics journal Vol. 11; no. 6; pp. 1 - 22
• Main Authors: Guan, Weipeng, Liu, Zhipeng, Wen, Shangsheng, Xie, Hongyun, Zhang, Xingjie
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.12.2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Algorithms; Bhattacharyya coefficient; Camshift algorithm; image sensor (IS); Image sensors; Kalman filter; Kalman filters; Light emitting diodes; Real time; real-time ability; Real-time systems; Robustness; Target tracking; Visible light positioning (VLP)
• ISSN: 1943-0655; 1943-0647; 1943-0647
• DOI: 10.1109/JPHOT.2019.2944080

There are three critical elements in the visible light positioning (VLP) system: Positioning accuracy, real-time ability and robustness. However, few existing VLP studies consider these three critical elements at the same time. Especially, robustness is usually ignored in VLP system, which has a great influence on positioning performance or even leads to the failure of positioning. Therefore, we propose a novel VLP method based on image sensor (as positioning terminal), using improved Camshift-Kalman algorithm. The proposed algorithm not only combines Camshift algorithm with Kalman filter, but also introduces the Bhattacharyya coefficient innovatively. It can realize high positioning accuracy, good robustness and good real-time ability. Experiments showed that the positioning accuracy of our proposed VLP algorithm was 0.55 cm, which realized high positioning accuracy. Besides, the average processing time per frame was 27.2 ms, which realized good real-time ability. Also, whether the LED was shielded, the interference occurred or the target's velocity changed suddenly, the proposed algorithm still maintained good positioning effect, which showed that it had good robustness.