The Improvement of Mean - Shift Algorithm and Kalman Filter of Tracking Moving Targets in the video of Robotic Fish
- Jingying Wu, Xiang Wei
- Corresponding Author
- Jingying Wu
Available Online December 2015.
- https://doi.org/10.2991/icmmcce-15.2015.89How to use a DOI?
- Mean shift algorithm; Kalman residuals; Kalman filter; Target tracking
- Several defects have been found in the former experiments of basic Mean shift algorithm, such as the erroneous judgments when target blocked by a large proportion of barrier. Thus, a fast target tracking algorithm based on Mean shift combined Kalman filter is proposed for the theoretical defect of basic Mean shift. The blocking issue and tracking of fast moving target are discussed in this paper and applied in the independent visual robotic fish tracking. Here are our conceptions: Firstly, the initial position of Mean shift is provided with Kalman filter in every frame, then Mean shift algorithm is used to track the position of target. Secondly, the calculation of the Kalman residuals is applied to turn on and turn off the Kalman filter under a large proportion of blocking situation, at the same time, the linear prediction of the target position is replaced by Kalman's function test. Finally, we conduct a real-time tracking experiment on the independent visual robotic fish tracking a moving ball, it is proved that the algorithm can achieve the tracking of fast moving target and robust against the barrier as well.
- Open Access
- This is an open access article distributed under the CC BY-NC license.
Cite this article
TY - CONF AU - Jingying Wu AU - Xiang Wei PY - 2015/12 DA - 2015/12 TI - The Improvement of Mean - Shift Algorithm and Kalman Filter of Tracking Moving Targets in the video of Robotic Fish BT - 2015 4th International Conference on Mechatronics, Materials, Chemistry and Computer Engineering PB - Atlantis Press UR - https://doi.org/10.2991/icmmcce-15.2015.89 DO - https://doi.org/10.2991/icmmcce-15.2015.89 ID - Wu2015/12 ER -