Abstract—In this paper, we propose combining the mixture color model with Kalman filter method. The purpose is to enable forklifts to search for pallets, but it is able to meet fully automated system with real-time. We focused on pallets for image feature and tracking. First, we manually segmented 30 pallet images and statistics of the best color threshold, this method must find the threshold of different color space and mixture of two important color spaces containing HSV and YCbCr, we extracted the H and the Cb composition mixtures to find the best color threshold, and using a combination of Kalman filter(KF) and the color model method to track pallet images, we then used the logic function to keep our information after obtaining the color image segmentation, the noise of the image must be removed, this algorithm can be used on video sequences efficiently. Finally, experimental results show that the method has effective tracking pallet images in the video sequences.
Index Terms—Kalman filter, color space, object tracking, image detection.
Ssu-Wei Chen, Luke K. Wang, and Jen-Hong Lan are with the
Department of Electrical Engineering of National Kaohsiung University of
Applied Sciences, Kaohsiung, 80778 Taiwan, R.O.C (corresponding author
to provide phone:+886-6-5931668; fax:+886-6-5932265 ; e-mail:
1098404109@cc.kuas.edu.tw, lwang@mail.ee.kuas.edu.tw,
obiwan@adm.cgmh.org.tw ).
Jia-Lin Tu is with the Department of Biomedical Informatics of Asia
University, Taichung, 41354 Taiwan, R.O.C (e-mail:
forushapes@hotmail.com ).
Cite: Ssu-Wei Chen, Luke K. Wang, Jen-Hong Lan, and Jia-Lin Tu, "Combining Kalman Filter with Mixture Color Model Tracking of Pallet Image," International Journal of Information and Education Technology vol. 2, no. 2, pp. 88-93, 2012.