Electro-Optic Technology Application, Volume. 26, Issue 6, 75(2011)
Tracking Method of CamShift Sequence Image Based on Incomplete Object
The algorithm can not adjust in real time the shape of the tracking window when the incomplete object is tracked by CamShift algorithm. The background image of the incomplete area is included in the tracking window, and the ability of the object matching is reduced, so CamShift algorithm can not track the incomplete object accurately. A new method of CamShift sequence measurement image is proposed, taking the similarity between the object model and the candidate model as the weight, the probability range of the kernel function is limited, the data volume of data set is reduced and the shape of the tracking window is adjusted in real time. The method can improve the matching similarity and solve the problem of stable tracking incomplete object in the theodolite measurement image. The experimental results show that the method can track steadily even when the loss of the object reaches 50%.
Get Citation
Copy Citation Text
LI Xiao-bing, WANG Yu-mei, ZHANG Hong-yi. Tracking Method of CamShift Sequence Image Based on Incomplete Object[J]. Electro-Optic Technology Application, 2011, 26(6): 75
Category:
Received: Nov. 24, 2011
Accepted: --
Published Online: Jan. 11, 2012
The Author Email: