运动目标跟踪在工业过程控制、医学研究、成像制导等领域具有重要的实用价值。目前的研究多基于背景静止的情况，对背景发生移动的情况研究较少。提出了一套完整的移动背景下的目标跟踪算法，首先使用基于互信息的方法配准序列图像的背景。然后使用差分的方法进行运动区域检测，并将其与图像分割技术相结合，得到目标跟踪模板。目标的跟踪基于Kalman滤波的预测，其匹配过程仍基于互信息理论，实验结果证明：该算法具有较高的计算效率和准确性，应用前景广泛。-it is of important practical value to

track moving objects in the fields of industrial process control，medical research，imaging guidance，etc．At present most research in this field are often based on static background，while those based on moving background

ale rather less．A complete algorithm

of object tracking in the moving background

was

proposed．Firstly，the

background

of

sequence images

were

registered by

using

the mutual

information．Secondly，a

differential method

was

used to detect

moving object in combination

with

image segmentation technology

to

get

the

object template．Object

tracking

was based on Kalman filter and the

object matching

was still based on the mutual information

theory．The experiments

show that

object tracking

is

rapid，accurate

and robust

by using

this

algorithm．

Computational efficiency

and

accuracy

of this new

tracking algorithm

is SO

high，and

its

application

prospect

will be broad．