论文部分内容阅读
图像光流的计算不需要在图像序列中建立特征之间的对应关系,但是直接利用光流法进行目标的运动参数估计和跟踪运算量很大,而且受噪声等影响而比较困难。首先采用聚类分割方法提取出目标区域,然后对目标区计算光流场,得到目标运动速度估计值,实现目标跟踪。此方法对比较复杂的红外图像也可以进行较好的分割,跟踪时无需建立图像间的特征对应关系,计算量大大缩小,目标运动速度的估计误差也很小。
The calculation of image optical flow does not need to establish the correspondence between the features in the image sequence. However, it is difficult to directly estimate and track the motion parameters of the target using the optical flow method, and is affected by noise and the like. Firstly, the target area is extracted by clustering segmentation method, and then the optical flow field is calculated for the target area to obtain the target velocity estimation to achieve the target tracking. In this method, the relatively complex infrared images can be well segmented. The tracking does not need to establish the correspondence between the features of the images, the calculation amount is greatly reduced, and the estimation error of the target motion speed is also small.