给出了一种对光照变化等实际环境中干扰不敏感的基于水平集和背景减的运动目标检测与分割方法。
Based on level set method and background subtraction, this paper proposed an approach to (realize) robust detection and segmentation of multiple motion objects.
该文借助知识理论，提出一种新的基于子图像块的区域分割与边缘检测相结合的目标提取方法。
Firstly, we coarsely detect the feature image obtained from subimage blocks, the target regions are identified according to prior knowledge.
一般采用差分法作为多目标检测与跟踪的初始分割，或者与其他方法结合使用。
Difference method can also be used for initial segmentation in multi-object detection and tracking or be integrated with other methods.
应用推荐