论文部分内容阅读
提出了一种基于差分图像运动检测和轮廓提取的跟踪与分割图像的方法。该方法首先从初始图像中提取目标的轮廓,然后利用相邻帧之间的差分图像初步确定目标在每帧图像中的粗略位置,最后把从上一帧图像中得到的目标轮廓置于该位置,并作为轮廓提取的初始值,由此可得到对目标的准确分割。
A method of tracking and segmenting images based on differential image motion detection and contour extraction is proposed. The method first extracts the outline of the target from the initial image, and then uses the difference image between the adjacent frames to determine the rough position of the target in each frame image. Finally, the target contour from the previous frame image is placed in the position , And extracted as the initial value of the contour, which can be accurate target segmentation.