提出了一种在原始梯度水线划分约束下应用于图象灰度表面的层化变换。经此变换之后，在多尺度测地重建滤波下，不同尺度的梯度水线将存在严格的因果关系。在此基础上能够建立测地重建这种形态滤波器下的梯度水线的尺度空间，以及多尺度计算框架。作为它的一个应用例子，讨论了一种基于区域跟踪的序列图象分析方法。这种多尺度框架能有效提高算法的稳定性、可靠性和计算效率。

In this paper, we propose a stratification transformation on the intensity surface of grayscale image. This transformation is confined by the original watershed partition, and after this transformation, there exists a strict causality between gradient watershed under different scales of geodesic reconstruction. In this way, the scale space of gradient watershed can be established. Some possible applications of watershed transformation on image sequence analysis are discussed, and by the utilization of the multiscale framework of gradient watershed, the efficiency and stability can be greatly improved.