为了提取血管扩张收缩的变化趋势，在采集原始视频AVI图像，进行帧序列转换后，提取血管轮廓成为研究

的重点。文中提出了一种基于标记分水岭对血管超声帧图像的分割方法，该法针对超声图像对比度低、噪声强的特点在标记

之前进行小波域增强。实验证明该法较传统的区域生长法更能准确地提取目标轮廓。基于该法得到的血管在连续时间内截面

积变化曲线，更方便医师进行精确的病理分析。由此可见，该法更适用于超声血管图像的边界提取。

-In order to extract the trend of contraction of vascular expansion, in the acquisition of the original video AVI images were converted fr a me sequence extracted from a focus of the study of vascular contours. In this paper, based on markers of vascular ultrasound fr a me watershed image segmentation method, the method for ultrasound image contrast is low, the characteristics of strong noise in the wavelet domain before marked increase. Experiments show that the method is more traditional region growing method to contour more accurately. The law is based on the blood vessels in the continuous-time cross-sectional area within the curve, more convenient and accurate pathological analysis of physicians. Thus, the method is more suitable for vascular ultrasound image boundary extraction.