Abstract: There are always lots of speckle noises in medical ultrasonic images because of the imaging mechanism.These speckle noises significantly reduce the clarity of the image and as a result，cause great difficulties in ultrasound diagnosis.To solve this problem，this paper presents a new de-noising method for medical ultrasonic image based on ROF model and adaptive wavelet threshold de-noising method.In addition to inhibiting the speckle noises in the medical ultrasonic image，this method can retain and even enhance the image detail information as much as possible.

Key words: speckle noise, frame correlation, Global Phase Coherence（GPC）, ROF model, adaptive wavelet threshold

摘要： 由于超声成像机制使医学超声图像中存在着大量的斑点噪声，这些斑点噪声大大降低了图像的清晰度和质量，给超声诊断带来很大的困难。针对医学超声图像的斑点噪声去噪问题，提出了一种基于帧相关处理、ROF分解和自适应小波阈值的去噪方法，能够在抑制超声图像斑点噪声的同时，尽可能地保留甚至增强图像的细节信息，大大提高图像质量，取得了很好的效果。

关键词: 斑点噪声, 帧相关, 全局相位相干性, ROF模型, 自适应小波阈值