Abstract: As a kind of non-linear filter，morphology filter is applied widely in image processing such as object extraction and noise removing etc.The paper proposes a new adaptive approach for selecting the scale of structuring element in binary images，which depends on the research of the element decomposition and morphology filter.The experiments show that the method can select the scale of structuring element exactly，and at the same time remove the noise completely and effectively.

Key words: mathematical morphology, filter, structuring elements

摘要： 形态学滤波作为一种非线性滤波方法在目标提取、抑制噪声等图像处理领域有着广泛的应用。针对形态学滤波中的一个关键问题—结构元的选择，通过对结构元分解和形态学滤波运算的研究，提出了一种自适应选择结构元尺度的二值形态学滤波方法。通过实验表明该方法可以快速、准确地确定滤波中结构元的尺度，具有较强的去噪性和抗干扰性。

关键词: 数学形态学, 滤波, 结构元