钙化点是乳腺癌早期的一个主要放射学征象。为了实现数字乳腺图像中钙化点的自动检测，提出了一种基于Top—hat算子的钙化点检测方法。该方法采用形态学中的Top—hat算子对图像的背景进行抑制，然后结合孤立钙化点的灰度、纹理和对比度等特征对乳腺图像中的钙化点进行检测。与一般的检测方法相比，这种方法能够有效地检测到强背景中的弱小钙化点目标，而且检测结果贴近于钙化点病灶的真实形状。

Microcalcifications are the primary radiological indicator of early breast cancers. A method for the automatic detection of microcalcifications in digital mammograms is presented in this paper. The algorithm based on top-hat morphologic operator is developed to restrain the background, then the features of single microcalcifications such as gray, texture, and contrast, etc. are extracted to detect the microcalcifications in mammograms. The proposed method is based on the Top-hat operator. Comparing to other approaches, this method can more efficiently detect weak microcaleifications in the high background, and the detected results are much close to the actual shapes of the calcifications.