Abstract: In order to detect lung nodules in chest CT images，an automatic recognition lung nodules system using rules is proposed.In the system，automatic threshold and contour tracking method are adopted to segment lung parenchyma at first;OTSU algorithm is used to extract Region Of Interested（ROI） subsequently，and then the features of ROI are extracted;at last some features are chosen which can distinguish between lung nodules and blood vessels.Rules are defined to determine candidate nodules based on the chosen features.Experiment results show that this system has better recognition performance with the diameter above 1 cm.

Key words: medical image, lung nodule, based on rules, recognition, OTSU algorithm

摘要： 为了检测胸部CT图像中的肺结节，提出一种基于应用规则的自动识别肺结节的系统。在识别系统中通过自动阈值法和轮廓跟踪法分割肺实质;采用OTSU算法分割肺实质中的感兴趣区域，对感兴趣区域的特征进行提取;选择对肺结节和血管区别度较大的特征。根据选取的这些特征设定识别肺结节的规则来确定肺结节的候选区域。实验结果表明，该系统对直径1 cm以上的结节具有较好的识别性能。

关键词: 医学图像, 肺结节, 基于规则, 识别, OTSU算法