Segmentation method of CT lung image based on multi-threshold and marker watershed algorithm

Since segmentation is confined by low conditions such as intensity non-homogeneity within different organizations, complex structures and other factors in lung CT image, it is difficult to segment lung boundary exactly. To improve the segmentation results, an integrated method of multi-threshold and marker controlled watershed segmentation is proposed. Firstly, multi-threshold method is used to solve image coarse segmentation and eliminate the residual trachea and main bronchus of image, and then marker controlled watershed segmentation method is used to fine segment and morphological operations is used to repair lung edge, finally clinical CT cardiac images are used to test the performance on matlab 2012 platform. The experimental results show that, this proposed method can better preserve the image edge information, effectively improve the accuracy of lung CT image segmentation, has obtained the ideal segmentation effect, it can provides valuable reference information for the clinical diagnosis of pulmonary diseases.