Research Of Lung CT Image Three-dimensional Reconstruction

Three-dimensional reconstruction technique is based on the image processing technologyof computer graphics in the medical field, which has a very wide range of applications. It canreconstruct the three-dimensional graphics of various organs, not only can be used to assistdoctors in clinical observation and diagnosis, but also can be used in the other fields, such astele-operation, teaching. It enhances the accuracy of the diagnosis, and offers learners withthree-dimensional, full range of visual sensory models, which has significant meaning inpractical applications.This paper based on the incidence rates of lung disease is increasing rise and thedifficulty of lung examination and treatment, obtaining three-dimensional display of lungs byusing computer software technology to construct human lung CT images. In the beginning,preprocessing lung CT images, which in the format of DICOM3.0format, including imageenhancement, image denoising and image interpolation. And then segment these images.Lastly, reconstructing three-dimension.This paper aims at the feature that edge of the lung is easily mixing with the other organs,selecting segmentation method of minimum spanning tree, which is similar to the nature ofclustering. Separating the area of the image, which has the same features as the minimumspanning tree with other areas. On this basis, improving the minimum spanning tree algorithm.Because the surrounding of chest in the lung CT images has many useless areas，usingmorphology to deal with. Isolating the chest of lung CT images and getting rid of edge, thusreducing the generation number of minimum spanning tree, shortening the operation time andimproving the efficiency of segmentation. Selecting the classical mobile cube reconstructionalgorithm in surface rendering in the reconstruction part. This algorithm based on theestablishment of the voxel model, constructing three-dimensional images of the lungs surfaceby topology constructing isosurface. This paper adds operation of grid smooth on the basis ofreconstructing to remove the noise and burr of the three-dimensional model, so image surfaceis smooth and improving the effects of reconstruction.Simulation experiment verifies the fast and effectiveness of the improving minimumspanning tree algorithm, and compares the effect of three-dimensional reconstruction based on improved minimum spanning tree with the simulation result of three-dimension based onthreshold segmentation. After that, it shows that the improved algorithm improves efficiencyand integrity of lung CT image three-dimensional reconstruction. People can get a better three-dimensional model of lungs. In the meantime of reconstructing three-dimension quickly, theeffect of three-dimensional reconstruction is better, and it will provide a powerful basis ofjudgment for diagnosis.