Study On Three-dimensional Reconstruction Of Lung Trachea Based On CT Image

With the development of medical CT imaging technology more and more mature,their use is also more and more common.High-precision digital images are come out one after the other,they provide a clearer diagnostic for imaging analysts.Because of high resolution and powerful function,computed tomography(C T)is the most common in current clinical diagnosis.But on the other hand,a large number of digital images increase the work burden on imaging analysts,and the nature of single reading work does not only take up a lot of medical resources,but also lead to the doctors' misdiagnosis rate increased.In this case,based on computer intelligence's CT image analysis and processing methods came into being.In the analysis of lung CT images,the segment of pulmonary trachea tree provides an important foundation for the diagnosis of pulmonary disease,and provides three-dimensional image navigation function for operation.Therefore,the trachea tree segment method in pulmonary CT images is an important field in computer intelligence CT image's research.First,this paper analysis the more classic a variety of methods of pulmonary tracheal extraction so far,as well as the difficulties in lung tracheal remodeling.After that,carry out the pretreatment of CT images,and carry out the analysis of the DICOM image format,the original CT data of the trachea were used to study the follow-up algorithm.Secondly,this paper realizes the extraction of the main trachea,and proposes a three-dimensional region grow algorithm based on the stack,the rough main trachea structure is obtained,and then carry out the morphological optimization to the result,and fill the small voids that exist,and finally the smooth main trachea structure is obtained.Once again,the extraction and splicing of the bronchus was achieved,and the extraction was performed on the basis of the extracted lung parenchyma,and the bronchioles were obtained by connecting domain labeling,dividing the lung parenchyma,adaptive binary and inner product.Finally,take the main trachea and bronchioles structure carrying out Or operation,and proposed a splicing tracheal algorithm,the results of the Or operation is still not connected through the two principles to judge,and finally use the splicing algorithm to link together,get a complete lung trachea tree.The correctness and superiority of the method of pulmonary trachea extraction were showed by comparing the experimental results with the results of manual marking.The main innovations of this paper are:1.This paper realizes the three-dimensional region grow algorithm based on the stack.The general two-dimensional region grow algorithm belongs to the recursive algorithm,it will consume a lot of time and memory.In order to solve this problem,the algorithm proposed in this paper is to create an array of stacks on the heap to simulate the advanced idea of the stack,and extends the neighborhood to twenty-six neighborhoods.2.In this paper,a splicing algorithm for bronchioles is proposed.First,according to the splicing criteria to match the main trachea and bronchioles,and then through the two-dimensional shape interpolation to stitch them together.It is found that there are still a few unconnected missing trachea regions after splicing,through the whole trachea is connected,using the region grow strategy to remove the non-tracheal area has not yet connected,and ultimately get a complete lung trachea tree.