医学超声图像分割的分析与实现-计算机应用技术专业论文.docxVIP

摘要医学超声成像由于其独具的实时性、无损性、廉价性、可重复性好、灵敏度高等优 势广泛应用于临床诊断和治疗中，常需要提取特定组织器官的信息，其不可缺少的手段 就是图像分割，通过图像分割，可以从医学超声图像中分离出感兴趣的解剖结构或定位 出病源的位置和形状，它常直接决定着后续分析的准确性。但是由于医学超声图像固有 的斑点噪声和组织相关的纹理，图像进行分割还达不到临床应用上的要求，在研究方面 也没有提出一个速度快准确率高的方法 。本文从医学超声图像的去噪入手，通过去噪，增强等预处理后，提出和改进了医学 超声图像分割算法，通过得到较为精确的分割效果，为后续分析提供可靠的保障。这也 是对低信噪比图像分割方法的丰富和有效探索，具有一定的理论意义和应用价值。论文 首先介绍了医学超声图像分割领域的研究现状、分析了目前医学超声图像分割中存在的 不足之处，而后在预处理阶段小波变换对专门针对原始医学超声图像中存在的斑纹噪声 选择合理的阈值进行去噪，在增强阶段利用直方图均衡化的理论，达到图像对比度非线 性增强，在分割阶段算法中结合梯度值确定生长条件完成区域生长，利用数学形态学对 生长后的图像进行调整方法，得出了较好的分割结果，在后处理阶段利用三次 B 样条 曲线模拟了分割边界，最后通过实验验证了算法的有效性。关键词：医学超声图像；图像分割；小波变换；直方图均衡；区域生长；数学形态学；AbstractThe medical image technology has been used more and more fields, because the medical ultrasound image has many characters of real time, lossless, cheapness, well repetition super sensitive. We often need to pick-up much communication of special apparatus, image segmentation apply to it appropriately. Across segmentation, we can receive apartment of anatomical proportion and get the position and figure of the focus. It often decides the veracity of next analysis. As the result of immanent speckle and veins of apparatus of medical ultrasound image, image segmentation fall short of clinical request and there is not effective and speediness method in the research field.In this thesis we find ways and means from wipe off speckle to region- growing segmentation, finally we received better results of segmentation. It is not only the guarantees of next analyze but also the effective explore of poor quality image segmentation method. It has definite theory meaning and applied value. At first, the thesis introduces the present condition of the research of medical ultrasound image segmentation, analyzes the feathers of the medical ultrasound image segmentation, then in the moment of pretreatment we make use of wavelet transform wipe off speckle and use histogram proportion, receive the non-linear enhancing the contract of medical ultrasound image, in the moment of segmentation we ma