X光图像拼接方法的研究

发布时间：2018-01-15 14:34

本文关键词：X光图像拼接方法的研究　出处：《东南大学》2017年硕士论文　论文类型：学位论文

【摘要】：在人体脊椎等长骨手术中,需要获取完整的解剖结构图像。DR(Digital Radiography)一次只能采集有限视野的图像,为了获得大视野范围的图像,需要对X光图像进行拼接。现有的拼接方法或者操作复杂,或者不能对存在透视变形的X光图像很好的拼接。本文通过对X光图像透视变形的原因进行分析,提出基于深度的X光图像拼接方法。首先,从图像中提取出能代表关键解剖结构的特征窗口图像,并对特征窗口图像采用多分辨率方法进行匹配。然后,根据匹配的结果计算特征窗口图像的视差,利用视差法得到特征窗口图像所代表的关键解剖结构的深度信息。最后,根据提取出的关键深度信息,选择感兴趣的深度范围,针对该深度范围内的图像实现拼接。实验结果表明,该深度信息提取方法可以从X光图像中提取出人体的关键深度信息;基于深度的拼接方法在不增加硬件成本的情况下,通过简便的操作,能得到拼接效果良好的结果图像。
[Abstract]:In the operation of human spine isobarous bone, we need to obtain complete anatomical structure image. DRN Digital Radiography) can only collect images with limited visual field at a time. In order to obtain images with large field of vision, X-ray images need to be stitched. The existing stitching methods or operations are complex. Or the X-ray images with perspective deformation can not be well stitched. Through the analysis of the causes of X-ray image perspective deformation, the depth based X-ray image stitching method is proposed. First of all. The feature window image which can represent the key anatomical structure is extracted from the image, and the feature window image is matched by multi-resolution method. Then, the parallax of the feature window image is calculated according to the matching results. The parallax method is used to obtain the depth information of the key anatomical structure represented by the feature window image. Finally, according to the extracted key depth information, the depth range of interest is selected. The experimental results show that the depth information extraction method can extract the key depth information of human body from X-ray images. The depth-based stitching method can get a good result image by simple operation without increasing the hardware cost.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【参考文献】