超声图像辅助手术器械跟踪技术研究

发布时间：2018-04-12 06:12

本文选题：前列腺根切术 + 超声图像　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：随着前列腺肿瘤的发病率逐年提升,发病群体趋于年轻化,对前列腺的精确治疗逐渐受到人们的关注。在微创手术机器人辅助前列腺根切术中,医生通过观看病灶部位的图像来开展相应的手术操作。由于前列腺周围组织内部的血管神经束(Neurovascular Bundle,NVB)不能在内窥镜图像中呈现,在切割过程中容易对NVB造成损伤,影响术后患者排尿及性功能。因此,本文通过微器械与组织接触变形,进行超声探头操作机构与手术机器人的配准,实现对微器械的实时跟踪,获取器械操作部位组织的超声图像,根据超声图像提供的信息减少对NVB的损伤。本文根据前列腺微创手术过程中的患者体位、直肠超声探头的尺寸及成像要求,对直肠超声探头操作机构进行结构和控制单元的设计。将超声探头操作机构、超声设备与主控台集成,构建术中超声图像辅助系统,制定了超声图像辅助系统在前列腺根切手术中的操作流程。由于人工提取微器械特征点的过程受医生主观经验影响,提取精度随机性较大,本文采用自动提取方法来获取超声图像中微器械的特征点。在该方法中,利用HAMMER算法进行有器械和无器械图像的弹性配准,对配准后的图像进行减影,获得器械的大致轮廓和位置。提出基于J散度的各向异性扩散去噪算法,对超声图像进行去噪处理。结合微器械末端先验轮廓知识,提出基于先验知识的参数活动轮廓算法,以减影后的轮廓为初始轮廓,对微器械轮廓进行分割。通过计算分割后的微器械轮廓内部区域像素坐标均值,提取出微器械特征点。本文采用经典的N型模板标定结构,完成对超声探头的内外参标定。利用标定得到的超声图像与超声探头操作机构末端坐标系的位姿转换矩阵,多次获得微器械末端特征点在超声探头操作机构与微创手术机器人基坐标系下的坐标,采用Levenberg-Marquardt算法完成超声探头操作机构与微创手术机器人的配准。在获得探头内外参标定矩阵和配准结果的基础上,结合微创手术机器人与超声探头操作机构运动学,实现超声图像对微器械的跟踪。为了验证超声图像辅助手术器械跟踪技术的可行性,对术中超声图像辅助系统及其关键技术开展实验研究。结合实验室现有的微创手术机器人,搭建超声图辅助手术器械跟踪实验平台。开展超声图像中微器械特征点提取实验,验证超声图像中微器械特征点自动提取方法的精确性和稳定性。开展超声探头内外参数的标定以及手术机器人与探头操作机构的配准实验,得到标定矩阵和配准结果。开展超声图像对微器械的跟踪实验,验证超声图像辅助手术器械跟踪技术的有效性。
[Abstract]:With the incidence of prostate cancer increasing year by year, the incidence of prostate cancer population tends to be younger, and the accurate treatment of prostate has been paid more and more attention.In minimally invasive robotic assisted prostatectomy, the doctor performs the operation by looking at the image of the lesion.Because the neurovascular bundles of nerves and nerves in periprostatic tissue can not be seen in endoscope images, it is easy to damage NVB in the course of cutting, which affects the urination and sexual function of the patients after operation.Therefore, through the contact deformation between the microinstruments and tissues, the ultrasonic probe operating mechanism is registered with the surgical robot, and the real-time tracking of the microinstruments is realized, and the ultrasonic images of the operating parts of the instruments are obtained.According to the information provided by ultrasound images, the damage to NVB is reduced.According to the position of the patient during the minimally invasive prostate surgery, the size and imaging requirements of the rectal ultrasound probe, the structure and control unit of the rectal ultrasonic probe operating mechanism were designed.The ultrasonic probe operating mechanism, ultrasonic equipment and the main control desk were integrated to construct an intraoperative ultrasound image assistant system, and the operation flow of the ultrasonic image assistant system in prostatectomy was established.Because the process of manually extracting the feature points of microinstruments is influenced by the doctors' subjective experience and the accuracy of extraction is randomness, this paper adopts the automatic extraction method to obtain the feature points of microdevices in ultrasonic images.In this method, the elastic registration of the images with and without instruments is carried out by using HAMMER algorithm, and the subtraction of the images after registration is carried out to obtain the approximate contour and position of the instruments.An anisotropic diffusion denoising algorithm based on J divergence is proposed for ultrasonic image denoising.Combining with the prior contour knowledge of the end of microinstruments, a parameter active contour algorithm based on the prior knowledge is proposed. The contour of the microdevice is segmented with the contour of the subtraction as the initial contour.By calculating the average pixel coordinates of the inner region of the segmented micro device contour, the feature points of the micro device are extracted.In this paper, the internal and external parameters of ultrasonic probe are calibrated by using the classical N-template calibration structure.By using the position and pose transformation matrix of the calibrated ultrasonic image and the end-coordinate system of the ultrasonic probe operating mechanism, the coordinates of the terminal characteristic points of the micro-instrument in the ultrasonic probe operating mechanism and the minimally invasive surgical robot coordinate system are obtained many times.The Levenberg-Marquardt algorithm is used to complete the registration of ultrasonic probe operating mechanism and minimally invasive surgical robot.On the basis of the calibration matrix and registration results of the internal and external parameters of the probe, combined with the kinematics of the micro-invasive surgical robot and the ultrasonic probe, the tracking of microinstruments by ultrasonic images is realized.In order to verify the feasibility of ultrasonic image assisted surgical instrument tracking technique, an experimental study was carried out on the intraoperative ultrasound image aid system and its key techniques.Combined with the existing minimally invasive surgical robot in the laboratory, an experimental platform for ultrasonic assisted surgical instrument tracking was built.In order to verify the accuracy and stability of the automatic extraction method of microinstrument feature points in ultrasonic images, an experiment was carried out to extract the feature points of microinstruments in ultrasonic images.The calibration of internal and external parameters of ultrasonic probe and the registration experiment of surgical robot and probe operating mechanism are carried out, and the calibration matrix and registration results are obtained.The ultrasonic image tracking experiment was carried out to verify the effectiveness of the ultrasonic image assisted surgical instrument tracking technique.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R737.25;R445.1;TP391.41

【参考文献】