WO2022160384A1 - 图像引导机器人微创外科手术的物理空间与图像空间配准方法 - Google Patents
图像引导机器人微创外科手术的物理空间与图像空间配准方法 Download PDFInfo
- Publication number
- WO2022160384A1 WO2022160384A1 PCT/CN2021/076194 CN2021076194W WO2022160384A1 WO 2022160384 A1 WO2022160384 A1 WO 2022160384A1 CN 2021076194 W CN2021076194 W CN 2021076194W WO 2022160384 A1 WO2022160384 A1 WO 2022160384A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- phantom
- coordinate system
- space
- physical space
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/20—Surgical navigation systems; Devices for tracking or guiding surgical instruments, e.g. for frameless stereotaxis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/30—Surgical robots
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
- A61B2034/107—Visualisation of planned trajectories or target regions
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
- A61B2034/108—Computer aided selection or customisation of medical implants or cutting guides
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/20—Surgical navigation systems; Devices for tracking or guiding surgical instruments, e.g. for frameless stereotaxis
- A61B2034/2046—Tracking techniques
- A61B2034/2065—Tracking using image or pattern recognition
Definitions
- the invention relates to the technical field of X-ray imaging, in particular to a physical space and image space registration method for image-guided robotic minimally invasive surgery.
- the registration of physical space and image space is a key link in image-guided robotic minimally invasive surgery.
- the registration process is a process of associating both the robot and the patient in the physical space and the CT image in the image space.
- Image-guided robotic minimally invasive surgery such as minimally invasive cochlear implant surgery
- the distance between the drilled cochlear implant channel and the important anatomical structures such as the facial nerve and cord tympani is about 0.5mm. Structures are identified and acquired in the image space, so the registration accuracy between the physical space and the image space largely determines the success of robotic minimally invasive surgery.
- the registration of physical space and image space is usually realized by means of point registration or surface registration.
- the most accurate point registration method is to implant fiducial markers in the bone by surgery.
- the existing literature (Gerber, N., et al. (2013). "High-Accuracy Patient-to-Image Registration for the Facilitation of Image-Guided Robotic Microsurgery on the Head", IEEE Transactions on Biomedical Engineering, 60(4):960-968.) discloses a temporal bone incision screwed into ten anchor points as implanted fiducial markers.
- the physical space coordinates of the above-mentioned fiducial markers are obtained by optical or magnetic position sensors, such as high-precision optical trackers and dedicated 6-DOF force sensing devices, but the 3D measurement process is complicated, the accuracy is limited, and the cost is high.
- the surface registration method uses laser surface scanning or tracking pointer to detect surface point cloud coordinates and then performs registration through the iterative nearest point algorithm, and its registration accuracy is not as good as point registration, such as existing literature (Fan, Y., et al. (2020).
- the existing physical space and image space registration methods have shortcomings such as low accuracy, complex process, high cost, and damage to the body, which seriously limit the application of image-guided robotic minimally invasive surgery in complex and delicate surgery. development and application.
- the purpose of the present invention is to provide an image-guided robot minimally invasive surgery physical space and image space registration method, which effectively avoids the complex process of physical space and image space paired mark point coordinate acquisition and registration in conventional methods.
- the method for registering the physical space and the image space of an image-guided robotic minimally invasive surgery includes the following steps:
- step 2) For each projection image obtained in step 2), according to the phantom-based CBCT imaging geometric parameter calibration algorithm, identify the imaging geometric parameters at each angle in the intraoperative CBCT scanning process, and obtain the phantom coordinate system and the phantom coordinate system. Conversion relationship between detector coordinate systems;
- step 3 According to the intraoperative CBCT imaging geometric parameters obtained in step 3), modify the CBCT three-dimensional reconstruction algorithm, and establish a three-dimensional reconstruction algorithm model under the phantom coordinate system;
- Intraoperative CBCT scans the surgical object to obtain human body projections at different angles, and the angles of the collected human projections correspond one-to-one with the angles of the phantom projections in step 2);
- step 6) Substitute the human body projection diagram obtained in step 5) into the three-dimensional reconstruction algorithm model established in step 4), and reconstruct the three-dimensional CT image of the surgical object in the model body coordinate system;
- the phantom is scanned and calibrated by intraoperative CBCT, and the imaging geometric parameters are automatically identified, and a three-dimensional reconstruction algorithm model under the phantom coordinate system is established.
- the CBCT scans the surgical object to reconstruct the 3D CT image in the phantom coordinate system, and the calibration phantom is fixedly connected to the robot through the base, and the relatively fixed transformation relationship between the phantom coordinate system and the robot coordinate system can be calibrated once. .
- the calibration mold body includes a cylindrical barrel and a sphere arranged in the barrel according to a certain rule, and the bottom of the barrel is provided with a base for connecting to the robot. the fixed seat.
- Said spheres are arranged in the following way:
- Two ends of the central axis of the cylinder body are respectively provided with a sphere, and a sphere is set at the center of the cylinder body;
- a circle of spheres is evenly arranged around the top end cap and the bottom end cap of the cylinder.
- step 3) includes:
- the segmentation methods in step 3-1) include a segmentation method based on gray threshold, a segmentation method based on edge detection and Hough circle detection, and a segmentation method based on U-Net fully convolutional network.
- step 3-2 the center coordinate of each spherical projection is calculated by the method of pixel gray value weighting.
- step 4 adopt the form of homogeneous transformation matrix to revise the described CBCT three-dimensional reconstruction algorithm, step 4) comprises:
- step 4-2) Using the imaging geometric parameters obtained in step 3-3) and the transformation matrix between the phantom coordinate system and the detector coordinate system, transform the focus coordinates and reconstructed voxel coordinates in the phantom coordinate system to the detection in the coordinate system of the device;
- the voxel-driven back-projection process modifies the weighted back-projection process in the form of intersection between the focus and the voxel point and the detector plane, and adopts a fast incremental algorithm
- the 3D reconstruction algorithm model is obtained by quickly traversing and solving each voxel.
- the imaging geometric parameters described in the present invention include the focus coordinates of the calibration phantom, the coordinates of the center point of the detector, and the deflection angle of the detector.
- the invention utilizes the calibration phantom to unify the physical space and the image space, identifies the geometric parameters of intraoperative CBCT imaging through the calibration phantom, establishes a three-dimensional reconstruction algorithm model under the phantom coordinate system based on the calibration phantom, and directly reconstructs the three-dimensional CT image of the surgical object.
- Based on the phantom coordinate system it avoids the complex process of obtaining and registering the coordinates of the physical space and image space paired markers in the conventional method, greatly improving the registration accuracy of the physical space and the image space, simplifying the registration process, and shortening the registration time. Accurate time, no damage to the surgical object, can effectively promote the development and application of image-guided robotic minimally invasive surgery in complex and delicate surgery.
- FIG. 1 is a flowchart of a method for registering a physical space and an image space in an image-guided robotic minimally invasive surgical operation according to an embodiment of the present invention
- FIG. 2 is a schematic structural diagram of a calibration phantom in an embodiment of the present invention.
- FIG. 3 is a flowchart of obtaining imaging geometric parameters in an embodiment of the present invention.
- FIG. 4 is a flowchart of obtaining a CT image of a surgical object in an embodiment of the present invention
- FIG. 5 is a schematic diagram of an intraoperative CBCT scanning calibration phantom in an embodiment of the present invention.
- FIG. 6 is a schematic diagram of an intraoperative CBCT scanning operation object in an embodiment of the present invention.
- the calibration mold structure 100 used in this embodiment includes a cylindrical barrel 4 and a sphere arranged in the barrel according to a certain rule.
- the top of the barrel 4 is provided with an upper end cover 1
- the bottom is provided with a lower end cover 9 and a sphere.
- the sphere includes: the upper center ball 3 and the lower center ball 7 located at both ends of the middle column 6 of the cylinder, the middle center ball 5 located at the center of the middle column 6 of the cylinder; the upper layer evenly distributed around the upper end cover 1
- the small balls 2 are uniformly distributed in the lower layer around the lower end cover 9 and the small balls 8 are evenly distributed.
- the present invention is not limited to using the above calibration phantom, and the above calibration phantom is only one embodiment of the present invention.
- the method for registering a physical space and an image space for image-guided robotic minimally invasive surgery in this embodiment includes the following steps:
- the calibration phantom 100 can determine all imaging geometric parameters according to a projection icon, including focus coordinates, detector center point coordinates, and detector deflection angle.
- Intraoperative CBCT400 scans and calibrates the phantom 100 to obtain several projection images at different angles within a 360° range. Each projection angle value of the collected projection data can be acquired by the inclination sensor installed on the intraoperative CBCT400.
- step S3 As shown in Figure 2, for each projection image obtained in step S2, according to the phantom-based CBCT imaging geometric parameter calibration algorithm, the imaging geometric parameters at each angle in the intraoperative CBCT400 scanning process are automatically identified. Referring to Figure 3, this step specifically includes the following sub-steps:
- step S4 As shown in Figure 4, according to the actual imaging geometric parameters of the intraoperative CBCT400 obtained in step S3, the CBCT 3D reconstruction algorithm is modified in the form of a homogeneous transformation matrix, and a 3D reconstruction algorithm model in the phantom coordinate system is established.
- This step specifically includes the following sub-steps:
- step S41 In the phantom coordinate system, use the imaging geometric parameters obtained in step S33 to perform cosine correction and line-by-line linear shift-invariant filtering operations on the projection data;
- step S42 Using the transformation matrix between the focus coordinates, the phantom coordinate system and the detector coordinate system obtained in step S33, transform the focus coordinates and reconstructed voxel coordinates in the phantom coordinate system into the detector coordinate system;
- the intraoperative CBCT 400 scans the surgical object 300 to obtain several projection maps at different angles within a 360° range, and the angles of the collected projection maps correspond one-to-one with the projection map angles in step S2.
- the position of the intraoperative CBCT 400 when scanning the surgical object 300 is consistent with the position when scanning the calibration phantom 100 in step S2.
- each projection angle value of the acquired projection data can be acquired by the inclination sensor installed on the intraoperative CBCT400.
- step S6 Substitute the projection image obtained in step S5 into the 3D reconstruction algorithm model in the phantom coordinate system established in step S4, and reconstruct the 3D CT image of the surgical object 300 in the phantom coordinate system.
- the three-dimensional CT image of the surgical object 300 is directly reconstructed in the phantom coordinate system, realizing the unification of the physical space and the image space, and the anatomical structure obtained in the image space is directly positioned in the physical space determined by the phantom coordinate system.
- This embodiment also presents the implementation result of the target test ball in the case of scanning the calibration phantom with the physical space and image space registration method.
- the target test ball is the upper center ball 3 and the lower center ball 7 in the calibration phantom shown in Figure 2.
- the origin O of the phantom coordinate system is set at the center of the middle central sphere 5, and the calibration phantom axis is set as the Z axis.
- the midplane is set as the XOY plane
- the coordinates of the center of the upper center ball 3 of the calibration model in the physical space are (0,0,39.5) (that is, the gold standard)
- the coordinates of the center of the lower center ball 7 are (0,0,- 39.5) (ie the gold standard).
- the line attenuation of the acrylic material used in the upper end cap 1, the lower end cap 9, the cylinder 4, the middle column 6, the fixing seat 10, etc. is about 0.021, and the line attenuation of the ball It is about 0.56, so by setting the voxel threshold (0.1), the area where the small sphere voxel is located can be segmented from the reconstructed volume, and the image space coordinates of the center of the small sphere can be obtained by weighting the voxel value.
- the coordinates of the center ball 3 of the upper layer of the calibration phantom in the image space are (-0.0253, -0.0474, 39.5163), and the coordinates of the center of the lower layer of the center ball 7 are (-0.0254, -0.0299, -39.4921).
- the target registration error (Target Registration Error, TRE) between the physical space and the image space
- the registration accuracy obtained by applying the registration method based on the calibration phantom coordinate system in this embodiment in the registration process between the physical space and the image space of the target test ball is obviously better than the registration accuracy obtained by the conventional method (usually a few millimeters). , individually to a few tenths of a millimeter).
- the registration method in this embodiment has a simple registration process, does not require surgical implantation of fiducial markers, and does not require the aid of a high-precision three-dimensional measurement device, which effectively shortens the registration time and ensures the accuracy of the registration results. Repeatability and reliability.
- the conventional point registration method needs to increase the operation time of the pre-surgery for the surgical implantation of fiducial markers.
- the surface registration method is a manual operation for the three-dimensional measurement of the physical space surface point cloud, and the general registration time is 20 to 30 minutes.
Landscapes
- Health & Medical Sciences (AREA)
- Surgery (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Robotics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
一种图像引导机器人微创外科手术的物理空间与图像空间配准方法，属于X射线成像技术领域。通过术中CBCT（400）扫描标定模体（100），自动识别成像几何参数，建立模体坐标系下的三维重建算法模型，术中CBCT（400）再扫描手术对象（300）即可重建出模体坐标系中的三维CT图像，而标定模体（100）通过基座（201）与机器人（200）固定连接，模体坐标系与机器人坐标系之间相对固定的转换关系标定一次即可。利用标定模体（100）将物理空间和图像空间统一，避免了常规方法中物理空间和图像空间配对标记点坐标获取和配准的复杂过程，大幅提升了物理空间和图像空间的配准精度、简化了配准流程、缩短了配准时间，对手术对象（300）亦无损伤。
Description
本发明涉及X射线成像技术领域，具体地说，涉及一种图像引导机器人微创外科手术的物理空间与图像空间配准方法。
物理空间与图像空间配准是图像引导机器人微创外科手术的关键环节，配准过程是将物理空间包含机器人和患者与图像空间包含CT图像两者相关联的过程。图像引导机器人微创外科手术，例如微创人工耳蜗植入手术，钻制的人工耳蜗通道与面神经、鼓索等重要解剖结构的距离约0.5mm，手术路径在物理空间中规划与实施，而解剖结构在图像空间中识别与获取，因此物理空间与图像空间的配准精度在很大程度上决定了机器人微创外科手术成功与否。
目前物理空间与图像空间配准通常采用点配准或面配准的方法实现。点配准方法中精度较高的是通过手术的方式在骨中植入基准标记点，现有文献(Gerber，N.，et al.(2013).“High-Accuracy Patient-to-Image Registration for the Facilitation of Image-Guided Robotic Microsurgery on the Head”，IEEE Transactions on Biomedical Engineering，60(4):960-968.)公开了一种颞骨切口旋入十个锚点作为植入的基准标记点，目标配准误差(Target Registration Error)，TRE＝0.2mm，但是对患者有侵入性损伤；而贴皮基准标记点和解剖基准标记点，获得的配准精度较差。此外，上述基准标记点的物理空间坐标通过光学或磁性位 置传感器获取，如高精度光学跟踪仪和专用六自由度力传感设备，但三维测量过程复杂、精度受限且成本较高。面配准方法采用激光表面扫描或跟踪指针检测表面点云坐标再通过迭代最近点算法进行配准，其配准精度比不上点配准，如现有文献(Fan，Y.，et al.(2020).“A robust automated surface-matching registration method for neuronavigation”，Medical Physics,47(7):2755-2767.)公开的采用三维白光扫描仪获取头部表面点云，经过粗精两次配准后目标配准误差TRE＝1.17mm。
综上所述，现有物理空间与图像空间配准方法存在精度偏低、过程复杂、成本较高、损伤机体等不足之处，严重限制了图像引导机器人微创外科手术在复杂精细手术方面的发展与应用。
发明内容
本发明的目的是提供一种图像引导机器人微创外科手术物理空间与图像空间配准方法，该方法有效避免了常规方法中物理空间和图像空间配对标记点坐标获取和配准的复杂过程。
为了实现上述目的，本发明提供的图像引导机器人微创外科手术物理空间与图像空间配准方法，包括以下步骤：
1)将标定模体固连在机器人上，确定模体坐标系与机器人坐标系之间的转换关系；
2)术中CBCT扫描标定模体，获得不同角度的模体投影图；
3)针对步骤2)中获得的每张投影图，依据基于模体的CBCT成像几何参数标定算法，识别术中CBCT扫描过程中每个角度下的成像几何参数，并求得模体坐标系与探测器坐标系之间的转换关系；
4)根据步骤3)中获得的术中CBCT成像几何参数，修正CBCT三维重建算法，建立模体坐标系下的三维重建算法模型；
5)术中CBCT扫描手术对象，获得不同角度的人体投影图，采集人体投影图的角度与步骤2)中模体投影图的角度一一对应；
6)将步骤5)中获得的人体投影图代入步骤4)中建立的三维重建算法模型，重建模体坐标系中手术对象的三维CT图像；
7)利用模体坐标系与机器人坐标系之间的转换关系，获得机器人与手术对象内部结构的相对位置关系。
上述技术方案中，不需要设定基准标记点，也不需要借助三维测量装置，通过术中CBCT扫描标定模体，自动识别成像几何参数，建立模体坐标系下的三维重建算法模型，术中CBCT再扫描手术对象即可重建出模体坐标系中的三维CT图像，而标定模体通过基座与机器人固定连接，模体坐标系与机器人坐标系之间相对固定的转换关系标定一次即可。
可选地，在一个实施例中，所述的标定模体包括圆柱形的筒体和按照一定规律设置在所述筒体内的球体，所述筒体的底部设有用于连接所述机器人基座的固定座。
所述的球体按照以下方式布置：
所述筒体中心轴的两端各设置一个球体，筒体的中心位置处设置一个球体；
筒体的顶部端盖和底部端盖的周围均匀布置一圈球体。
进一步地，步骤3)包括：
3-1)对于每张投影图，分割出标定模体中各个球体的投影；
3-2)计算每个球体投影的中心坐标；
3-3)根据球体投影中心坐标与球体投影之间的几何关系，求解成像几何参数、以及模体坐标系与探测器坐标系之间的变换矩阵。
步骤3-1)中的分割方法包括基于灰度阈值的分割方法、基于边缘检测和霍夫圆检测的分割方法、基于U-Net全卷积网络的分割方法。
步骤3-2)中，采用像素灰度值加权的方法计算每个球体投影的中心坐标。
进一步地，步骤4)中，采用齐次变换矩阵的形式修正所述的CBCT三维重建算法，步骤4)包括：
4-1)在模体坐标系中，利用步骤3-3)中获得的成像几何参数对投影数据进行余弦修正和逐行线性移不变滤波操作；
4-2)利用步骤3-3)中获得的成像几何参数、以及模体坐标系与探测器坐标系之间的变换矩阵，将模体坐标系中的焦点坐标和重建体素坐标变换至探测器坐标系中；
4-3)在探测器坐标系中，基于体素驱动的反投过程按照焦点和体素点连线与探测器平面求交的形式，对加权反投过程进行修正，并采用快速增量算法对各体素进行快速遍历及求解，得到三维重建算法模型。
本发明中所述的成像几何参数包括标定模体的焦点坐标、探测器中心点坐标、以及探测器偏转角度。
与现有技术相比，本发明的有益之处在于：
本发明利用标定模体将物理空间与图像空间统一，通过标定模体识别出术中CBCT成像几何参数，并据此建立模体坐标系下的三维重建算法模型，将手术对象三维CT图像直接重建于模体坐标系，避免了常规方法中物理空间和图像空间配对标记点坐标获取和配准的复杂过程，大幅提升了物理空间和图像空间的配准精度、简化了配准流程、缩短了配准时间，对手术对象亦无损伤，可有效促进图像引导机器人微创外科手术在复杂精细手术方面的发展与应用。
图1为本发明实施例中图像引导机器人微创外科手术物理空间与图像空间配准方法的流程图；
图2为本发明实施例中标定模体的结构示意图；
图3为本发明实施例中获得成像几何参数的流程图；
图4为本发明实施例中获得手术对象CT图像的流程图；
图5为本发明实施例中术中CBCT扫描标定模体的示意图；
图6为本发明实施例中术中CBCT扫描手术对象的示意图。
为使本发明的目的、技术方案和优点更加清楚，以下结合实施例及其附图对本发明作进一步说明。显然，所描述的实施例是本发明的一部分实施例，而不是全部的实施例。基于所描述的实施例，本领域普通技术人员在无需创造性劳动的前提下所获得的所有其他实施例，都属于本发明保护的范围。
除非另外定义，本发明使用的技术术语或者科学术语应当为本发明所属领域内具有一般技能的人士所理解的通常意义。本发明中 使用的“包括”或者“包含”等类似的词语意指出现该词前面的元件或者物件涵盖出现在该词后面列举的元件或者物件及其等同，而不排除其他元件或者物件。
实施例
本实施例所使用的标定模体结构100参见图2，包括圆柱形的筒体4和按照一定规律设置在筒体内的球体，筒体4顶部设有上端盖1，底部设有下端盖9和用于连接机器人200的固定座10。球体包括：位于筒体中间立柱6的两端的上层中心小球3和下层中心小球7，位于筒体中间立柱6中心处的中间中心小球5；均匀布置在上端盖1周围的上层均布小球2，均匀布置在下端盖9周围的下层均布小球8。
当然，本发明并不局限于使用以上标定模体，以上标定模体仅是本发明的一个实施方式。
参见图1，本实施例中图像引导机器人微创外科手术的物理空间与图像空间配准方法包括以下步骤：
S1：参见图5，将标定模体100固定在与机器人200连接的基座201上，标定模体100所处的位置需在术中CBCT400扫描成像视野范围内。标定模体100可以根据一张投影图标定出所有成像几何参数，包括焦点坐标、探测器中心点坐标、探测器偏转角。
S2：术中CBCT400扫描标定模体100，获得360°范围内不同角度的若干张投影图。采集投影数据的每个投影角度值可由安装在术中CBCT400上的倾角传感器获取。
S3：如图2所示，针对步骤S2中获得的每张投影图，依据基于模体的CBCT成像几何参数标定算法，自动识别术中CBCT400扫 描过程中每个角度下的成像几何参数。参见图3，该步骤具体包括以下子步骤：
S31：对于每张投影图，采用基于灰度阈值的分割方法或基于边缘检测和霍夫圆检测的分割方法或基于U-Net全卷积网络的分割方法，分割出标定模体中各个小球的投影；
S32：对于每个小球投影采用像素灰度值加权的方法计算小球投影中心坐标；
S33：根据小球投影中心坐标以及小球投影之间的几何关系，求解焦点坐标、探测器中心点坐标、探测器偏转角等成像几何参数，以及模体坐标系与探测器坐标系之间的变换矩阵。
S4：如图4所示，根据步骤S3中获得的术中CBCT400实际成像几何参数，采用齐次变换矩阵的形式，修正CBCT三维重建算法，建立模体坐标系下的三维重建算法模型。该步骤具体包括以下子步骤：
S41：在模体坐标系中，利用步骤S33中获得的成像几何参数对投影数据进行余弦修正和逐行线性移不变滤波操作；
S42：利用步骤S33中获得的焦点坐标、模体坐标系与探测器坐标系之间的变换矩阵，将模体坐标系中焦点坐标、重建体素坐标，变换至探测器坐标系中；
S43：在探测器坐标系中，基于体素驱动的反投过程按照焦点和体素点连线与探测器平面求交的形式实现，对加权反投过程进行修正，并采用快速增量算法对各体素进行快速遍历及求解。
S5：参见图6，术中CBCT400扫描手术对象300，获得360°范围内不同角度的若干张投影图，采集投影图的角度与步骤S2中的 投影图角度一一对应。术中CBCT400扫描手术对象300时的位置与步骤S2中扫描标定模体100时的位置保持一致。术中CBCT400扫描手术对象300时，采集投影数据的每个投影角度值可由安装在术中CBCT400上的倾角传感器获取。
S6：将步骤S5中获得的投影图代入步骤S4中建立的模体坐标系下的三维重建算法模型，重建出模体坐标系中手术对象300的三维CT图像。直接在模体坐标系中重建出手术对象300的三维CT图像，实现了物理空间与图像空间的统一，图像空间中获取的解剖结构直接定位在模体坐标系所确定的物理空间中。
S7：利用模体坐标系与机器人坐标系之间确定的转换关系，即可获得机器人200与手术对象300及手术对象300解剖结构的相对位置关系。物理空间中模体坐标系与机器人坐标系之间相对固定的转换关系，在手术机器人搭建时标定一次即可。
本实施例还给出该物理空间与图像空间配准方法在扫描标定模体情况下目标测试小球的实施结果。目标测试小球为图2所示标定模体中的上层中心小球3和下层中心小球7，模体坐标系原点O设在中间中心小球5球心，标定模体轴线设为Z轴，中平面设为XOY平面，物理空间中标定模体上层中心小球3球心坐标为(0,0,39.5)(即金标准)，下层中心小球7球心坐标为(0,0,-39.5)(即金标准)。模体坐标系下标定模体的三维CT重建图像中，上端盖1、下端盖9、筒体4、中间立柱6、固定座10等所用亚克力材质的线衰约为0.021，小球的线衰约为0.56，因此可以通过设置体素阈值(为0.1)，从重建体中分割出小球体素所在的区域，并通过体素值加权的方式得到小球球心的图像空间坐标。由此得到图像空间中标定模体上层中心小球3球心坐标为(-0.0253，-0.0474, 39.5163)，下层中心小球7球心坐标为(-0.0254，-0.0299，-39.4921)。本实施例中物理空间与图像空间的目标配准误差(Target Registration Error，TRE)，上层中心小球3：TRE
marker1＝0.056mm，下层中心小球7：TRE
marker2＝0.040mm。术中CBCT扫描标定模体耗时2分钟，几何参数识别和CT重建算法修正计算机程序自动运行耗时0.5分钟，术中CBCT扫描手术对象耗时2分钟，手术对象三维CT图像重建耗时0.5分钟，上述配准过程总体耗时5分钟。
上述目标测试小球物理空间与图像空间配准过程中应用本实施例中基于标定模体坐标系的配准方法得到的配准精度明显优于常规方法所得到的配准精度(一般为几毫米，个别为零点几毫米)。特别是本实施例中的配准方法其配准流程简单，不需要通过手术植入基准标记点，也不需要借助高精度三维测量装置，有效缩短配准时间的同时也确保了配准结果的可重复性和可靠性。常规点配准方法手术植入基准标记点需增加预手术环节操作时间较长，面配准方法物理空间表面点云三维测量为手动操作一般配准时间为20～30分钟。
显然，上述实施例仅仅是为清楚地说明所作的举例，而并非对实施方式的限定。对于所属领域的普通技术人员来说，在上述说明的基础上还可以做出其它不同形式的变化或变动。这里无需也无法对所有的实施方式予以穷举。而由此所引伸出的显而易见的变化或变动仍处于本发明创造的保护范围之中。
Claims (9)
- 一种图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，包括以下步骤：1)将标定模体固连在机器人上，确定模体坐标系与机器人坐标系之间的转换关系；2)术中CBCT扫描标定模体，获得不同角度的模体投影图；3)针对步骤2)中获得的每张投影图，依据基于模体的CBCT成像几何参数标定算法，识别术中CBCT扫描过程中每个角度下的成像几何参数，并求得模体坐标系与探测器坐标系之间的转换关系；4)根据步骤3)中获得的术中CBCT成像几何参数，修正CBCT三维重建算法，建立模体坐标系下的三维重建算法模型；5)术中CBCT扫描手术对象，获得不同角度的人体投影图，采集人体投影图的角度与步骤2)中模体投影图的角度一一对应；6)将步骤5)中获得的人体投影图代入步骤4)中建立的三维重建算法模型，重建模体坐标系中手术对象的三维CT图像；7)利用模体坐标系与机器人坐标系之间的转换关系，获得机器人与手术对象内部结构的相对位置关系。
- 根据权利要求1所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，所述的标定模体包括圆柱形的筒体和按照一定规律设置在所述筒体内的球体，所述筒体的底部设有用于连接所述机器人基座的固定座。
- 根据权利要求2所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，所述的球体按照以下方式布置：所述筒体中心轴的两端各设置一个球体，筒体的中心位置处设置一个球体；筒体的顶部端盖和底部端盖的周围均匀布置一圈球体。
- 根据权利要求2所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，步骤3)包括：3-1)对于每张投影图，分割出标定模体中各个球体的投影；3-2)计算每个球体投影的中心坐标；3-3)根据球体投影中心坐标与球体投影之间的几何关系，求解成像几何参数、以及模体坐标系与探测器坐标系之间的变换矩阵。
- 根据权利要求4所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，步骤3-1)中的分割方法包括基于灰度阈值的分割方法、基于边缘检测和霍夫圆检测的分割方法、基于U-Net全卷积网络的分割方法。
- 根据权利要求4所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，步骤3-2)中，采用像素灰度值加权的方法计算每个球体投影的中心坐标。
- 根据权利要求4所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，步骤4)包括：4-1)在模体坐标系中，利用步骤3-3)中获得的成像几何参数对投影数据进行余弦修正和逐行线性移不变滤波操作；4-2)利用步骤3-3)中获得的成像几何参数、以及模体坐标系与探测器坐标系之间的变换矩阵，将模体坐标系中的焦点坐标和重建体素坐标变换至探测器坐标系中；4-3)在探测器坐标系中，基于体素驱动的反投过程按照焦点和体素点连线与探测器平面求交的形式，对加权反投过程进行修正，并采用快速增量算法对各体素进行快速遍历及求解，得到三维重建算法模型。
- 根据权利要求1所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，步骤4)中，采用齐次变换矩阵的形式修正所述的CBCT三维重建算法。
- 根据权利要求1～8中任一权利要求所述的图像引导机器人微创外科手术的物理空间与图像空间配准方法，其特征在于，所述的成像几何参数包括标定模体的焦点坐标、探测器中心点坐标、以及探测器偏转角度。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110104422.3 | 2021-01-26 | ||
| CN202110104422.3A CN112451093B (zh) | 2021-01-26 | 2021-01-26 | 图像引导机器人微创外科手术的物理空间与图像空间配准方法 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2022160384A1 true WO2022160384A1 (zh) | 2022-08-04 |
Family
ID=74802758
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2021/076194 Ceased WO2022160384A1 (zh) | 2021-01-26 | 2021-02-09 | 图像引导机器人微创外科手术的物理空间与图像空间配准方法 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112451093B (zh) |
| WO (1) | WO2022160384A1 (zh) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115631170A (zh) * | 2022-10-28 | 2023-01-20 | 中山大学 | 一种高精度线位移测量方法及其系统 |
| CN115737122A (zh) * | 2022-10-28 | 2023-03-07 | 浙江久择医疗科技有限公司 | 一种借助x射线的统一断层影像和磁定位工具坐标的方法 |
| CN115795579A (zh) * | 2022-12-23 | 2023-03-14 | 岭南师范学院 | 一种无特征复杂曲面误差分析的快速坐标对齐方法 |
| CN115880469A (zh) * | 2023-02-20 | 2023-03-31 | 江苏省人民医院（南京医科大学第一附属医院） | 一种表面点云数据与三维影像的配准方法 |
| CN116616869A (zh) * | 2023-03-03 | 2023-08-22 | 无锡艾米特智能医疗科技有限公司 | 一种超声点云的精准模型构建方法 |
| CN116672082A (zh) * | 2023-07-24 | 2023-09-01 | 苏州铸正机器人有限公司 | 一种手术导航标尺的导航配准方法及装置 |
| CN116993791A (zh) * | 2023-09-28 | 2023-11-03 | 真健康（北京）医疗科技有限公司 | 一种基于体表定位装置的医学图像配准方法及设备 |
| CN117474966A (zh) * | 2023-11-21 | 2024-01-30 | 上海卓昕医疗科技有限公司 | 医学影像与显影设备的配准方法、装置、设备和介质 |
| CN118914253A (zh) * | 2024-10-12 | 2024-11-08 | 浙江双元科技股份有限公司 | 一种非理想条件下的平面ct系统参数标定方法及装置 |
| CN119139021A (zh) * | 2023-06-14 | 2024-12-17 | 武汉联影智融医疗科技有限公司 | 手术坐标系的配准方法、装置和存储介质 |
| CN119184851A (zh) * | 2023-06-26 | 2024-12-27 | 武汉联影智融医疗科技有限公司 | 空间配准方法及装置、机器人辅助系统、设备和介质 |
| CN119399308A (zh) * | 2024-11-06 | 2025-02-07 | 中国科学院高能物理研究所 | 一种成像几何在线校准方法 |
| CN120531421A (zh) * | 2025-02-08 | 2025-08-26 | 北京和华瑞博医疗科技有限公司 | 定位系统 |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113143463B (zh) * | 2021-03-16 | 2022-08-26 | 上海交通大学 | 一种手术导航装置、系统、标定方法、介质及电子设备 |
| CN113643428A (zh) * | 2021-08-17 | 2021-11-12 | 北京唯迈医疗设备有限公司 | 一种适用于多自由度锥形束ct的全参数几何校准方法 |
| CN113963056B (zh) * | 2021-09-07 | 2022-08-26 | 于留青 | Ct图像重建方法、装置、电子设备以及存储介质 |
| CN113855288B (zh) * | 2021-11-01 | 2023-05-05 | 杭州柳叶刀机器人有限公司 | 图像生成方法、装置、电子设备及存储介质 |
| CN115005991B (zh) * | 2022-08-03 | 2022-12-13 | 北京壹点灵动科技有限公司 | 手术导航装置的精度检测方法及手术导航精度检测装置 |
| CN115998432B (zh) * | 2022-12-20 | 2026-04-28 | 上海导向医疗系统有限公司 | 一种手术导航中坐标自动配准方法及系统 |
| CN117338422B (zh) * | 2023-10-30 | 2024-04-05 | 赛诺威盛医疗科技(扬州)有限公司 | 空间注册及运动学解算机器人控制方法、系统及装置 |
| CN117679178B (zh) * | 2023-12-22 | 2024-06-07 | 青岛大学附属医院 | 一种创伤骨科微创手术机器人系统 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104083216A (zh) * | 2014-07-03 | 2014-10-08 | 北京天智航医疗科技股份有限公司 | 手术定位标尺 |
| US20140312229A1 (en) * | 2011-12-21 | 2014-10-23 | Rolls-Royce Plc | Position measurement |
| EP3295887A1 (en) * | 2016-09-16 | 2018-03-21 | Globus Medical, Inc. | Robotic fluoroscopic navigation |
| CN111388091A (zh) * | 2020-03-17 | 2020-07-10 | 京东方科技集团股份有限公司 | 光标尺、坐标系配准的方法 |
| CN111388089A (zh) * | 2020-03-19 | 2020-07-10 | 京东方科技集团股份有限公司 | 治疗设备及其配准方法、配准装置 |
-
2021
- 2021-01-26 CN CN202110104422.3A patent/CN112451093B/zh active Active
- 2021-02-09 WO PCT/CN2021/076194 patent/WO2022160384A1/zh not_active Ceased
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140312229A1 (en) * | 2011-12-21 | 2014-10-23 | Rolls-Royce Plc | Position measurement |
| CN104083216A (zh) * | 2014-07-03 | 2014-10-08 | 北京天智航医疗科技股份有限公司 | 手术定位标尺 |
| EP3295887A1 (en) * | 2016-09-16 | 2018-03-21 | Globus Medical, Inc. | Robotic fluoroscopic navigation |
| CN111388091A (zh) * | 2020-03-17 | 2020-07-10 | 京东方科技集团股份有限公司 | 光标尺、坐标系配准的方法 |
| CN111388089A (zh) * | 2020-03-19 | 2020-07-10 | 京东方科技集团股份有限公司 | 治疗设备及其配准方法、配准装置 |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115631170A (zh) * | 2022-10-28 | 2023-01-20 | 中山大学 | 一种高精度线位移测量方法及其系统 |
| CN115737122A (zh) * | 2022-10-28 | 2023-03-07 | 浙江久择医疗科技有限公司 | 一种借助x射线的统一断层影像和磁定位工具坐标的方法 |
| CN115795579A (zh) * | 2022-12-23 | 2023-03-14 | 岭南师范学院 | 一种无特征复杂曲面误差分析的快速坐标对齐方法 |
| CN115880469A (zh) * | 2023-02-20 | 2023-03-31 | 江苏省人民医院（南京医科大学第一附属医院） | 一种表面点云数据与三维影像的配准方法 |
| CN116616869A (zh) * | 2023-03-03 | 2023-08-22 | 无锡艾米特智能医疗科技有限公司 | 一种超声点云的精准模型构建方法 |
| CN119139021A (zh) * | 2023-06-14 | 2024-12-17 | 武汉联影智融医疗科技有限公司 | 手术坐标系的配准方法、装置和存储介质 |
| CN119184851A (zh) * | 2023-06-26 | 2024-12-27 | 武汉联影智融医疗科技有限公司 | 空间配准方法及装置、机器人辅助系统、设备和介质 |
| CN116672082B (zh) * | 2023-07-24 | 2024-03-01 | 苏州铸正机器人有限公司 | 一种手术导航标尺的导航配准方法及装置 |
| CN116672082A (zh) * | 2023-07-24 | 2023-09-01 | 苏州铸正机器人有限公司 | 一种手术导航标尺的导航配准方法及装置 |
| CN116993791B (zh) * | 2023-09-28 | 2023-12-19 | 真健康（北京）医疗科技有限公司 | 一种基于体表定位装置的医学图像配准方法及设备 |
| CN116993791A (zh) * | 2023-09-28 | 2023-11-03 | 真健康（北京）医疗科技有限公司 | 一种基于体表定位装置的医学图像配准方法及设备 |
| CN117474966A (zh) * | 2023-11-21 | 2024-01-30 | 上海卓昕医疗科技有限公司 | 医学影像与显影设备的配准方法、装置、设备和介质 |
| CN118914253A (zh) * | 2024-10-12 | 2024-11-08 | 浙江双元科技股份有限公司 | 一种非理想条件下的平面ct系统参数标定方法及装置 |
| CN119399308A (zh) * | 2024-11-06 | 2025-02-07 | 中国科学院高能物理研究所 | 一种成像几何在线校准方法 |
| CN120531421A (zh) * | 2025-02-08 | 2025-08-26 | 北京和华瑞博医疗科技有限公司 | 定位系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112451093B (zh) | 2021-05-04 |
| CN112451093A (zh) | 2021-03-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112451093B (zh) | 图像引导机器人微创外科手术的物理空间与图像空间配准方法 | |
| US12106497B2 (en) | Method and navigation system for registering two-dimensional image data set with three-dimensional image data set of body of interest | |
| US8081808B2 (en) | Retinal thickness measurement by combined fundus image and three-dimensional optical coherence tomography | |
| US10074199B2 (en) | Systems and methods for tissue mapping | |
| Gerber et al. | High-accuracy patient-to-image registration for the facilitation of image-guided robotic microsurgery on the head | |
| CN109464196A (zh) | 采用结构光影像配准的手术导航系统及配准信号采集方法 | |
| CN113347937A (zh) | 参照系的配准 | |
| WO2018218611A1 (zh) | 锥形束计算机断层成像系统的几何参数确定方法 | |
| JP2019511268A (ja) | 脳深部刺激療法の電極の三次元画像における回転配向の決定 | |
| CN113892960B (zh) | X射线自成像几何标定方法及装置 | |
| CN112617877B (zh) | 移动ct系统的自主扫描方法、存储介质和ct扫描装置 | |
| CN116883471B (zh) | 面向胸腹部经皮穿刺的线结构光无接触点云配准方法 | |
| CN113450395B (zh) | 基于光学跟踪探针划取点的手术导航空间配准方法及装置 | |
| CN115775611B (zh) | 一种穿刺手术规划系统 | |
| US20140294276A1 (en) | Automatic estimation of anatomical extents | |
| CN107595287A (zh) | 一种将鲤鱼磁共振扫描坐标转换为脑立体定位坐标的方法 | |
| CN114947943B (zh) | 非接触式超声探头快速重标定方法及装置 | |
| US20080037848A1 (en) | System and Method for Corpus Callosum Segmentation in Magnetic Resonance Images | |
| CN114404039B (zh) | 三维模型的组织漂移校正方法、装置、电子设备及存储介质 | |
| WO2001057805A2 (en) | Image data processing method and apparatus | |
| Sun et al. | Stepwise local stitching ultrasound image algorithms based on adaptive iterative threshold Harris corner features | |
| CN115272288B (zh) | 一种医学图像标记点自动识别方法、电子设备及存储介质 | |
| CN121724982A (zh) | 一种基于人工智能的针灸穴位定位方法及系统 | |
| CN112132903A (zh) | 一种视觉系统与多轴运动系统的坐标系标定方法及系统 | |
| CN1794301A (zh) | 手术导航中基于基准面膜的多模式医学图像配准方法 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21921983 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 21921983 Country of ref document: EP Kind code of ref document: A1 |