Depth camera based radiotherapy position monitoring system

Abstract

The application relates to a radiotherapy body position monitoring system based on a depth camera, comprising a coordinate system conversion system and a body position monitoring module; the coordinate system conversion system is used for determining the conversion relationship between a world coordinate system and a positioning CT coordinate system; the body position monitoring module: acquires a patient body position point cloud through a depth camera, and separates a target point cloud P; positioning CT data of a patient acquired through a positioning CT system is introduced into a point cloud processing software to separate a positioning CT point cloud of the patient, and the positioning CT point cloud is converted into a world coordinate system through the conversion relationship between the world coordinate system and the positioning CT coordinate system to form a positioning point cloud P CT‑sim The target point cloud P and the positioning point cloud P CT‑sim are subjected to real-time point cloud accurate registration calculation of six-dimensional errors, and the results are output. The application can guide the adjustment of the patient body position with specific error values, and can also monitor the body position in real time during radiotherapy.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of body position monitoring in radiotherapy, and in particular to a radiotherapy body position monitoring system based on a depth camera. BACKGROUND

[0002] Radiotherapy is one of the three main methods of clinical treatment of tumors, and most patients with malignant tumors need to receive radiotherapy to some extent. The purpose of radiotherapy is to kill tumor cells as much as possible while sparing or minimizing the irradiation of surrounding normal tissues and organs. Studies have shown that inaccurate patient positioning during radiotherapy can greatly increase the recurrence rate of patients. Cone beam CT (CBCT) is the most commonly used image-guided radiotherapy technology. After CT image reconstruction before treatment, a three-dimensional image model is obtained, which is compared with the patient model of the treatment plan to calculate the parameters that need to be adjusted for the treatment bed. However, CBCT has the disadvantage of radiation.

[0003] However, radiotherapy is a long process with many factors affecting the accuracy of patient positioning. The positioning before radiotherapy is based on the marker lines on the patient's body surface, but the marker lines are easily blurred, and for obese patients or patients with loose skin, the positioning error based on the marker lines on the body surface is large, so new methods need to be proposed to assist technical personnel in positioning.

[0004] The factors affecting the change of patient position during radiotherapy mainly include the following: (1) Respiratory motion and other autonomous movements of tissues cause irregular movement of tumor tissues, which is mainly concentrated in the chest and abdomen. Patients may take chest breathing or abdominal breathing, or may have excessive inhalation and exhalation during radiotherapy, causing movement of the target area. (2) For non-rigid organs such as the breast, the degree of activity is large, and positioning errors are likely to occur between fractions, so during radiotherapy, the chest wall, arms, and other parts must be kept in relatively consistent positions and cannot be changed arbitrarily. The existing CBCT technology cannot confirm whether the patient's limbs remain in the same position during treatment. (3) For patients in the pelvic region, the degree of bladder filling has a great influence on the position of the tumor, the position of the organ at risk (OAR), and the position of the patient's surface positioning line. (4) Children, the elderly, or unconscious patients may have involuntary movements during treatment, and common radiotherapy guidance techniques are difficult to monitor such changes in body position. Therefore, new methods need to be proposed to monitor the patient's position in real time during radiotherapy and to reflect the changes in the patient's position to technical personnel in real time.

[0005] Patent No. 201610464508.6 proposes a position monitoring method for radiotherapy, which places target markers on the patient's body surface and reference markers in a local coordinate system. A monitoring module receives signals from the target marker blocks and reference marker blocks to complete position monitoring. However, the markers are prone to detachment and displacement, which greatly affects patient position tracking. Secondly, the displacement of individual markers is insufficient to replace the displacement of the entire target area, and the placement of the patient's limbs outside the target area also has a significant impact on the displacement of the target area, especially for breast and cervical patients.

[0006] Patent No. 201811579081.X proposes a radiotherapy-assisted system using infrared and visible light three-dimensional reconstruction. The patient's positioning CT data is imported into the treatment planning system for planning, including target delineation and dose calculation. Therefore, the patient's position during treatment should be based on the positioning CT position. However, this invention does not register the reconstructed body surface with the positioning CT, nor does it calculate the error between the actual and positioning positions. Furthermore, the reconstruction effects of visible and infrared light are significantly affected by the environment, making monitoring challenging. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a radiotherapy position monitoring system based on a depth camera, which can guide the adjustment of the patient's position with specific error values ​​and can also monitor the position in real time during radiotherapy.

[0008] The technical solution adopted by this invention to solve its technical problem is:

[0009] A depth camera-based system for monitoring patient positioning during radiotherapy.

[0010] Includes a coordinate system transformation system and a body position monitoring module;

[0011] The coordinate system transformation system is used to determine the transformation relationship between the world coordinate system and the positioning CT coordinate system.

[0012] The body position monitoring module acquires the patient's body position cloud using a depth camera and segments the target point cloud P. It then imports the patient's positioning CT data acquired through the positioning CT system into the point cloud processing software to segment the patient's positioning CT point cloud, and finally uses the transformation relationship between the world coordinate system and the positioning CT coordinate system to segment the patient's positioning CT point cloud. Transform into the world coordinate system to form a point cloud P CT-sim The target point cloud P is compared with the local point cloud P. CT-sim Perform real-time point cloud fine registration calculation of six-dimensional error and output the results;

[0013] The coordinate system of the depth camera is the world coordinate system, and the coordinate system of the localization CT system is the localization CT coordinate system.

[0014] Furthermore, the coordinate system transformation system determines the transformation relationship between the world coordinate system and the positioning CT coordinate system. The method is as follows:

[0015] The calibration phantom was fixed at the isocenter of the accelerator for CBCT scanning. The obtained CBCT data was reconstructed and the CBCT point cloud Q was segmented. phantom ;

[0016] A depth camera fixed above the treatment bed is used to acquire the surface point cloud of the calibration phantom, and the calibration phantom point cloud P is segmented. phantom ;

[0017] The calibration phantom point cloud P phantom Q with CBCT point cloud phantom Registration determines the transformation relationship between the world coordinate system and the CBCT coordinate system.

[0018] The bed movement relationship between the CBCT coordinate system and the localization CT coordinate system is obtained through the TPS treatment planning system. This determines the transformation relationship between the world coordinate system and the positioning CT coordinate system.

[0019]

[0020] Furthermore, the body position monitoring module includes a radiotherapy positioning verification module and / or a radiotherapy real-time monitoring module;

[0021] The radiotherapy positioning verification module is used to verify the patient's position before radiotherapy.

[0022] The radiotherapy real-time monitoring module is used for real-time monitoring of the patient's position during radiotherapy.

[0023] Furthermore, the radiotherapy positioning verification module: acquires point cloud data of the patient's pre-radiotherapy positioning using a depth camera, and manually segments the patient's positioning point cloud information P from the acquired point cloud information. patient The positioning CT data obtained from the patient's positioning CT system is imported into point cloud processing software to segment the patient's positioning CT point cloud, and then the transformation relationship between the world coordinate system and the positioning CT coordinate system is used. Transform into the world coordinate system to form a point cloud P CT-sim The segmented patient placement point cloud information P patient With positioning point cloud P CT-sim Perform point cloud fine registration calculations for six-dimensional errors and output the results.

[0024] Furthermore, the radiotherapy real-time monitoring module: acquires the patient's body point cloud during radiotherapy in real time using a depth camera, and automatically segments the real-time target point cloud P.realtime-patient ; the positioning CT data of the patient obtained by the positioning CT system is imported into the point cloud processing software to segment the positioning CT point cloud of the patient, and the conversion relationship between the world coordinate system and the positioning CT coordinate system is used to convert the positioning point cloud P into the world coordinate system CT-sim , the real-time target point cloud P realtime-patient is converted into the world coordinate system to form the positioning point cloud P CT-sim , and the real-time point cloud precision registration calculation six-dimensional error is carried out.

[0025] Further, the point cloud precision registration criterion is:

[0026] The minimum distance of the points in the point cloud to the plane where the corresponding points of the target point cloud are located is used as the registration criterion:

[0027]

[0028] Wherein, T represents a transformation matrix, p i , q i respectively represent the corresponding points in the source point cloud and the target point cloud, n i represents the normal vector of the corresponding point q i .

[0029] Further, the radiotherapy position monitoring system based on the depth camera further comprises

[0030] The mixed reality guided radiotherapy positioning module is used for guiding the radiotherapy positioning of the patient.

[0031] The mixed reality guided radiotherapy positioning module: the positioning CT system is used for positioning CT scanning of the patient to obtain the positioning CT data and reconstruct a three-dimensional virtual model of the patient, and a mixed reality model is generated by rendering at the isocenter of the accelerator.

[0032] The operator uses the HoloLens mixed reality glasses to guide the radiotherapy positioning by taking the mixed reality model as a standard.

[0033] The beneficial effects of the present application are:

[0034] (1) By registering the actual position of the patient with the upper surface point cloud and the positioning CT upper surface point cloud, the six-dimensional error is calculated, and the specific error value is used to guide the adjustment of the patient's position.

[0035] (2) In order to solve the problem of body position change in radiotherapy, the present application introduces a real-time body position monitoring module in radiotherapy, which automatically extracts the patient's surface point cloud at a speed of milliseconds / frame and automatically registers and calculates the error with the positioning CT point cloud. If the error is greater than the set error, an alarm is issued to realize real-time body position monitoring.

[0036] ​(3) The application adopts a fixed depth camera to acquire a patient body surface point cloud, takes the coordinate system of the depth camera as a real physical space, i.e. a world coordinate system, and determines the conversion relationship between the world coordinate system and the positioning CT coordinate system through a calibration phantom, which can be used for all patients, so that the coordinate system conversion relationship only needs to be calibrated and confirmed once, without the need for regular calibration.

[0037] (4) To solve the error caused by the operator positioning according to the patient body surface marker line before radiotherapy, the application introduces a mixed reality guided radiotherapy module. The mixed reality guided radiotherapy module can present the positioning CT body position of the patient in a visual form at the isocenter of the accelerator, realize mixed reality three-dimensional visualization, and the technician can intuitively see the positioning CT body position at the isocenter of the accelerator through the HoloLens glasses, and position the patient according to the positioning CT body position, which has better intuitiveness and effectively guides the positioning of the patient during radiotherapy. BRIEF DESCRIPTION OF DRAWINGS

[0038] The application will be further described below in combination with the drawings and embodiments.

[0039] Figure 1 A flowchart for point cloud real-time body position monitoring by the depth camera-based radiotherapy body position monitoring system of the application;

[0040] Figure 2 A conversion diagram of the relationship between the positioning CT coordinate system and the world coordinate system in the application;

[0041] Figure 3 A mixed reality model effect diagram for guiding the patient positioning by the application;

[0042] Figure 4 A structure diagram of the calibration phantom in the application; DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the application will be clearly and completely described below in combination with the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application.

[0044] As Figures 1-4 A depth camera-based radiotherapy body position monitoring system, comprising a coordinate system conversion system and a body position monitoring module.

[0045] The coordinate system conversion system is used to determine the conversion relationship between the world coordinate system and the positioning CT coordinate system.

[0046] The body position monitoring module: acquires the patient body position point cloud through the depth camera, and segments the target point cloud P; imports the positioning CT data of the patient acquired through the positioning CT system into the point cloud processing software to segment the positioning CT point cloud of the patient, and then converts the relationship between the world coordinate system and the positioning CT coordinate system Converts into the world coordinate system to form the positioning point cloud P CT-sim The target point cloud P and the positioning point cloud P CT-sim are matched, the six-dimensional error is calculated in real time, and the result is output;

[0047] The coordinate system of the depth camera is the world coordinate system, and the coordinate system of the positioning CT system is the positioning CT coordinate system.

[0048] The coordinate system conversion system determines the conversion relationship between the world coordinate system and the positioning CT coordinate system The method is as follows:

[0049] The calibration phantom is fixed at the isocenter of the accelerator for CBCT scanning, the obtained CBCT data is reconstructed and segmented to obtain the CBCT point cloud Q phantom ;

[0050] The surface point cloud of the calibration phantom is acquired by using the depth camera fixed above the treatment bed, and the calibration phantom point cloud P phantom is segmented;

[0051] The calibration phantom point cloud P phantom and the CBCT point cloud Q phantom are matched to determine the conversion relationship between the world coordinate system and the CBCT coordinate system

[0052] The bed-moving relationship between the CBCT coordinate system and the positioning CT coordinate system is acquired through the TPS treatment planning system Thus, the conversion relationship between the world coordinate system and the positioning CT coordinate system is determined

[0053]

[0054] Preferably, the body position monitoring module comprises a radiotherapy positioning verification module and / or a radiotherapy real-time monitoring module;

[0055] The radiotherapy positioning verification module is used for verifying the patient body position before radiotherapy.

[0056] The radiotherapy real-time monitoring module is used for real-time monitoring of the patient body position during radiotherapy.

[0057] The radiotherapy positioning verification module: acquires the point cloud of the patient body position before radiotherapy through the depth camera, and manually segments the positioning point cloud information P of the patient from the acquired point cloud informationpatient ; import the positioning CT data obtained by the patient through the positioning CT system into the point cloud processing software to segment the positioning CT point cloud of the patient, and convert the positioning CT point cloud into the positioning point cloud P in the world coordinate system through the conversion relationship between the world coordinate system and the positioning CT coordinate system CT-sim ; import the segmented positioning point cloud information P patient of the patient into the point cloud processing software, and perform point cloud fine registration calculation on the positioning point cloud P CT-sim and the real-time target point cloud P realtime-patient , and output the result.

[0058] The real-time monitoring module of radiotherapy: through a depth camera, real-time body position point cloud of a patient during radiotherapy is obtained, and real-time target point cloud P realtime-patient ; import the positioning CT data obtained by the patient through the positioning CT system into the point cloud processing software to segment the positioning CT point cloud of the patient, and convert the positioning CT point cloud into the positioning point cloud P in the world coordinate system through the conversion relationship between the world coordinate system and the positioning CT coordinate system CT-sim , real-time target point cloud P realtime-patient is imported into the point cloud processing software, and real-time point cloud fine registration calculation on the positioning point cloud P CT-sim and the real-time target point cloud P i is performed, and the result is output.

[0059] Preferably, the point cloud fine registration criterion is:

[0060] The registration criterion is to minimize the distance of the points in the point cloud to the plane on which the corresponding points of the target point cloud lie:

[0061]

[0062] Wherein, T represents a transformation matrix, p i , q i respectively represent corresponding points in the source point cloud and the target point cloud, n i represents the normal vector of the corresponding point q i .

[0063] Preferably, the radiotherapy body position monitoring system based on the depth camera further comprises

[0064] A mixed reality guided radiotherapy positioning module for guiding the radiotherapy positioning of the patient;

[0065] The mixed reality guided radiotherapy positioning module: a positioning CT system is used to perform positioning CT scanning on the patient to obtain positioning CT data and reconstruct a three-dimensional virtual model of the patient, and a mixed reality model is generated by rendering at the isocenter of the accelerator;

[0066] An operator uses HoloLens mixed reality glasses to guide the radiotherapy positioning according to the mixed reality model.

[0067] ReferenceFigure 1 The method for mixed reality guided positioning and point cloud real-time position monitoring by using the above-mentioned deep camera-based radiotherapy position monitoring system comprises the following steps:

[0068] Step S1, determine the conversion relationship between the world coordinate system and the positioning CT coordinate system:

[0069] The coordinate conversion relationship is shown in Figure 2 The conversion relationship between the world coordinate system and the CBCT coordinate system is determined by using the calibration phantom. The surface marker line of the calibration phantom is aligned with the laser lamp, the calibration phantom is ensured to be located at the isocenter of the accelerator to obtain CBCT data, and the surface point cloud of the calibration phantom at the isocenter of the accelerator is obtained by using the depth camera fixed above the treatment bed, and the calibration phantom point cloud P phantom is manually segmented. The CBCT data is reconstructed and the CBCT point cloud Q phantom is segmented. The coordinate system of the depth camera is taken as the world coordinate system, and the coordinate system of the positioning CT system is taken as the positioning CT coordinate system.

[0070] The calibration phantom point cloud P phantom is registered with the CBCT point cloud Q phantom , and the conversion relationship between the world coordinate system and the CBCT coordinate system is determined

[0071] The bed-moving relationship between the CBCT coordinate system and the positioning CT coordinate system is obtained by the TPS treatment planning system , so as to determine the conversion relationship between the world coordinate system and the positioning CT coordinate system

[0072] The TPS treatment planning system (radiotherapy planning system) is a kind of medical equipment, which models the radiation source and the patient to simulate the planned implementation of radiotherapy.

[0073] The point cloud registration method adopted by the application mainly includes principal component analysis (PCA) point cloud coarse registration and ICP fine registration. The registration root mean square error is between 0.5mm-1.5mm, which meets the coordinate system conversion accuracy requirement.

[0074] Further, the registration method based on PCA mainly utilizes the principal axis direction of the point cloud data for registration. First, the covariance matrix of the two groups of point cloud data is calculated, the main feature component is calculated according to the covariance matrix, that is, the principal axis direction of the point cloud data, then the rotation matrix is obtained through the principal axis direction, and the translation vector is obtained by calculating the center coordinates of the two groups of point cloud data.

[0075] Further, the application adopts point-to-plane ICP fine registration. The distance from the points in the point cloud to the plane where the corresponding points of the target point cloud are located is minimized as the registration criterion:

[0076]

[0077] where T represents a transformation matrix, p i and q i represent corresponding points in the source point cloud and the target point cloud respectively, n i represents the normal vector of the corresponding point q i Compared with the traditional ICP algorithm, the point-to-plane ICP can better reflect the spatial structure of the point cloud, can better resist false corresponding point pairs, and has faster iteration speed.

[0078] Step S2, mixed reality guided radiotherapy positioning:

[0079] The positioning CT data is imported into the TPS treatment planning system for treatment planning design, and the exported DICOM-RT data is three-dimensionally reconstructed to form a three-dimensional virtual model, mainly reconstructing patient contours, PTV, CTV, OAR, etc. The image marker is fixed at the center of the accelerator, and the HoloLens glasses capture the reconstructed three-dimensional virtual model fixed at the center of the accelerator, which is marked as the gold standard to guide the radiotherapy patient positioning.

[0080] Further, the HoloLens glasses used in the application are self-positioned by the simultaneous localization and mapping (SLAM) technology. The SLAM includes environment perception cameras, depth cameras, inertial measurement units (IMU) and other sensors. The IMU and other sensors perceive the direction of the HoloLens, the environment perception camera senses the offset of the relative position of the HoloLens, and the depth camera perceives the environment around the HoloLens.

[0081] Step S3, pre-radiotherapy body position positioning verification based on point cloud:

[0082] An Intel RealSense D435i depth camera is used to collect point clouds. Intel RealSenseViewer software is used to obtain scene point clouds, and MeshLab software is used to segment patient positioning point cloud information P patient The patient is scanned by a positioning CT system to obtain a positioning CT point cloud, which is converted into a world coordinate system through the conversion relationship between the world coordinate system and the CT coordinate system to form a positioning point cloud P CT-sim ; the segmented patient positioning point cloud information P patient is point cloud ICP precise registration calculated six-dimensional error with the positioning point cloud P CT-sim .

[0083] The calculation of the six-dimensional error is as follows: the rotation matrix R and the translation vector t of the two pieces of point clouds are obtained, and finally the rotation error alpha, beta and gamma are calculated according to the relationship between the rotation matrix R and the Euler angle, the translation error is calculated according to the translation vector t, and the patient position is further adjusted.

[0084] Further, the fine registration method adopted by the application is a point-to-surface fine registration iterative closest point (ICP).

[0085] Further, the curvature downsampling method is used to filter the point cloud before point cloud registration. That is, the more the curvature of the point cloud, the more the number of sampling points. This sampling method has the characteristics of high calculation efficiency, high stability, strong noise resistance and uniform local distribution of sampling points.

[0086] Further, the radius outlier removal method is used to remove point cloud noise before point cloud registration. The principle is that if a certain range of a point does not reach enough neighbors, the point will be considered as noise and removed.

[0087] Further, the relationship between the rotation matrix R and the Euler angle is:

[0088]

[0089]

[0090]

[0091] Wherein alpha is the x-axis deflection angle, beta is the y-axis deflection angle, and gamma is the z-axis deflection direction.

[0092] If the matrix obtained after ICP transformation is:

[0093]

[0094] The six-dimensional error to be solved is:

[0095]

[0096] Step S4, real-time monitoring of the body position in the point cloud-based radiotherapy:

[0097] The application adopts Intel RealSense D435i as a depth camera to obtain real-time point cloud information. The depth camera obtains an RGB-D image, wherein the RGB picture provides x, y coordinates in the pixel coordinate system, and the depth map directly provides z coordinates in the camera coordinate system, that is, the distance between the camera and the point. According to the information of the RGB-D image and the camera internal parameter, the coordinates of any pixel point in the camera coordinate system can be calculated. According to the camera external parameter, the coordinates of any pixel point in the world coordinate system can be calculated.

[0098] (1) Use a depth camera to acquire real-time point cloud: First, the depth camera acquires RGB-D images, which are then converted into three-dimensional point cloud coordinate information to complete the real-time acquisition of point cloud.

[0099] The specific process is as follows: First, the depth camera acquires RGB-D images; second, points are calculated using the realsnese2 library functions; next, the points and their corresponding pixel coordinates are obtained through the relevant properties of the points; finally, the point cloud is visualized by rendering the 3D point data to achieve real-time acquisition of the point cloud.

[0100] (2) Selection of point cloud processing area: There is a lot of noise, non-target point cloud and other interference information in the real-time point cloud acquired by the depth camera. By selecting the area, the range of the point cloud to be processed is defined. Generally, the area within the treatment bed is selected to remove a lot of non-target information, which helps the real-time operation of monitoring.

[0101] (3) Real-time segmentation of target point cloud: Real-time point cloud processing is performed on the selected point cloud processing area to segment the real-time target point cloud information of the patient's upper surface, mainly including planar segmentation and point cloud normal vector segmentation.

[0102] Specifically, the process involves: extracting the treatment bed board portion through point cloud planar segmentation; removing the treatment bed board yields the patient's point cloud portion. Then, the patient's point cloud portion is segmented using point cloud normal vectors. Based on the normal vector direction, the upper surface point cloud of the patient is selected to complete the extraction of the patient's target point cloud, forming the real-time target point cloud P. realtime-patient .

[0103] (4) Real-time calculation of the error between patient positioning and localization: Import the patient positioning CT data into the point cloud processing software to segment the positioning point cloud of the upper surface of the patient, and calculate the error between the world coordinate system and the positioning CT coordinate system through the transformation relationship. Transform into the world coordinate system to form a point cloud P CT-sim At this time, the real-time target point cloud P realtime-patient Positioning point cloud P CT-sim The positional relationship refers to the relationship between the patient's real-time position and the position at the time of positioning, which is the real-time target point cloud P. realtime-patient With positioning point cloud P CT-sim Calculate the six-dimensional error of point cloud fine registration.

[0104] (5) Determine the relationship between the error and the set threshold: Compare the calculated six-dimensional error with the set threshold. If the error is within the threshold range, continue radiotherapy. If it exceeds the threshold, issue a warning and reposition the patient.

[0105] The calculation of six-dimensional error is as follows: the rotation matrix R and the translation vector t of the two point clouds are obtained, the rotation error α, β and γ is calculated according to the relationship between the rotation matrix R and the Euler angle, the translation error is calculated according to the translation vector t, and the patient position is further adjusted.

[0106] After the real-time calculation of the translation and rotation errors is completed, it is necessary to compare with the set threshold. The present application sets that the translation error of each axis of the head and neck patient cannot be greater than 3mm, the translation error of each axis of the chest and abdomen patient cannot be greater than 5mm, and the rotation error of all patients cannot be greater than 3°. If the translation error and the rotation error are less than the set threshold, the radiotherapy is continued, and if the threshold is exceeded, a warning is issued and the patient is repositioned.

[0107] The present application uses a fixed depth camera to obtain the patient surface point cloud, takes the coordinate system of the depth camera as the real physical space, i.e. the world coordinate system, determines the conversion relationship between the world coordinate system and the positioning CT coordinate system through the calibration phantom, and this conversion relationship can be used for all patients, so that the coordinate system conversion relationship only needs to be calibrated and confirmed once, without the need for regular calibration.

[0108] To solve the error caused by the technician positioning according to the patient surface marker line before radiotherapy, the present application introduces the mixed reality guided radiotherapy technology and the pre-radiotherapy position verification technology. The mixed reality guided radiotherapy technology can present the positioning CT position at the isocenter of the accelerator in a visual form, realize mixed reality three-dimensional visualization, and the technician can intuitively see the positioning CT position at the isocenter of the accelerator through the HoloLens glasses, and position the patient according to the positioning CT position, which has better intuitiveness. The pre-radiotherapy position verification technology can calculate the six-dimensional error by registering the actual position point cloud of the patient with the upper surface point cloud of the positioning CT, and guide the adjustment of the patient position with specific error values.

[0109] To solve the problem of body position change during radiotherapy, the present application introduces the real-time monitoring technology of body position during radiotherapy to automatically extract the patient surface point cloud at a millisecond level / frame speed and automatically register and calculate the error with the positioning CT point cloud, and if the error is greater than the set error, an alarm is issued.

[0110] The present application uses a depth camera to obtain a large range of patient surface point clouds, and divides the region where the PTV is located into ROI1 region and the region outside the PTV into ROI2 region. The body position inside and outside the target region can be monitored at the same time, and the monitoring accuracy is increased.

[0111] The above merely describes preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art, according to the technical solution and inventive concept of the present application, makes equivalent replacement or change within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A radiotherapy position monitoring system based on a depth camera, characterized in that: Includes a coordinate system transformation system and a body position monitoring module; The coordinate system transformation system is used to determine the transformation relationship between the world coordinate system and the positioning CT coordinate system. The body position monitoring module acquires the patient's body position cloud using a depth camera and segments the target point cloud P. It then imports the patient's positioning CT data acquired through the positioning CT system into the point cloud processing software to segment the patient's positioning CT point cloud, and finally uses the transformation relationship between the world coordinate system and the positioning CT coordinate system to segment the patient's positioning CT point cloud. Transform into the world coordinate system to form a point cloud P CT-sim The target point cloud P is compared with the local point cloud P. CT-sim Perform real-time point cloud fine registration calculation of six-dimensional error and output the results; The coordinate system of the depth camera is the world coordinate system, and the coordinate system of the localization CT system is the localization CT coordinate system. The coordinate system transformation system determines the transformation relationship between the world coordinate system and the positioning CT coordinate system. The method is as follows: The calibration phantom was fixed at the isocenter of the accelerator for CBCT scanning. The obtained CBCT data was reconstructed and the CBCT point cloud Q was segmented. phantom ; A depth camera fixed above the treatment bed is used to acquire the surface point cloud of the calibration phantom, and the calibration phantom point cloud P is segmented. phantom ; The calibration phantom point cloud P phantom Q with CBCT point cloud phantom Registration determines the transformation relationship between the world coordinate system and the CBCT coordinate system. The bed movement relationship between the CBCT coordinate system and the localization CT coordinate system is obtained through the TPS treatment planning system. This determines the transformation relationship between the world coordinate system and the positioning CT coordinate system.

2. The radiotherapy position monitoring system based on a depth camera according to claim 1, characterized in that: The body position monitoring module includes a radiotherapy positioning verification module and / or a radiotherapy real-time monitoring module; The radiotherapy positioning verification module is used to verify the patient's position before radiotherapy. The radiotherapy real-time monitoring module is used for real-time monitoring of the patient's position during radiotherapy.

3. The radiotherapy position monitoring system based on a depth camera according to claim 2, characterized in that: The radiotherapy positioning verification module: uses a depth camera to collect point cloud data of the patient's pre-radiotherapy positioning, and manually segments the patient's positioning point cloud information P from the collected point cloud information. patient The positioning CT data obtained from the patient's positioning CT system is imported into point cloud processing software to segment the patient's positioning CT point cloud, and then the transformation relationship between the world coordinate system and the positioning CT coordinate system is used. Transform into the world coordinate system to form a point cloud P CT-sim The segmented patient placement point cloud information P patient With positioning point cloud P CT-sim Perform point cloud fine registration calculations for six-dimensional errors and output the results.

4. The radiotherapy position monitoring system based on a depth camera according to claim 2, characterized in that: The radiotherapy real-time monitoring module acquires the body point cloud of the patient during radiotherapy in real time using a depth camera, and automatically segments the real-time target point cloud P. realtime-patient The positioning CT data obtained from the patient's positioning CT system is imported into point cloud processing software to segment the patient's positioning CT point cloud, and then the transformation relationship between the world coordinate system and the positioning CT coordinate system is used. Transform into the world coordinate system to form a point cloud P CT-sim Real-time target point cloud P realtime-patient With positioning point cloud P CT-sim Perform real-time point cloud fine registration calculation of six-dimensional error and output the results.

5. The radiotherapy position monitoring system based on a depth camera according to claim 1, 3, or 4, characterized in that: The point cloud registration criteria are as follows: The registration criterion is to minimize the plane distance between a point in the point cloud and the corresponding point in the target point cloud. Where T represents the transformation matrix, p i q i Let n represent the corresponding points in the source point cloud and the target point cloud, respectively. i q i The normal vector of the corresponding point.

6. The radiotherapy position monitoring system based on a depth camera according to claim 1, characterized in that: Also includes A mixed reality-guided radiotherapy positioning module is used to guide patients in radiotherapy positioning. Mixed Reality Guided Radiotherapy Positioning Module: The module uses a positioning CT system to perform a positioning CT scan on the patient to obtain positioning CT data and reconstruct a three-dimensional virtual model of the patient. A mixed reality model is then rendered and generated at the center of the accelerator and other locations. Operators used HoloLens mixed reality glasses to guide radiotherapy positioning using a mixed reality model as a standard.

Images