Radiotherapy attitude monitoring method and system based on three-dimensional human skeleton tracking, electronic equipment and storage medium

Abstract

The invention belongs to the field of radiotherapy, and provides a radiotherapy attitude monitoring method and system based on three-dimensional human skeleton tracking, electronic equipment and a storage medium, and the method comprises the steps: skeleton node information identification, projection coordinate data calculation, angle data calculation, risk point identification, rotation angle collection and collision risk judgment. According to the invention, space pose information analysis of key skeleton nodes is carried out by adopting an unmarked three-dimensional vision technology, and collision risk assessment is carried out in combination with the motion trail of the linear accelerator, so that a dangerous angle interval can be accurately predicted, a multi-stage early warning mechanism is triggered, and the safety and continuity of a treatment process are ensured; and treatment interruption caused by body position deviation or equipment interference is avoided.

Description

Technical Field

[0001] The present invention relates to the field of radiotherapy, and in particular to a radiotherapy posture monitoring method, system, electronic equipment and storage medium based on three-dimensional human skeleton tracking. Background Art

[0002] In recent years, breakthroughs in cutting-edge technologies such as artificial intelligence, the Internet of Things, and big data are accelerating the intelligent transformation of radiotherapy systems. As a core component of precision oncology treatment, modern radiotherapy systems deeply integrate multidisciplinary technologies such as medical imaging, radiation physics, and automatic control. Through intelligent optimization of treatment planning systems, image-guided real-time positioning, and precise delivery of linear accelerators, they significantly improve the effectiveness and safety of cancer treatment. However, in clinical practice, human-machine collaborative management of radiotherapy still faces significant challenges. On the one hand, prolonged radiotherapy can easily lead to unconscious positional shifts or voluntary movements in patients, directly affecting radiation delivery accuracy and increasing the risk of normal tissue exposure. On the other hand, while traditional rigid fixation devices can restrict patient movement, they can reduce treatment comfort and cause adverse reactions such as skin compression. How to achieve dynamic patient posture monitoring and intelligent obstacle avoidance for radiotherapy equipment without relying on physical restraints, thereby building a safe, precise, and user-friendly closed-loop radiotherapy control system, has become a technical bottleneck that urgently needs to be overcome in the development of intelligent radiotherapy. Currently, research on non-contact human body monitoring methods for radiotherapy is still in the exploratory stage, and relatively few related technologies are available globally. Based on existing publicly available technologies, a few companies abroad have conducted preliminary research and development of optical body surface monitoring systems. For example, the existing AlignRT and CatalystHD systems integrate body monitoring capabilities during radiotherapy. However, it should be noted that these systems are primarily limited to respiratory motion tracking and alignment deviation correction, enabling only single-dimensional monitoring of physiological signals and not dynamic capture and analysis of full-body posture. To enable body monitoring during radiotherapy, researchers have proposed a contact-based body position monitoring system that uses pressure sensors and artificial intelligence to achieve non-invasive, radiation-free, and real-time monitoring. Regarding non-contact radiotherapy body monitoring technology, researchers have developed a 2D image-based patient posture recognition system using deep learning and ResNet-18 to detect incorrect patient posture and alert technicians. Researchers compare the current laser position of a marker with an initial reference position acquired during simulation and the first irradiation session to perform patient alignment and position monitoring, aiming to measure the accuracy of traditional laser alignment techniques for patient repositioning. Researchers have developed an infrared interactive patient position guidance and acquisition control system for clinical use, consisting of an infrared camera, infrared markers, and specialized software. The use of multiple positioning markers visualizes patient posture in real time and provides feedback on misalignment and required postural adjustments.

[0003] However, overall, this field still has significant technical gaps in multimodal data fusion, robustness in complex scenarios, and clinical multi-center verification. In particular, there is a lack of a monitoring-prediction-control closed-loop solution covering the entire radiotherapy process. Summary of the Invention

[0004] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide a radiotherapy posture monitoring method, system, electronic device and storage medium based on three-dimensional human skeleton tracking to solve the problems of skin damage and respiratory function restriction caused by rigid fixation.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A radiotherapy posture monitoring method based on three-dimensional human skeleton tracking, comprising:

[0007] Using a 3D vision sensor to capture images of the target area to obtain a three-dimensional depth image containing the target human body, and using an image recognition algorithm to detect the three-dimensional depth image to obtain skeleton node information;

[0008] Mapping the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data;

[0009] Calculating the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data;

[0010] Calculating a straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determining a projection point whose distance data does not meet a safety distance threshold as a risk point;

[0011] Acquiring the rotation angle of the linear accelerator treatment head at the current moment;

[0012] When the rotation angle and the angle data of the risk point meet the collision risk range, a collision risk warning is triggered.

[0013] Preferably, mapping the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data includes:

[0014] Pre-calibrating the isocenter coordinates, the motion plane, and the safety range of the linear accelerator;

[0015] Projecting the corresponding nodes of the skeleton node information onto the motion plane.

[0016] Preferably, mapping the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data includes:

[0017] Calculating the distance between the node corresponding to the skeleton node information and the motion plane to obtain the node distance;

[0018] The coordinates of the projection points corresponding to the skeleton node information are calculated according to the node distance to obtain the projection coordinate data; the expression of the projection coordinate data is: ;in, is the projection coordinate data; The node corresponding to the skeleton node information; is the node distance; is the normal vector of the motion plane.

[0019] Preferably, calculating the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data includes:

[0020] Construct an angle calculation formula; the expression of the angle calculation formula is:

[0021] ;

[0022] in, is the angle data; 、 are the coordinate data of the projection points to be calculated; 、 are the abscissa and ordinate of the isocenter coordinates respectively;

[0023] The angle calculation formula is used to calculate the angle between the projection coordinate data and the preset zero scale reference line to obtain the angle data.

[0024] Preferably, calculating the straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determining the projection point whose distance data does not meet the safety distance threshold as a risk point includes:

[0025] Construct a risk judgment formula; the expression of the risk judgment formula is: ;in, is the safety distance threshold;

[0026] The projection coordinate data is screened using the risk judgment formula to obtain the risk points.

[0027] Preferably, when the rotation angle and the angle data of the risk point meet the collision risk range, triggering a collision risk warning includes:

[0028] Calculating the absolute value of the difference between the angle data and the rotation angle at the current stage to obtain the absolute value of the angle difference;

[0029] When the absolute value of the angle difference is less than the preset collision threshold, the angle data is judged using the collision risk range according to the rotation angle of the current stage and the rotation angle of the next stage to obtain a collision risk judgment result; the expression of the collision risk range is: ;in, 、 are the rotation angle of the current stage and the rotation angle of the next stage respectively;

[0030] When the collision risk judgment result is a collision potential state, a collision risk warning is triggered.

[0031] Preferably, a radiotherapy posture monitoring system based on three-dimensional human skeleton tracking comprises:

[0032] A node information extraction module is used to use a 3D vision sensor to collect images of the target area to obtain a three-dimensional depth image containing the target human body, and use an image recognition algorithm to detect the three-dimensional depth image to obtain skeleton node information;

[0033] A projection model is used to map the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data;

[0034] An angle calculation module, used to calculate the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data;

[0035] a distance judgment module, configured to calculate the straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determine a projection point whose distance data does not meet a safety distance threshold as a risk point;

[0036] An angle acquisition module, used to acquire the rotation angle of the linear accelerator treatment head at the current moment;

[0037] The risk judgment module is used to trigger a collision risk warning when the rotation angle and the angle data of the risk point meet the collision risk range.

[0038] Preferably, an electronic device comprises: at least one processor, and a memory communicatively connected to the processor; wherein the memory stores instructions that can be executed by the processor, and the instructions are executed by the processor so that the processor can execute the aforementioned radiotherapy posture monitoring method based on three-dimensional human skeleton tracking.

[0039] Preferably, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the aforementioned radiotherapy posture monitoring method based on three-dimensional human skeleton tracking.

[0040] The present invention discloses the following technical effects:

[0041] The present invention provides a radiotherapy posture monitoring method, system, electronic device and storage medium based on three-dimensional human skeleton tracking. By using markerless three-dimensional vision technology to analyze the spatial posture information of key skeletal nodes and combining the motion trajectory of the linear accelerator to perform collision risk assessment, the problems of skin damage and respiratory function restriction caused by rigid fixation are solved, and the prediction of dangerous angle intervals and a multi-level early warning mechanism are realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 A schematic diagram of a radiotherapy posture monitoring process based on three-dimensional human skeleton tracking provided by an embodiment of the present invention;

[0044] Figure 2 A schematic diagram of a radiotherapy system provided by an embodiment of the present invention;

[0045] Figure 3 A flowchart of the system operation provided by an embodiment of the present invention;

[0046] Figure 4 A system block diagram provided for an embodiment of the present invention;

[0047] Figure 5 A schematic diagram of modeling a projection plane coordinate system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0049] The purpose of the present invention is to provide a radiotherapy posture monitoring method, system, electronic device and storage medium based on three-dimensional human skeleton tracking to solve the problems of skin damage and respiratory function restriction caused by rigid fixation.

[0050] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0051] Figure 1 A schematic diagram of a radiotherapy posture monitoring process based on three-dimensional human skeleton tracking provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the present invention provides a radiotherapy posture monitoring method based on three-dimensional human skeleton tracking, comprising:

[0052] Using a 3D vision sensor to capture images of the target area to obtain a three-dimensional depth image containing the target human body, and using an image recognition algorithm to detect the three-dimensional depth image to obtain skeleton node information;

[0053] Mapping the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data;

[0054] Calculating the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data;

[0055] Calculating a straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determining a projection point whose distance data does not meet a safety distance threshold as a risk point;

[0056] Acquiring the rotation angle of the linear accelerator treatment head at the current moment;

[0057] When the rotation angle and the angle data of the risk point meet the collision risk range, a collision risk warning is triggered.

[0058] Furthermore, the skeleton node information is mapped to the motion plane of the linear accelerator according to the pre-calibrated space conversion parameters to obtain projection coordinate data, including:

[0059] Pre-calibrating the isocenter coordinates, the motion plane, and the safety range of the linear accelerator;

[0060] Projecting the corresponding nodes of the skeleton node information onto the motion plane.

[0061] Specifically, the skeleton node information is mapped to the motion plane of the linear accelerator according to the pre-calibrated space conversion parameters to obtain projection coordinate data, including:

[0062] Calculating the distance between the node corresponding to the skeleton node information and the motion plane to obtain the node distance;

[0063] The coordinates of the projection points corresponding to the skeleton node information are calculated according to the node distance to obtain the projection coordinate data; the expression of the projection coordinate data is: ;in, is the projection coordinate data; The node corresponding to the skeleton node information; is the node distance; is the normal vector of the motion plane.

[0064] Furthermore, the angle between the projection coordinate data and a preset zero scale reference line is calculated to obtain angle data, including:

[0065] Construct an angle calculation formula; the expression of the angle calculation formula is:

[0066] ;

[0067] in, is the angle data; 、 are the coordinate data of the projection points to be calculated; 、 are the abscissa and ordinate of the isocenter coordinates respectively;

[0068] The angle calculation formula is used to calculate the angle between the projection coordinate data and the preset zero scale reference line to obtain the angle data.

[0069] Specifically, calculating the straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determining the projection point whose distance data does not meet the safety distance threshold as a risk point includes:

[0070] Construct a risk judgment formula; the expression of the risk judgment formula is: ;in, is the safety distance threshold;

[0071] The projection coordinate data is screened using the risk judgment formula to obtain the risk points.

[0072] Furthermore, when the rotation angle and the angle data of the risk point meet the collision risk range, a collision risk warning is triggered, including:

[0073] Calculating the absolute value of the difference between the angle data and the rotation angle at the current stage to obtain the absolute value of the angle difference;

[0074] When the absolute value of the angle difference is less than the preset collision threshold, the angle data is judged using the collision risk range according to the rotation angle of the current stage and the rotation angle of the next stage to obtain a collision risk judgment result; the expression of the collision risk range is: ;in, 、 are the rotation angle of the current stage and the rotation angle of the next stage respectively;

[0075] When the collision risk judgment result is a collision potential state, a collision risk warning is triggered.

[0076] refer to Figures 2 to 4 A radiotherapy posture monitoring device based on 3D human skeleton tracking includes a 3D human skeleton extraction module, a projection calculation module, a collision risk assessment module, and a linear accelerator control module. The 3D depth camera captures the spatial coordinates of human joints in real time to construct a high-precision 3D human skeleton model. The projection calculation module uses this model to project key human body parts onto the motion space of the radiotherapy equipment, generating a dynamic safety boundary. The collision risk assessment module combines the gantry motion parameters with the safety boundary to assess collision risk in real time and output risk angle data. The gantry control module dynamically adjusts the linear accelerator's motion trajectory based on this data, achieving precise monitoring during radiotherapy and proactive human-machine collision avoidance.

[0077] Specifically, the image recognition algorithm for extracting human skeleton points adopts the method used by Microsoft Kinect cameras: first, foreground extraction is performed by acquiring a depth image, and the depth value is used to distinguish the foreground human body from the background. Then, a refinement algorithm is used to extract the human skeleton lines, making the extracted skeleton lines cleaner and clearer. To address the problem of human self-occlusion, the depth value distribution is used to distinguish between occluded parts and body parts, and the skeletons are extracted separately and then fused. Finally, referring to human body proportion data, the location of skeletal joints is determined using the maximum triangle algorithm, for example, using the center of gravity of the head as the head joint point, and the elbow joint point is determined based on the principle of maximum triangle area.

[0078] Furthermore, the human skeleton extraction module is used to receive and process three-dimensional depth images containing the human body in real time without adding additional labels. It only detects the human body area in the three-dimensional depth image through an image recognition algorithm, extracts the skeleton node information corresponding to the human body area, and transmits the skeleton node information to the projection calculation module.

[0079] Specifically, the projection calculation module is used to receive skeleton nodes from the human skeleton extraction module in real time. Based on pre-calibrated spatial transformation parameters, the skeleton node information is mapped to the motion plane of the linear accelerator, generating projection coordinate data, and transmitting the projection coordinate data to the collision risk assessment module. Specifically, it includes:

[0080] S1: Pre-calibrate the isocenter coordinates of the linear accelerator , rotational motion plane , safety range .

[0081] S2: Receive N skeleton node coordinates transmitted from the 3D human skeleton extraction module to form a point set P:

[0082]

[0083] S3: Combine the pre-calibrated data to get the skeleton node coordinates Projection on the plane of rotation The plane of rotation The normal vector is ,flat It can be defined as:

[0084]

[0085] S4: Calculate skeleton nodes To the rotation plane distance :

[0086]

[0087] S5: Calculate the coordinates of the projection point :

[0088]

[0089] S6: Output all projection point coordinates and pre-calibration data to the risk assessment module.

[0090] refer to Figure 5 The collision risk assessment module is used to receive the projection coordinate data from the projection calculation module in real time, model the motion range of the linear accelerator as a two-dimensional plane angle coordinate system, establish a polar coordinate system (range -180° to +180°) based on the isocenter point, and calculate the angle between each projection point and the preset zero-scale reference line of the human 3D skeleton node monitored in real time to obtain angle data, and calculate the straight-line distance from each projection point to the isocenter point; based on the pre-calibrated safety distance threshold, determine whether each projection point is in the safe area, and output the angle data of the projection point in the risk area; at the same time, receive the current angle parameters of the linear accelerator treatment head sent by the linear accelerator's rack control module, and generate and send a warning signal when it detects that there is a collision risk between the projection point and the motion trajectory of the linear accelerator to avoid human-machine collision caused by human movement when the linear accelerator is not rotating. The specific steps include:

[0091] S1: Receive the projection coordinate data of the projection calculation module in real time.

[0092] S2: Calculate the distance from the projection point to the isocenter point, based on the safety distance threshold Classify projected points into safe points and risky points:

[0093]

[0094] S3: Calculate the angle between each risk point and the zero-scale baseline:

[0095]

[0096] S4: Convert the included angle into an angle value in the linear accelerator coordinate system, and generate a 360-bit bit sequence accordingly. If the angle of the risk point is L degrees, set the Lth value in the bit sequence to 1.

[0097] S5: Receive the angle value of the treatment head transmitted by the linear accelerator rack control module in real time, calculate the difference between the angle and the angle of each risk point, and Triggering collision risk warning:

[0098]

[0099] Furthermore, the linear accelerator gantry control module is used to accurately control the rotational movement of the linear accelerator during radiotherapy, receive the core data of the radiotherapy plan (rotation path, etc.) transmitted by the TPS system (radiotherapy planning system) in real time, and parse the bit sequence containing the collision risk angle data transmitted by the collision risk assessment module in real time, and transmit the current angle parameters of the linear accelerator treatment head (such as gantry angle, collimator angle, etc.) to the collision risk assessment module in real time. When the next stage rotation angle in the radiotherapy plan data is With the current rotation angle , there is a projection point on the rotation path whose angle is , when:

[0100]

[0101] Issue a collision warning.

[0102] As an optional implementation, this embodiment further provides a radiotherapy posture monitoring system based on three-dimensional human skeleton tracking, comprising:

[0103] A node information extraction module is used to use a 3D vision sensor to collect images of the target area to obtain a three-dimensional depth image containing the target human body, and use an image recognition algorithm to detect the three-dimensional depth image to obtain skeleton node information;

[0104] A projection model is used to map the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data;

[0105] An angle calculation module, used to calculate the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data;

[0106] a distance judgment module, configured to calculate the straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determine a projection point whose distance data does not meet a safety distance threshold as a risk point;

[0107] An angle acquisition module, used to acquire the rotation angle of the linear accelerator treatment head at the current moment;

[0108] The risk judgment module is used to trigger a collision risk warning when the rotation angle and the angle data of the risk point meet the collision risk range.

[0109] As an optional implementation, this embodiment also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the processor; wherein the memory stores instructions that can be executed by the processor, and the instructions are executed by the processor so that the processor can execute the aforementioned radiotherapy posture monitoring method based on three-dimensional human skeleton tracking.

[0110] As an optional implementation, this embodiment further provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the aforementioned radiotherapy posture monitoring method based on three-dimensional human skeleton tracking.

[0111] The beneficial effects of the present invention are as follows:

[0112] This invention accurately predicts and avoids collision risks without physical contact, eliminating adverse reactions such as skin damage and respiratory impairment caused by rigid fixation while significantly improving patient comfort and compliance during treatment. Furthermore, real-time monitoring data can be fed back into the radiotherapy planning system to optimize dose distribution and irradiation angles, achieving both improved treatment accuracy and efficiency.

[0113] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0114] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A radiotherapy posture monitoring method based on three-dimensional human skeleton tracking, characterized in that: include: Using a 3D vision sensor to capture images of the target area to obtain a three-dimensional depth image containing the target human body, and using an image recognition algorithm to detect the three-dimensional depth image to obtain skeleton node information; Mapping the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data; Calculating the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data; Calculating a straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determining a projection point whose distance data does not meet a safety distance threshold as a risk point; Acquiring the rotation angle of the linear accelerator treatment head at the current moment; When the rotation angle and the angle data of the risk point meet the collision risk range, a collision risk warning is triggered.

2. The radiotherapy posture monitoring method based on three-dimensional human skeleton tracking according to claim 1, characterized in that: The skeleton node information is mapped to the motion plane of the linear accelerator according to the pre-calibrated space conversion parameters to obtain projection coordinate data, including: Pre-calibrating the isocenter coordinates, the motion plane, and the safety range of the linear accelerator; Projecting the corresponding nodes of the skeleton node information onto the motion plane.

3. The radiotherapy posture monitoring method based on three-dimensional human skeleton tracking according to claim 1, characterized in that: The skeleton node information is mapped to the motion plane of the linear accelerator according to the pre-calibrated space conversion parameters to obtain projection coordinate data, including: Calculating the distance between the node corresponding to the skeleton node information and the motion plane to obtain the node distance; The coordinates of the projection points corresponding to the skeleton node information are calculated according to the node distance to obtain the projection coordinate data; the expression of the projection coordinate data is: ;in, is the projection coordinate data; The node corresponding to the skeleton node information; is the node distance; is the normal vector of the motion plane.

4. The radiotherapy posture monitoring method based on three-dimensional human skeleton tracking according to claim 1, characterized in that: Calculating the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data includes: Construct an angle calculation formula; the expression of the angle calculation formula is: ； in, is the angle data; 、 are the coordinate data of the projection points to be calculated; 、 are the abscissa and ordinate of the isocenter coordinates respectively; The angle calculation formula is used to calculate the angle between the projection coordinate data and the preset zero scale reference line to obtain the angle data.

5. The radiotherapy posture monitoring method based on three-dimensional human skeleton tracking according to claim 1, characterized in that: Calculating the straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determining the projection point whose distance data does not meet the safety distance threshold as a risk point, including: Construct a risk judgment formula; the expression of the risk judgment formula is: ;in, is the safety distance threshold; The projection coordinate data is screened using the risk judgment formula to obtain the risk points.

6. The radiotherapy posture monitoring method based on three-dimensional human skeleton tracking according to claim 1, characterized in that: When the rotation angle and the angle data of the risk point meet the collision risk range, a collision risk warning is triggered, including: Calculating the absolute value of the difference between the angle data and the rotation angle at the current stage to obtain the absolute value of the angle difference; When the absolute value of the angle difference is less than the preset collision threshold, the angle data is judged using the collision risk range according to the rotation angle of the current stage and the rotation angle of the next stage to obtain a collision risk judgment result; the expression of the collision risk range is: ;in, 、 are the rotation angle of the current stage and the rotation angle of the next stage respectively; When the collision risk judgment result is a collision potential state, a collision risk warning is triggered.

7. A radiotherapy posture monitoring system based on three-dimensional human skeleton tracking, characterized in that: include: A node information extraction module is used to use a 3D vision sensor to collect images of the target area to obtain a three-dimensional depth image containing the target human body, and use an image recognition algorithm to detect the three-dimensional depth image to obtain skeleton node information; A projection model is used to map the skeleton node information to the motion plane of the linear accelerator according to pre-calibrated space conversion parameters to obtain projection coordinate data; An angle calculation module, used to calculate the angle between the projection coordinate data and a preset zero scale reference line to obtain angle data; a distance judgment module, configured to calculate the straight-line distance between the projection coordinate data and the isocenter of the motion plane to obtain distance data, and determine a projection point whose distance data does not meet a safety distance threshold as a risk point; An angle acquisition module, used to acquire the rotation angle of the linear accelerator treatment head at the current moment; The risk judgment module is used to trigger a collision risk warning when the rotation angle and the angle data of the risk point meet the collision risk range.

8. An electronic device, characterized in that: include: At least one processor and a memory communicatively connected to the processor; wherein the memory stores instructions that can be executed by the processor, and the instructions are executed by the processor so that the processor can execute a radiotherapy posture monitoring method based on three-dimensional human skeleton tracking as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to enable a computer to execute a radiotherapy posture monitoring method based on three-dimensional human skeleton tracking according to any one of claims 1 to 6.

Images