Shoulder joint puncture guiding system and method based on intelligent ultrasonic image

Abstract

The invention discloses a shoulder joint puncture guiding system and method based on an intelligent ultrasonic image, and the system comprises an ultrasonic image collection and preprocessing module, an anatomical structure recognition and positioning module, a puncture path planning and guiding module, a puncture operation real-time monitoring and feedback module, and a puncture result evaluation and recording module. Through intelligent ultrasonic image recognition and path planning, the technical problems that traditional shoulder joint puncture depends on doctor experience and is insufficient in precision are solved, through real-time monitoring and feedback, the technical problem that important anatomical structures are prone to being accidentally injured in the traditional puncture process is solved, the operation safety is improved, and through the intelligent puncture guiding system, the puncture guiding accuracy is improved. The technical problems that a traditional method lacks a standardized process and is poor in operation consistency are solved, and the technical problems that a traditional puncture method is long in time consumption and low in efficiency are solved through automatic anatomical structure recognition and path planning.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent ultrasonic analysis puncture guidance, and in particular to a shoulder joint puncture guidance system and method based on intelligent ultrasonic imaging. Background Art

[0002] At present, with the rapid development of medical technology, the demand for shoulder joint puncture in diagnosis and treatment is increasing. However, traditional puncture guidance technology usually relies on mechanical adjustment devices. When doctors need to achieve precise puncture in complex anatomical structures, these methods often cannot meet the needs. In other words, this puncture guidance method based on mechanical adjustment will inevitably lack intelligent support in the process of achieving precise puncture, resulting in insufficient puncture accuracy and subsequent compensatory processing of the puncture results.

[0003] After searching, Chinese patent publication number: CN119606499A discloses an interventional ultrasound puncture auxiliary guide. This technical solution is a puncture guidance method based on mechanical adjustment. Although it can improve the movement flexibility of the ultrasound probe, it cannot realize intelligent anatomical structure recognition, path planning and real-time feedback, nor can it guarantee the accuracy and safety during the puncture process. Therefore, it is necessary to design a shoulder joint puncture guidance system and method based on intelligent ultrasound imaging to improve the above problems. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides a shoulder joint puncture guidance system and method based on intelligent ultrasound imaging, aiming to solve the problems in the existing technology that intelligent anatomical structure recognition, path planning and real-time feedback cannot be achieved during joint puncture guidance, and the accuracy and safety of the puncture process cannot be guaranteed.

[0005] The present invention provides a shoulder joint puncture guidance system based on intelligent ultrasonic imaging, the system comprising:

[0006] Ultrasonic image acquisition and preprocessing module, which is responsible for using ultrasonic imaging equipment to collect ultrasonic image data of the shoulder joint area in real time and perform preprocessing operations such as denoising, image enhancement, and edge detection to improve image quality and provide a clear and accurate image foundation for subsequent processing;

[0007] The anatomical structure recognition and positioning module, based on a deep learning model, analyzes pre-processed ultrasound images to accurately identify the key anatomical structures of the shoulder joint. It then locates the puncture target area based on the recognition results and generates coordinate information for the target area.

[0008] The puncture path planning and guidance module uses a path planning algorithm to calculate the optimal puncture path based on the coordinates of the target area and the patient's position, ensuring that the path avoids important anatomical structures. The planned puncture path is then superimposed on the ultrasound image in real time, providing doctors with intuitive and clear puncture guidance through a visual interface.

[0009] The real-time monitoring and feedback module for puncture operations collects ultrasound image data in real time during the puncture process, dynamically monitors the position and trajectory of the puncture needle, compares the actual position of the puncture needle with the planned path through an intelligent algorithm, and provides real-time adjustment suggestions to the doctor to ensure the accuracy and safety of the puncture operation.

[0010] Preferably, the step of collecting real-time ultrasonic image data of the shoulder joint area by the ultrasonic imaging device includes:

[0011] A high-frequency linear array probe is used to scan the shoulder joint area. The transmission frequency and focal depth are adaptively adjusted according to the patient's body fat thickness for dynamic parameter configuration. When the fat layer thickness is ≤2cm, the high-frequency mode is used to improve the resolution of superficial tissues. When the fat layer thickness is greater than 2cm, the probe switches to the penetration enhancement mode and combines coded excitation technology to improve the signal-to-noise ratio of deep tissues.

[0012] Through the beamforming algorithm of the probe array, sagittal, coronal and transverse images are acquired simultaneously to generate three-dimensional volume data and obtain ultrasound image data of multi-plane synchronous imaging;

[0013] The built-in pressure sensor in the ultrasound imaging device monitors the contact pressure between the probe and the skin in real time. When the pressure exceeds the threshold, a warning is triggered to avoid image distortion caused by tissue deformation.

[0014] Preferably, the step of performing preprocessing operations on the ultrasound image data includes:

[0015] A K-SVD dictionary learning algorithm is used to construct a joint dictionary containing normal tissue and noise features. Sparse coding is used to separate noise components, and non-local mean constraints are introduced to preserve the continuity of fine structures such as tendons and nerves. Speckle noise is suppressed in ultrasound image data. The relative displacement of the probe and skin is tracked in real time using the optical flow method. A displacement field model is established, and a backprojection algorithm is used to dynamically correct the spatial alignment of the image sequence. Inertial measurement data is combined to compensate for rotation artifacts caused by probe jitter, motion artifacts are compensated for in the ultrasound image data, and a dynamic time gain compensation curve is designed. The gain parameters are adjusted in layers according to the echo signal strength, and gain equalization optimization is performed on the ultrasound image data.

[0016] Anisotropic diffusion filtering is used to enhance the edges along the direction of tendon fibers, suppress lateral diffusion, and enhance the high-frequency details of ultrasound image data. Low-frequency components are extracted through wavelet multi-resolution decomposition, which enhances the contours of large-scale anatomical structures such as joint capsules and bursae, and strengthens the low-frequency structure of ultrasound image data.

[0017] Preferably, the training method of the deep learning model is:

[0018] Collecting a shoulder joint ultrasound image dataset requires covering ultrasound images of different body positions, different pathological conditions, and different equipment parameters to enhance the generalization ability of the model. Radiologists perform pixel-level annotation of key anatomical structures, generate segmentation masks, and input them into the deep learning model for training.

[0019] Preferably, the steps of performing recognition analysis on the pre-processed ultrasound image to accurately identify key anatomical structures of the shoulder joint, then locating the puncture target area according to the recognition result, and generating coordinate information of the target area include:

[0020] The preprocessed ultrasound image data is input into the deep learning model, which outputs the probability map of each anatomical structure. Connected region analysis is applied to remove the noise area, and the largest connected region is retained as the final recognition result to identify the key anatomical structures of the shoulder joint.

[0021] Based on the identified key anatomical structures of the shoulder joint, its geometric center or specific point is calculated as the puncture target. A three-dimensional coordinate system is established with the ultrasound probe as the origin, and the target area is converted into spatial coordinates.

[0022] Preferably, the steps of calculating the optimal puncture path using a path planning algorithm based on the coordinate information of the target area and the patient's position to ensure that the path avoids important anatomical structures, and then superimposing the planned puncture path on the ultrasound image in real time to provide the doctor with intuitive and clear puncture guidance through a visual interface are as follows:

[0023] The spatial coordinates are input into the puncture device, and degree-of-freedom sensors are installed on the puncture needle and probe of the puncture device to track the spatial coordinates in real time. The three-dimensional contour of the patient's body surface is reconstructed by the binocular infrared camera on the puncture device and aligned with the ultrasound image coordinate system to reduce positioning errors and form a guided puncture path. The puncture path planning is performed using a three-level path planning mode, which includes an automatic mode, a semi-automatic mode, and an emergency mode. The automatic mode generates the shortest obstacle avoidance path based on the A* algorithm and is suitable for standard anatomical structures. The semi-automatic mode allows the doctor to manually set key path points, and the system performs local optimization. When a sudden displacement is detected in the emergency mode, the RRT* algorithm is enabled to replan the path within 50ms.

[0024] Preferably, the guidance interface of the system uses augmented reality (AR) overlay technology to fuse the planned path, risk heat map and real-time ultrasound image for display.

[0025] Preferably, the steps of real-time acquisition of ultrasound image data, dynamic monitoring of the position and trajectory of the puncture needle, comparison of the actual position of the puncture needle with the planned path through an intelligent algorithm, and providing real-time adjustment suggestions to the doctor to ensure the accuracy and safety of the puncture operation are as follows:

[0026] The ultrasound probe is used to obtain real-time three-dimensional image data of the puncture area. The optical positioning system and electromagnetic sensor are used to track the spatial coordinates of the puncture needle. A mapping relationship between the image coordinate system and the patient's anatomical structure is established to generate real-time trajectory data with millimeter-level accuracy. The real-time trajectory data is then compared and calculated with the guided puncture path.

[0027] Calculate the Euclidean distance and angular offset between the actual position of the puncture needle and the guided puncture path. If the offset exceeds the preset threshold, the system triggers an alarm and generates correction suggestions;

[0028] Dynamic correction arrows or warning messages are superimposed on the guidance interface of the system to prompt doctors to adjust the needle insertion angle or depth.

[0029] Preferably, the system also includes a puncture result evaluation and recording module. After the puncture is completed, ultrasound images are used to evaluate the puncture results to confirm whether the puncture needle has reached the target area. The image data, path planning information and operation records during the puncture process are saved in the database to provide a reference for subsequent diagnosis and treatment.

[0030] The present invention also provides a shoulder joint puncture guidance method based on intelligent ultrasound imaging. According to the shoulder joint puncture guidance system based on intelligent ultrasound imaging as described above, the method comprises the following steps:

[0031] S1. Acquisition and preprocessing of ultrasound images of shoulder joint;

[0032] S1.1. Use ultrasound imaging equipment to collect real-time ultrasound image data of the shoulder joint area;

[0033] S1.2. Preprocess the ultrasound image to improve image quality;

[0034] S2. Identification and positioning of shoulder joint anatomy;

[0035] S2.1. Analyze pre-processed ultrasound images using a deep learning model to identify key anatomical structures of the shoulder joint.

[0036] S2.2. Locate the puncture target area based on the recognition results and generate coordinate information of the target area;

[0037] S3, puncture path planning and guidance;

[0038] S3.1. Calculate the optimal puncture path using a path planning algorithm based on the target area's coordinates and the patient's position, avoiding critical anatomical structures.

[0039] S3.2. Superimpose the planned puncture path on the ultrasound image in real time, providing doctors with intuitive puncture guidance through a visual interface;

[0040] S4, real-time monitoring and feedback of puncture operation;

[0041] S4.1. During the puncture process, real-time ultrasound imaging data is collected to dynamically monitor the position and trajectory of the puncture needle;

[0042] S4.2. An intelligent algorithm compares the actual position of the puncture needle with the planned path, providing real-time feedback and adjustment suggestions to ensure the accuracy and safety of the puncture.

[0043] S5. Evaluation and recording of puncture results;

[0044] S5.1. After the puncture is completed, use ultrasound imaging to evaluate the puncture results and confirm whether the puncture needle has reached the target area;

[0045] S5.2. Save the imaging data, path planning information, and operation records during the puncture process to the database to provide a reference for subsequent diagnosis and treatment.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] The present invention solves the technical problem of traditional shoulder joint puncture relying on the doctor's experience and insufficient accuracy through intelligent ultrasonic image recognition and path planning. It solves the technical problem of easy accidental injury to important anatomical structures during traditional puncture through real-time monitoring and feedback, and improves the safety of operation. It solves the technical problem of the lack of standardized process and poor operation consistency of traditional methods through the intelligent puncture guidance system. It solves the technical problem of long time consumption and low efficiency of traditional puncture methods through automated anatomical structure recognition and path planning. DETAILED DESCRIPTION

[0048] To make the objectives, technical solutions, and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other implementations obtained by those of ordinary skill in the art without inventive effort are within the scope of protection of the present invention.

[0049] Example: This example provides a shoulder joint puncture guidance system based on intelligent ultrasound imaging, which includes an ultrasound image acquisition and preprocessing module, an anatomical structure recognition and positioning module, a puncture path planning and guidance module, a real-time monitoring and feedback module for puncture operations, and a puncture result evaluation and recording module.

[0050] In this embodiment, the ultrasound image acquisition and preprocessing module is responsible for using ultrasound imaging equipment to acquire ultrasound image data of the shoulder joint area in real time and performing preprocessing operations such as denoising, image enhancement, and edge detection to improve image quality and provide a clear and accurate image foundation for subsequent processing. The specific steps are as follows:

[0051] A high-frequency linear array probe is used to scan the shoulder joint area, and the transmission frequency and focal depth are adaptively adjusted according to the patient's body fat thickness for dynamic parameter configuration. When the fat layer thickness is ≤2cm, the high-frequency mode is used to improve the resolution of superficial tissue. When the fat layer thickness is greater than 2cm, it switches to the penetration enhancement mode and combines the coded excitation technology to improve the signal-to-noise ratio of deep tissue. Through the beam synthesis algorithm of the probe array, sagittal, coronal and cross-sectional images are obtained simultaneously to generate three-dimensional volume data and obtain multi-plane synchronous imaging ultrasound image data. A built-in pressure sensor in the ultrasound imaging device monitors the contact pressure between the probe and the skin in real time. When the pressure exceeds the threshold, a warning is triggered to avoid image distortion caused by tissue deformation.

[0052] The ultrasound image data were preprocessed using the K-SVD dictionary learning algorithm to construct a joint dictionary containing normal tissue and noise features. The noise components were separated by sparse coding, and non-local mean constraints were introduced to retain the continuity of fine structures such as tendons and nerves. The ultrasound image data were subjected to speckle noise suppression. After speckle noise suppression, the contrast-to-noise ratio of the ultrasound image data was improved by ≥15dB. The relative displacement of the probe and the skin was tracked in real time based on the optical flow method, and a displacement field model was established. The spatial alignment of the image sequence was dynamically corrected by the back-projection algorithm. Combined with the inertial measurement single data, the image sequence was corrected. To compensate for rotation artifacts caused by probe jitter, the ultrasound image data is compensated for motion artifacts to ensure that the image registration error is ≤0.2mm. A dynamic time gain compensation curve is designed, and the gain parameters are adjusted in layers according to the echo signal intensity. The ultrasound image data is optimized for gain equalization. Anisotropic diffusion filtering is used to enhance the edges along the tendon fiber direction and suppress lateral diffusion, thereby improving the sharpness of the rotator cuff edge by 50%. The high-frequency details of the ultrasound image data are enhanced, and the low-frequency components are extracted through wavelet multi-resolution decomposition. The contours of large-scale anatomical structures such as the joint capsule and bursa are enhanced, and the low-frequency structure of the ultrasound image data is strengthened.

[0053] In this embodiment, the anatomical structure recognition and positioning module is based on a deep learning model. This module recognizes and analyzes the pre-processed ultrasound images to accurately identify the key anatomical structures of the shoulder joint. Subsequently, based on the recognition results, it locates the puncture target area and generates the coordinate information of the target area.

[0054] Among them, the training method of the deep learning model is: collect shoulder joint ultrasound image datasets, which must cover ultrasound images of different body postures, different pathological conditions and different equipment parameters to enhance the model's generalization ability. Radiologists perform pixel-level annotation of key anatomical structures, generate segmentation masks, and input them into the deep learning model for training.

[0055] The specific operation steps are as follows: collect a shoulder joint ultrasound image dataset, which must cover ultrasound images of different body positions, different pathological conditions, and different equipment parameters to enhance the generalization ability of the model; have radiologists perform pixel-level annotation of key anatomical structures, generate segmentation masks, and input them into the deep learning model for training; input the preprocessed ultrasound image data into the deep learning model, output the probability map of each anatomical structure, apply connected region analysis to remove the noise area, retain the largest connected domain as the final recognition result, and identify the key anatomical structure of the shoulder joint; based on the identified key anatomical structure of the shoulder joint, calculate its geometric center or specific point as the puncture target, establish a three-dimensional coordinate system with the ultrasound probe as the origin, and convert the target area into spatial coordinates (such as (x, y, z)).

[0056] In this embodiment, the puncture path planning and guidance module uses a path planning algorithm to calculate the optimal puncture path based on the coordinate information of the target area and the patient's position, ensuring that the path avoids important anatomical structures. The planned puncture path is then superimposed on the ultrasound image in real time, providing the doctor with intuitive and clear puncture guidance through a visual interface.

[0057] The specific operating steps are: input the spatial coordinates into the puncture device, install degree of freedom sensors on the puncture needle and probe of the puncture device, track the spatial coordinates in real time, reconstruct the three-dimensional contour of the patient's body surface through the binocular infrared camera on the puncture device, align it with the ultrasound image coordinate system, reduce positioning errors, form a guided puncture path, and use a three-level path planning mode for puncture path planning. The three-level path planning mode includes automatic mode, semi-automatic mode and emergency mode. The automatic mode generates the shortest obstacle avoidance path based on the A* algorithm and is suitable for standard anatomical structures. The semi-automatic mode allows the doctor to manually set key path points, and the system performs local optimization. When the emergency mode detects sudden displacement, the RRT* algorithm is enabled to re-plan the path within 50ms. The guidance interface of the system uses augmented reality (AR) overlay technology to fuse the planned path, risk heat map and real-time ultrasound image for display.

[0058] In this embodiment, the puncture operation real-time monitoring and feedback module collects ultrasound image data in real time during the puncture process, dynamically monitors the position and trajectory of the puncture needle, compares the actual position of the puncture needle with the planned path through an intelligent algorithm, and provides real-time adjustment suggestions to the doctor to ensure the accuracy and safety of the puncture operation;

[0059] The specific operation steps are as follows: using an ultrasound probe to obtain real-time three-dimensional image data of the puncture area, combining an optical positioning system and an electromagnetic sensor to track the spatial coordinates of the puncture needle, establishing a mapping relationship between the image coordinate system and the patient's anatomical structure, generating real-time trajectory data with millimeter-level accuracy, and comparing and calculating the real-time trajectory data with the guided puncture path;

[0060] Calculate the Euclidean distance and angular offset between the actual position of the puncture needle and the guided puncture path. If the offset exceeds the preset threshold, the system triggers an alarm and generates correction suggestions;

[0061] Dynamic correction arrows or warning messages are superimposed on the guidance interface of the system to prompt doctors to adjust the needle insertion angle or depth.

[0062] In this embodiment, the system also includes a puncture result evaluation and recording module. After the puncture is completed, ultrasound images are used to evaluate the puncture results to confirm whether the puncture needle has reached the target area. The image data, path planning information and operation records during the puncture process are saved in the database to provide a reference for subsequent diagnosis and treatment.

[0063] This embodiment provides a shoulder joint puncture guidance method based on intelligent ultrasound imaging, including the shoulder joint puncture guidance system based on intelligent ultrasound imaging as described in the above embodiment. The method includes the following steps:

[0064] S1. Acquisition and preprocessing of ultrasound images of shoulder joint;

[0065] S1.1. Use ultrasound imaging equipment to collect real-time ultrasound image data of the shoulder joint area;

[0066] S1.2. Preprocess the ultrasound image to improve image quality;

[0067] S2. Identification and positioning of shoulder joint anatomy;

[0068] S2.1. Analyze pre-processed ultrasound images using a deep learning model to identify key anatomical structures of the shoulder joint.

[0069] S2.2. Locate the puncture target area based on the recognition results and generate coordinate information of the target area;

[0070] S3, puncture path planning and guidance;

[0071] S3.1. Calculate the optimal puncture path using a path planning algorithm based on the target area's coordinates and the patient's position, avoiding critical anatomical structures.

[0072] S3.2. Superimpose the planned puncture path on the ultrasound image in real time, providing doctors with intuitive puncture guidance through a visual interface;

[0073] S4, real-time monitoring and feedback of puncture operation;

[0074] S4.1. During the puncture process, real-time ultrasound imaging data is collected to dynamically monitor the position and trajectory of the puncture needle;

[0075] S4.2. An intelligent algorithm compares the actual position of the puncture needle with the planned path, providing real-time feedback and adjustment suggestions to ensure the accuracy and safety of the puncture.

[0076] S5. Evaluation and recording of puncture results;

[0077] S5.1. After the puncture is completed, use ultrasound imaging to evaluate the puncture results and confirm whether the puncture needle has reached the target area;

[0078] S5.2. Save the imaging data, path planning information, and operation records during the puncture process to the database to provide a reference for subsequent diagnosis and treatment.

[0079] The present invention solves the technical problem of traditional shoulder joint puncture relying on the doctor's experience and insufficient accuracy through intelligent ultrasonic image recognition and path planning. It solves the technical problem of easy accidental injury to important anatomical structures during traditional puncture through real-time monitoring and feedback, and improves the safety of operation. It solves the technical problem of the lack of standardized process and poor operation consistency of traditional methods through the intelligent puncture guidance system. It solves the technical problem of long time consumption and low efficiency of traditional puncture methods through automated anatomical structure recognition and path planning.

[0080] All technical features in this embodiment can be freely combined according to actual needs.

[0081] The above embodiments are preferred implementation schemes of the present invention. In addition, the present invention can also be implemented in other ways. Any obvious replacement without departing from the concept of the present technical solution is within the scope of protection of the present invention.

Claims

1. A shoulder joint puncture guidance system based on intelligent ultrasound imaging, characterized in that: The system includes: Ultrasonic image acquisition and preprocessing module, which is responsible for using ultrasonic imaging equipment to collect ultrasonic image data of the shoulder joint area in real time and perform preprocessing operations such as denoising, image enhancement, and edge detection to improve image quality and provide a clear and accurate image foundation for subsequent processing; The anatomical structure recognition and positioning module, based on a deep learning model, analyzes pre-processed ultrasound images to accurately identify the key anatomical structures of the shoulder joint. It then locates the puncture target area based on the recognition results and generates coordinate information for the target area. The puncture path planning and guidance module uses a path planning algorithm to calculate the optimal puncture path based on the coordinates of the target area and the patient's position, ensuring that the path avoids important anatomical structures. The planned puncture path is then superimposed on the ultrasound image in real time, providing doctors with intuitive and clear puncture guidance through a visual interface. The real-time monitoring and feedback module for puncture operations collects ultrasound image data in real time during the puncture process, dynamically monitors the position and trajectory of the puncture needle, compares the actual position of the puncture needle with the planned path through an intelligent algorithm, and provides real-time adjustment suggestions to the doctor to ensure the accuracy and safety of the puncture operation.

2. The intelligent ultrasound imaging-based shoulder joint puncture guidance system according to claim 1, wherein the step of the ultrasound imaging device collecting real-time ultrasound image data of the shoulder joint area comprises: A high-frequency linear array probe is used to scan the shoulder joint area. The transmission frequency and focal depth are adaptively adjusted according to the patient's body fat thickness for dynamic parameter configuration. When the fat layer thickness is ≤2cm, the high-frequency mode is used to improve the resolution of superficial tissues. When the fat layer thickness is greater than 2cm, the probe switches to the penetration enhancement mode and combines coded excitation technology to improve the signal-to-noise ratio of deep tissues. Through the beamforming algorithm of the probe array, sagittal, coronal and transverse images are acquired simultaneously to generate three-dimensional volume data and obtain ultrasound image data of multi-plane synchronous imaging; The built-in pressure sensor in the ultrasound imaging device monitors the contact pressure between the probe and the skin in real time. When the pressure exceeds the threshold, a warning is triggered to avoid image distortion caused by tissue deformation.

3. The intelligent ultrasound imaging-based shoulder joint puncture guidance system according to claim 2, wherein the step of preprocessing the ultrasound image data comprises: A K-SVD dictionary learning algorithm is used to construct a joint dictionary containing normal tissue and noise features. Sparse coding is used to separate noise components, and non-local mean constraints are introduced to preserve the continuity of fine structures such as tendons and nerves. Speckle noise is suppressed in ultrasound image data. The relative displacement of the probe and skin is tracked in real time using the optical flow method. A displacement field model is established, and a backprojection algorithm is used to dynamically correct the spatial alignment of the image sequence. Inertial measurement data is combined to compensate for rotation artifacts caused by probe jitter, motion artifacts are compensated for in the ultrasound image data, and a dynamic time gain compensation curve is designed. The gain parameters are adjusted in layers according to the echo signal strength, and gain equalization optimization is performed on the ultrasound image data. Anisotropic diffusion filtering is used to enhance the edges along the direction of tendon fibers, suppress lateral diffusion, and enhance the high-frequency details of ultrasound image data. Low-frequency components are extracted through wavelet multi-resolution decomposition, which enhances the contours of large-scale anatomical structures such as joint capsules and bursae, and strengthens the low-frequency structure of ultrasound image data.

4. The intelligent ultrasound imaging-based shoulder joint puncture guidance system according to claim 3, wherein the training method of the deep learning model is: Collecting a shoulder joint ultrasound image dataset requires covering ultrasound images of different body positions, different pathological conditions, and different equipment parameters to enhance the generalization ability of the model. Radiologists perform pixel-level annotation of key anatomical structures, generate segmentation masks, and input them into the deep learning model for training.

5. The intelligent ultrasound imaging-based shoulder joint puncture guidance system according to claim 4, wherein the steps of performing recognition and analysis on the preprocessed ultrasound image to accurately identify key anatomical structures of the shoulder joint, then locating the puncture target area based on the recognition results, and generating coordinate information of the target area include: The preprocessed ultrasound image data is input into the deep learning model, which outputs the probability map of each anatomical structure. Connected region analysis is applied to remove the noise area, and the largest connected region is retained as the final recognition result to identify the key anatomical structures of the shoulder joint. Based on the identified key anatomical structures of the shoulder joint, its geometric center or specific point is calculated as the puncture target. A three-dimensional coordinate system is established with the ultrasound probe as the origin, and the target area is converted into spatial coordinates.

6. The intelligent ultrasound imaging-based shoulder joint puncture guidance system according to claim 5, characterized in that the optimal puncture path is calculated using a path planning algorithm based on the coordinate information of the target area and the patient's position, ensuring that the path avoids important anatomical structures. Subsequently, the planned puncture path is superimposed on the ultrasound image in real time, providing the physician with intuitive and clear puncture guidance through a visual interface, the steps of: The spatial coordinates are input into the puncture device, and degree-of-freedom sensors are installed on the puncture needle and probe of the puncture device to track the spatial coordinates in real time. The three-dimensional contour of the patient's body surface is reconstructed by the binocular infrared camera on the puncture device and aligned with the ultrasound image coordinate system to reduce positioning errors and form a guided puncture path. The puncture path planning is performed using a three-level path planning mode, which includes an automatic mode, a semi-automatic mode, and an emergency mode. The automatic mode generates the shortest obstacle avoidance path based on the A* algorithm and is suitable for standard anatomical structures. The semi-automatic mode allows the doctor to manually set key path points, and the system performs local optimization. When a sudden displacement is detected in the emergency mode, the RRT* algorithm is enabled to replan the path within 50ms.

7. A shoulder joint puncture guidance system based on intelligent ultrasound imaging according to claim 6, characterized in that the guidance interface of the system uses augmented reality (AR) overlay technology to fuse the planned path, risk heat map and real-time ultrasound image for display.

8. The intelligent ultrasound imaging-based shoulder joint puncture guidance system according to claim 7, characterized in that the steps of real-time acquisition of ultrasound image data, dynamic monitoring of the position and trajectory of the puncture needle, comparison of the actual position of the puncture needle with the planned path through an intelligent algorithm, and real-time provision of adjustment suggestions to the physician to ensure the accuracy and safety of the puncture operation are as follows: The ultrasound probe is used to obtain real-time three-dimensional image data of the puncture area. The optical positioning system and electromagnetic sensor are used to track the spatial coordinates of the puncture needle. A mapping relationship between the image coordinate system and the patient's anatomical structure is established to generate real-time trajectory data with millimeter-level accuracy. The real-time trajectory data is then compared and calculated with the guided puncture path. Calculate the Euclidean distance and angular offset between the actual position of the puncture needle and the guided puncture path. If the offset exceeds the preset threshold, the system triggers an alarm and generates correction suggestions; Dynamic correction arrows or warning messages are superimposed on the guidance interface of the system to prompt doctors to adjust the needle insertion angle or depth.

9. A shoulder joint puncture guidance system based on intelligent ultrasound imaging according to claim 8, characterized in that the system also includes a puncture result evaluation and recording module. After the puncture is completed, the puncture result is evaluated using ultrasound imaging to confirm whether the puncture needle has reached the target area. The image data, path planning information and operation records during the puncture process are saved in a database to provide a reference for subsequent diagnosis and treatment.

10. A shoulder joint puncture guidance method based on intelligent ultrasound imaging, according to the shoulder joint puncture guidance system based on intelligent ultrasound imaging according to claims 1-9, characterized in that: The method comprises the following steps: S1. Acquisition and preprocessing of ultrasound images of shoulder joint; S1.

1. Use ultrasound imaging equipment to collect real-time ultrasound image data of the shoulder joint area; S1.

2. Preprocess the ultrasound image to improve image quality; S2. Identification and positioning of shoulder joint anatomy; S2.

1. Analyze pre-processed ultrasound images using a deep learning model to identify key anatomical structures of the shoulder joint. S2.

2. Locate the puncture target area based on the recognition results and generate coordinate information of the target area; S3, puncture path planning and guidance; S3.

1. Calculate the optimal puncture path using a path planning algorithm based on the target area's coordinates and the patient's position, avoiding critical anatomical structures. S3.

2. Superimpose the planned puncture path on the ultrasound image in real time, providing doctors with intuitive puncture guidance through a visual interface; S4, real-time monitoring and feedback of puncture operation; S4.

1. During the puncture process, real-time ultrasound imaging data is collected to dynamically monitor the position and trajectory of the puncture needle; S4.

2. An intelligent algorithm compares the actual position of the puncture needle with the planned path, providing real-time feedback and adjustment suggestions to ensure the accuracy and safety of the puncture. S5. Evaluation and recording of puncture results; S5.

1. After the puncture is completed, use ultrasound imaging to evaluate the puncture results and confirm whether the puncture needle has reached the target area; S5.

2. Save the imaging data, path planning information, and operation records during the puncture process to the database to provide a reference for subsequent diagnosis and treatment.