Acupuncture point positioning method and system based on human body key points

Abstract

The invention relates to the technical field of acupoint recognition, in particular to an acupoint positioning method based on key points of a human body, which comprises the following steps: S1, capturing infrared radiation of a testee through infrared thermal imaging equipment, generating thermal image data, and transmitting the data to an upper computer system in a wired or wireless manner; s2, the upper computer system receives and preprocesses the generated thermal image data, including denoising, non-uniformity correction and image quality enhancement; s3, performing human body key point detection on the preprocessed thermal image data by using a YOLO v11 algorithm; s4, according to the two-dimensional localizer and the relative pitch between each acupuncture point and the key point of the human body, determining the accurate acupuncture point of the testee in combination with a triangulation positioning method; s5, storing the positioned acupuncture points as coordinates in the form of a dictionary, superposing the coordinates on the source image, and then uploading the coordinates to a cloud and a client; according to the invention, the accuracy of acupoint positioning is improved, and the probability of acupoint positioning errors caused by human body posture problems is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of acupoint recognition, and particularly to an acupoint positioning method and system based on human key points. Background Art

[0002] With the development of the times, the inheritance of traditional Chinese medicine faces challenges, and it takes a lot of time and effort to systematically learn Chinese medicine knowledge. To lower the medical threshold and promote the application of Chinese medicine, we can use artificial intelligence technology to assist acupoint positioning, simplify the learning process, improve the treatment accuracy, and enable more people to benefit from the wisdom and efficacy of Chinese medicine. In an artificial intelligence-assisted disease prediction system, it is necessary to add an acupoint superposition module to increase the prediction accuracy rate, which involves positioning acupoints. Existing acupoint positioning generally relies on the rich experience of Chinese medicine practitioners, which requires a large amount of labor costs and there is also a phenomenon of positioning errors due to subjective differences.

[0003] With the continuous progress of intelligent medical technology, algorithms for human acupoint recognition have emerged continuously. Currently, acupoint positioning algorithms mainly rely on optical positioning and vision technology, but due to differences in human body postures, sizes, etc., the positioning accuracy of existing algorithms is low. Therefore, how to solve the problem of inaccurate acupoint positioning caused by individual body postures has become an urgent issue to be solved in intelligent human acupoint recognition. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide an acupoint positioning method based on human key points. The present invention improves the accuracy of acupoint positioning and reduces the probability of acupoint positioning errors caused by human body posture problems.

[0005] The purpose of the present invention is achieved by the following measures: An acupoint positioning method based on human key points, comprising the following steps:

[0006] S1, requiring the person to be measured to stand still at a specified position and maintain the posture, capturing the infrared radiation of the person to be measured through an infrared thermal imaging device and generating thermal image data, and the data is transmitted to the host computer system by wired or wireless means;

[0007] S2, the host computer system receives and preprocesses the generated thermal image data, including denoising, non-uniformity correction, and image quality enhancement;

[0008] S3, using the YOLO v11 algorithm to detect human key points in the preprocessed thermal image data, forming a human skeleton with the detected key points, identifying the key parts on the surface of the person to be measured, and representing them with two-dimensional calibration;

[0009] S4. Based on the two-dimensional positioning marks and the relative cun distances between each acupoint and the key points of the human body, combined with the triangulation method, determine the accurate acupoint points of the measured person, reducing the positioning errors caused by body posture differences;

[0010] S5. Store the located acupoint points as coordinates in dictionary form, superimpose them on the source image, and then upload them to the cloud and the client.

[0011] Preferably, the host computer system includes a data receiving module, a preprocessing module for preprocessing and parsing the received data, a security control module for encrypting the data transmission and storage process, and an interactive interface for data display, device operation control, and system setting adjustment.

[0012] Preferably, the step S2 further includes performing data preprocessing using the contrast-limited adaptive histogram equalization algorithm.

[0013] Preferably, the preprocessing using the adaptive histogram equalization algorithm includes:

[0014] Image reading, obtaining the original infrared image from the infrared thermal imaging device;

[0015] Dividing the image area, dividing the entire image into multiple small blocks, and independently applying histogram equalization to each small block;

[0016] Histogram calculation and equalization, for each small block, calculate the pixel value distribution, and according to the calculated histogram, redistribute the pixel values within each small block to expand its dynamic range and enhance the local contrast, while applying a limiting condition to prevent excessive amplification of noise;

[0017] Limiting the contrast, setting a contrast upper limit in the adaptive histogram equalization algorithm. If the count in a certain bin exceeds this limit, the excess part will be evenly distributed to other bins;

[0018] Result fusion, recombining all the equalized small blocks into a complete image, and ensuring the continuity and consistency between the small blocks through appropriate algorithms to avoid artificial traces;

[0019] Preferably, in the step S3, the YOLO v11 improves the Backbone and Neck architectures, and adds C3k2 and C2PSA components.

[0020] Preferably, in step S3, the preprocessed thermal image data is input into the trained YOLO v11 model. The model will automatically perform feature extraction and object detection on the thermal image data, and output the position information of the bounding box and key points of the human body in the thermal image data. According to the output results, 17 key points of the human body can be located and recognized in the thermal image: 0: nose, 1: left eye, 2: right eye, 3: left ear, 4: right ear, 5: left shoulder, 6: right shoulder, 7: left elbow, 8: right elbow, 9: left wrist, 10: right wrist, 11: left hip, 12: right hip, 13: left knee, 14: right knee, 15: left ankle, 16: right ankle.

[0021] Preferably, the two-dimensional positioning markers in step S3:

[0022] A two-dimensional rectangular coordinate system is established with a certain fixed reference point of the human body. For each detected key point of the human body, the coordinate values (x_key, y_key) in the above coordinate system are recorded. The coordinate values are obtained by the YOLO v11 model outputting the key point position information and combining the definition of the coordinate system for conversion;

[0023] According to the coordinate values (x_key, y_key) of each key point and the relative position vector (Δx, Δy), the positions of each acupoint (x_acupoint, y_acupoint) are determined:

[0024] x_acupoint = Δx + x_key,

[0025] y_acupoint = Δy + y_key.

[0026] Preferably, the steps of the triangulation method in step S4 are as follows: Select easily recognizable key points of the human body as the reference, construct a triangle or polygon, and obtain the exact position of the target acupoint by measuring the lengths of each side or angles.

[0027] After receiving the thermal image data, the host computer system preprocesses the data in combination with CLAHE (Contrast Limited Adaptive Histogram Equalization). CLAHE divides the image into multiple small blocks / regions (tiles) and applies histogram equalization independently to each region. The purpose of this is to locally adjust the brightness and contrast rather than globally change the characteristics of the entire image. To avoid obvious artifacts at the tile boundaries, a certain overlapping area can be set between adjacent tiles, and a smooth transition can be made when merging the results finally. The specific steps include: First, read and convert the infrared image into a suitable format for processing; further, divide the image into overlapping small blocks, calculate the pixel value distribution within each small block and perform equalization processing, while applying a contrast limit to prevent noise amplification; furthermore, splice the processed small blocks back together and eliminate boundary artifacts through smooth transition; finally, perform additional post-processing optimization as needed. This method can significantly improve the visibility of local details while maintaining the overall structure of the image, and is particularly suitable for preprocessing infrared images under uneven lighting or complex backgrounds.

[0028] CLAHE can significantly improve the contrast of local regions while maintaining the overall image structure, making details that were originally difficult to distinguish more clearly visible. This is particularly important for infrared images because they often exhibit uneven brightness distributions due to temperature differences;

[0029] In the contrast limit step, by clipping the histogram, the abnormal highlights or dark spots caused by noise are effectively suppressed, thereby improving the quality and readability of the image;

[0030] CLAHE can automatically adjust the equalization degree according to the specific characteristics of different regions, and is suitable for processing infrared images with complex backgrounds or multiple lighting conditions.

[0031] Compared with the previous YOLO model, YOLO v11 has improved the Backbone and Neck architectures and added C3k2 and C2PSA components. C3K2 is optimized in the shallow layer of the network and can extract features more effectively; C2PSA better captures spatial context information by embedding a multi-head attention mechanism inside the C2 mechanism, making feature extraction more efficient and of higher quality, thereby improving the accuracy and robustness of object detection. Different from traditional CNNs, YOLOv11 uses a Transformer-based backbone network that can capture long-range dependencies and performs better in detecting small objects and objects in complex scenes, can better understand the global context information of the image, and makes the detection results more accurate.

[0032] Input the preprocessed thermal image data into the trained YOLO v11 model. The model will automatically extract features and detect objects from the thermal image data, and then output the position information of the bounding box and key points of the human body in the thermal image data. According to the output results, 17 key points of the human body can be located and recognized in the thermal image. Post-process the position information of the key points output by the model, such as filtering key points with low confidence and smoothing the key points.

[0033] Use the methods of triangulation and relative measurement, introduce two-dimensional positioning marks, construct triangles or other polygon structures based on the detected key points of the human body, and use the geometric characteristics of these structures to calculate the specific coordinates of the target acupoints. The application of the same body inch includes:

[0034] The middle finger same body inch is defined as the distance between the two end points of the medial side when the middle phalanx of the patient's middle finger is flexed, which is 1 inch. This method is often used for the location of acupoints on the limbs. After determining the key points such as the wrist and elbow, the middle finger same body inch can be used to locate acupoints on the arm.

[0035] The thumb same body inch is defined as the width of the patient's thumb joint, which is 1 inch. It is suitable for locating acupoints in narrow areas (such as between fingers and toes).

[0036] The horizontal finger same body inch (one husband method): It is to let the patient close the index finger, middle finger, ring finger and little finger together. Taking the transverse striation of the middle phalanx of the middle finger as the standard, the width of the four fingers is 3 inches. This method is often used for locating acupoints in relatively broad areas such as the abdomen and back.

[0037] This involves the methods of determining acupoint positions based on key points, including:

[0038] The ratio method uses some fixed proportional relationships in human anatomy to locate acupoints. Taking the arm as an example, after knowing the positions of key points such as the shoulder, elbow, and wrist, acupoints can be located according to the proportional relationships such as the length of the forearm and the upper arm. For example, if a certain acupoint is located at the midpoint of the forearm, then after determining the positions of both ends of the forearm (elbow and wrist) through key points, the approximate position of this acupoint can be calculated.

[0039] The angle method locates acupoints according to the angles formed between human bones. Taking the knee joint as an example, after determining the bone key points of the thigh and calf (such as the hip joint, knee joint, and ankle joint), the angle relationship between the thigh and calf is used to locate acupoints. For example, if a certain acupoint is located on the outside of the knee joint, in a direction forming a certain angle with the calf and thigh, the specific position of the acupoint can be determined by calculating this angle and the known limb length.

[0040] The distance addition method starts from the known key points and locates acupoints according to a fixed distance. This distance can be an absolute distance or a relative distance.

[0041] An acupoint positioning system based on human key points, comprising:

[0042] A data acquisition module for acquiring the infrared thermal image data of the person to be measured;

[0043] A data preprocessing module for transmitting the acquired data to the upper computer system and simultaneously using the CLAHE algorithm for denoising, non-uniformity correction, and image quality enhancement;

[0044] A human key point detection module, where the YOLO v11 model automatically performs feature extraction and target detection on the preprocessed thermal image data, and locates and identifies the key points of the human body in the thermal image.

[0045] An acupoint determination module that uses a method combining triangular positioning and relative cun distance, introduces a two-dimensional positioning mark, and obtains the final acupoint positioning points.

[0046] An acupoint data visualization and transmission module that stores the located acupoint points and corresponding coordinates in a dictionary, superimposes them on the source image, and then transmits the superimposed result to the cloud and the client.

[0047] Advantages of the present invention:

[0048] 1. The invention combines with the YOLOvll modern computer vision technology, providing a modern means for traditional Chinese medicine acupoint positioning, and contributing to the promotion of the modernization development of traditional Chinese medicine.

[0049] 2. The invention uses a method combining triangular positioning and relative cun distance, which can more accurately describe the spatial position relationship between key points and acupoints, and is more accurate than traditional qualitative descriptions. This quantitative method helps to reduce subjective judgment errors, especially in complex human postures and anatomical structures, and can better determine the acupoint positions.

[0050] 3. The use of the same body cun in the invention takes into account individual differences. Since everyone's body size is different, using relative cun distance can locate acupoints according to the patient's own body size, making the positioning more in line with the actual situation of the individual, improving the accuracy of acupoint positioning, and reducing the probability of acupoint positioning errors caused by human body posture problems. Description of the Drawings

[0051] Figure 1 Is a flowchart of the acupoint positioning method based on human key points;

[0052] Figure 2 Is a schematic diagram of the acupoint positioning system based on human key points. Detailed Embodiments

[0053] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0054] Embodiment 1: As Figure 1 described, a method for acupoint location based on human key points, and the method of the present invention is implemented based on the following system devices:

[0055] An infrared thermal imaging device for capturing the infrared radiation of the person to be measured and generating corresponding thermal image data.

[0056] A host computer system, which has advanced programming languages and technical frameworks, and is composed of modules such as data reception, preprocessing, parsing, user interaction, and security control.

[0057] The CLAHE algorithm divides the image into small regions and performs histogram equalization on each region respectively, while restricting the contrast to avoid noise amplification, thereby enhancing local details and improving the visual effect.

[0058] The YOLO v11 algorithm can automatically extract features and detect targets from the thermal image data, and then output the position information of the bounding box and key points of the human body in the thermal image data. According to the output results, the key points of the human body can be located and recognized in the thermal image.

[0059] The acupoint determination module determines the positions of acupoints according to the detected human key points by using the methods of triangulation and relative inch distance, and introducing the two-dimensional calibration method.

[0060] The location method of the present invention includes the following steps:

[0061] S1. Require the person to be measured to stand still at the specified position and maintain the posture. The infrared thermal imaging device captures the infrared radiation of the person to be measured and generates thermal image data, and the data is transmitted to the host computer system by wired or wireless means;

[0062] S2. The host computer system receives and preprocesses the generated thermal image data, including denoising, non-uniformity correction, and image quality enhancement;

[0063] S3. Use the YOLO v11 algorithm to detect the key points of the human body from the preprocessed thermal image data, form the human skeleton with the detected key points, identify the key parts on the surface of the person to be measured, and represent them with two-dimensional calibration;

[0064] S4. Based on the two-dimensional positioning marks and the relative cun distances between each acupoint and the key points of the human body, combined with the triangulation method, determine the accurate acupoint points of the measured person, reducing the positioning errors caused by body posture differences;

[0065] S5. Store the located acupoint points as coordinates in dictionary form, overlay them on the source image, and then upload them to the cloud and the client.

[0066] In step S1, the infrared thermal imaging device is a device used to capture infrared radiation and generate thermal images, used to collect infrared data and transmit it to the host computer system. It is connected to the host computer system through different physical interfaces, such as Ethernet (LAN), RS232, RS485, TCP / IP, UDP, etc., and can realize data transmission according to the specified communication protocol. At the same time, the infrared thermal imaging device ensures real-time performance and can stably output according to the set frame rate.

[0067] The host computer system includes a data receiving module, a preprocessing module for preprocessing and parsing the received data, a security control module for encrypting the data transmission and storage process, and an interactive interface for data display, device operation control, and system setting adjustment.

[0068] Preferably, step S2 further includes using the contrast-limited adaptive histogram equalization algorithm for data preprocessing.

[0069] Using the adaptive histogram equalization algorithm for preprocessing includes:

[0070] Image reading, obtaining the original infrared image from the infrared thermal imaging device;

[0071] Dividing the image area, dividing the entire image into multiple small blocks, and applying histogram equalization to each small block independently;

[0072] Histogram calculation and equalization, for each small block, calculate the pixel value distribution, and according to the calculated histogram, redistribute the pixel values within each small block to expand its dynamic range and enhance local contrast, while applying a limiting condition to prevent excessive amplification of noise;

[0073] Limiting the contrast, setting a contrast upper limit in the adaptive histogram equalization algorithm. If the count in a certain bin exceeds this limit, the excess part will be evenly distributed to other bins;

[0074] Result fusion, recombining all the equalized small blocks into a complete image, and ensuring the continuity and consistency between the small blocks through appropriate algorithms to avoid artificial traces;

[0075] That is, after collecting the infrared thermal image of the subject, it enters step S2 for data preprocessing, including preprocessing by the host computer system and preprocessing by the CLAHE algorithm (Contrast Limited Adaptive Histogram Equalization algorithm). The preliminary processing is completed by CLAHE. After the host computer receives the image, it immediately applies the CLAHE algorithm to enhance the local contrast of the image. This step can significantly improve the image quality, making subsequent key point detection and other analysis tasks more accurate. The image data after preliminary processing is segmented into small blocks suitable for further processing. Part of it is directly used for immediate applications, and the other part is sent to the host computer system for more complex analysis. The advanced processing is completed by the host computer. After receiving the preprocessed image data, the host computer uses its powerful computing resources and flexibility to perform more complex and time-consuming tasks, such as human key point detection and pose estimation using deep learning algorithms. At the same time, the host computer is also responsible for functions such as storing historical data, recording logs, and user interface display. This data preprocessing method highlights the advantages of the CLAHE algorithm and the host computer system. The CLAHE algorithm is good at processing images with uneven illumination or complex backgrounds and can significantly improve the visibility of local details while maintaining the overall structure. By using the CLAHE algorithm to complete most of the preprocessing work, the load on the host computer can be reduced, enabling it to focus on more complex computing tasks, thereby improving the efficiency of the entire system.

[0076] In step S3, YOLO v11 improves the Backbone and Neck architectures and adds C3k2 and C2PSA components.

[0077] In step S3, the preprocessed thermal image data is input into the trained YOLO v11 model. The model will automatically perform feature extraction and object detection on the thermal image data and output the position information of the bounding box and key points of the human body in the thermal image data. According to the output results, 17 key points of the human body can be located and recognized in the thermal image: 0: nose, 1: left eye, 2: right eye, 3: left ear, 4: right ear, 5: left shoulder, 6: right shoulder, 7: left elbow, 8: right elbow, 9: left wrist, 10: right wrist, 11: left hip, 12: right hip, 13: left knee, 14: right knee, 15: left ankle, 16: right ankle.

[0078] In step S3, YOLO v11 adopts an improved backbone and neck architecture, enhancing the feature extraction ability. It can more precisely capture the features of human key points in thermal image data, achieve more accurate key point localization, and contribute to improving the accuracy of subsequent applications such as acupoint localization. The preprocessed thermal image data is input into the trained YOLO v11 model. The model will automatically perform feature extraction and object detection on the thermal image data, and then output the position information of the bounding box and key points of the human body in the thermal image data. According to the output results, 17 key points of the human body can be located and recognized in the thermal image. Post-process the position information of the key points output by the model, such as filtering out key points with low confidence and performing smoothing processing on the key points. If there is a situation where key points are missing, consider throwing an exception and repeating this step.

[0079] Two-dimensional positioning marks in step S3:

[0080] Establish a two-dimensional rectangular coordinate system with a certain fixed reference point of the human body. For each detected human key point, record the coordinate values (x_key, y_key) in the above coordinate system. The coordinate values are obtained by the YOLO v11 model outputting the key point position information and combining the definition of the coordinate system for conversion.

[0081] Determine each acupoint (x_acupoint, y_acupoint) according to the coordinate values (x_key, y_key) of each key point and the relative position vector (Δx, Δy):

[0082] x_acupoint = Δx + x_key,

[0083] y_acupoint = Δy + y_key.

[0084] The steps of the triangulation method in step S4 are as follows: Select easily recognizable human key points as benchmarks, construct triangles or polygons, and obtain the exact position of the target acupoint by measuring the lengths of each side or angles.

[0085] Specifically, when selecting reference points, a set of stable and easily recognizable human body key points need to be selected as references. These points usually include but are not limited to the acromion, elbow, wrist, hip joint, knee joint, and ankle joint, etc. These key points are relatively fixed in different postures and can therefore be used as reliable references; construct triangles or polygons, use three or more selected reference points to construct one or more triangles. For example, to locate the Zusanli acupoint (located on the anterior lateral side of the lower leg, 3 cun below Dubi), three points, namely the upper outer edge of the patella (knee), the head of the fibula (below the outer side of the knee), and the tip of the medial malleolus, can be selected to form a triangle. Similarly, for other acupoints, corresponding combinations of reference points can also be found; measure and calculate, after establishing the triangle, the lengths of each side or angles can be measured to obtain the exact position of the target acupoint.

[0086] Whether it is the change in body size during the growth and development of the human body or the change in the local limb morphology caused by the change in posture, the method combining triangular positioning and relative cun distance can be flexibly adjusted to ensure the rationality of acupoint positioning.

[0087] Finally, in step S5, store the acupoint and its coordinate information, generate the corresponding dictionary and overlay it on the source image, and output the content after overlay.

[0088] As Figure 2 shown, the acupoint positioning system based on human body key points includes:

[0089] A data acquisition module for acquiring the infrared thermal image data of the measured person;

[0090] A data preprocessing module for transmitting the acquired data to the host computer system and simultaneously using the CLAHE algorithm for denoising, non-uniformity correction, and image quality enhancement;

[0091] A human body key point detection module, the YOLO v11 model will automatically perform feature extraction and target detection on the preprocessed thermal image data, and locate and identify the key points of the human body in the thermal image.

[0092] An acupoint determination module, using the method combining triangular positioning and relative cun distance, introducing a two-dimensional positioning mark to obtain the final acupoint positioning point.

[0093] An acupoint data visualization transmission module, storing the located acupoints and the corresponding coordinates in a dictionary, overlaying them on the source image, and then transmitting the result after overlay to the cloud and the client.

[0094] This system can achieve real-time dynamic monitoring of multiple acupoints on the human body surface. Long-term tracking and recording of the changing trends of these data helps to discover early warning signals. For example, a gradual increase in the temperature of certain acupoints may indicate the occurrence of inflammation; while fluctuations in resistance values may be related to changes in the nervous system. Precise acupoint positioning provides important clues for revealing the scientific principles behind traditional Chinese medicine theory. This invention is used to assist in disease prediction, promote personalized medical services, and drive the progress of medical research.

[0095] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for locating acupoints based on key points of the human body, characterized in that: The steps include: S1, requires the subject to stand at a designated position and maintain a posture, and the infrared thermal imaging device captures the subject's infrared radiation and generates thermal imaging data, which is transmitted to the host computer system via wired or wireless means; S2, the host computer system receives and preprocesses the generated thermal imaging data, including denoising, non-uniformity correction, and image quality enhancement; S3, using the YOLO v11 algorithm to detect the key points of the human body on the preprocessed thermal image data, composing the detected key points into a human skeleton, identifying the key parts of the subject's body surface, and representing them with two-dimensional calibration; S4, based on the two-dimensional positioning mark and the relative distance between each acupoint and the key points of the human body, combined with the triangulation positioning method, the precise acupoint of the subject is determined to reduce the positioning error caused by body posture differences; S5, the located acupuncture points are stored as coordinates in a dictionary form, superimposed on the source image, and then uploaded to the cloud and the client.

2. The acupoint locating method based on key points of the human body according to claim 1, characterized in that: The host computer system includes a data receiving module, a preprocessing module for preprocessing and parsing the received data, a security control module for encrypting the data transmission and storage process, and an interactive interface for data display, equipment operation control and system setting adjustment.

3. The acupoint locating method based on key points of the human body according to claim 1, characterized in that: The step S2 also includes performing data preprocessing using a contrast-limited adaptive histogram equalization algorithm.

4. The acupoint locating method based on key points of the human body according to claim 3, characterized in that: The preprocessing using the adaptive histogram equalization algorithm includes: Image reading, obtaining the original infrared image from the infrared thermal imaging device; Divide the image area, divide the entire image into multiple small blocks, and apply histogram equalization to each small block independently; Histogram calculation and equalization: For each small block, calculate the distribution of its pixel values, and redistribute the pixel values ​​in each small block according to the calculated histogram to expand its dynamic range and enhance the local contrast, while imposing a restriction to prevent excessive amplification of noise; Limit the contrast. Set a contrast upper limit in the adaptive histogram equalization algorithm. If the count in a bin exceeds this limit, the excess will be evenly distributed to other bins. The results are fused to reassemble all the equalized small blocks into a complete image, and appropriate algorithms are used to ensure the continuity and consistency between the small blocks to avoid artificial traces.

5. The acupoint locating method based on key points of the human body according to claim 1, characterized in that: In step S3, YOLO v11 improves the Backbone and Neck architectures and adds C3k2 and C2PSA components.

6. The acupoint locating method based on key points of the human body according to claim 5, characterized in that: In step S3, the preprocessed thermal imaging data is input into the trained YOLO v11 model. The model automatically performs feature extraction and target detection on the thermal imaging data, and outputs the position information of the bounding box and key points of the human body in the thermal imaging data. According to the output results, 17 key points of the human body can be located and identified in the thermal imaging image: 0: nose, 1: left eye, 2: right eye, 3: left ear, 4: right ear, 5: left shoulder, 6: right shoulder, 7: left elbow, 8: right elbow, 9: left wrist, 10: right wrist, 11: left hip, 12: right hip, 13: left knee, 14: right knee, 15: left ankle, 16: right ankle.

7. The acupoint locating method based on key points of the human body according to claim 1, characterized in that: The two-dimensional positioning mark in step S3: A two-dimensional rectangular coordinate system is established with a fixed reference point of the human body. For each detected key point of the human body, the coordinate value (x_key, y_key) in the above coordinate system is recorded. The coordinate value is obtained by converting the key point position information output by the YOLO v11 model in combination with the definition of the coordinate system. Determine each acupoint (x_acupoint, y_acupoint) based on the coordinate value (x_key, y_key) and relative position vector (Δx, Δy) of each key point: x_acupoint=Δx+x_key, y_acupoint=Δy+y_key.

8. The acupoint locating method based on key points of the human body according to claim 1, characterized in that: The steps of the triangulation positioning method in step S4 are as follows: select easily identifiable key points of the human body as a reference, construct a triangle or a polygon, and obtain the exact position of the target acupuncture point by measuring the length of each side or angle.

9. An acupoint positioning system based on key points of the human body, characterized in that: include: A data acquisition module, used to collect infrared thermal imaging data of the subject; The data preprocessing module is used to transmit the collected data to the host computer system and use the CLAHE algorithm to perform denoising, non-uniformity correction and image quality enhancement; Human key point detection module, the YOLO v11 model automatically performs feature extraction and target detection on the preprocessed thermal image data, and locates and identifies the key points of the human body in the thermal image. The acupuncture point determination module uses a method that combines triangulation positioning and relative distance, introduces a two-dimensional positioning mark, and obtains the final acupuncture point positioning point. The acupoint data visualization transmission module stores the located acupoints and their corresponding coordinates in a dictionary, superimposes them on the source image, and then transmits the superimposed results to the cloud and client.

Images