Intelligent moxibustion control system and method based on AR acupoint navigation

Abstract

The invention discloses an intelligent moxibustion control system and method based on AR acupoint navigation, and belongs to the technical field of data fusion processing, and the system comprises a moxibustion scheme generation module which is used for generating a corresponding basic moxibustion scheme according to obtained user disease data, and determining a target acupoint; the moxibustion heating path generation module is used for generating a heating path of a moxibustion head at each target acupuncture point during moxibustion; the interaction module is used for acquiring the disease data of the user and receiving, outputting and displaying the basic moxibustion scheme and the heating path; according to the scheme, the area where the target acupuncture point is located is reflected through multiple positioning modes in a probability distribution mode, then a point heating mode is changed into a line heating mode, the moxibustion head heating path is planned according to the moxibustion heating requirement, and finally the heating path is directly displayed in the real body image in an overlapped mode by means of the AR navigation unit. It is ensured that the target acupuncture point is subjected to sufficient dose of heat radiation without causing heat diffusion, the complexity of moxibustion operation is simplified, and meanwhile the moxibustion treatment effect is improved.

Description

Technical Field

[0001] This application belongs to the field of data fusion processing technology, and relates to an intelligent moxibustion control system and method based on AR acupoint navigation. Background Technology

[0002] Moxibustion is widely used in clinical practice in Traditional Chinese Medicine (TCM). It is a TCM therapy that uses the heat and medicinal properties generated by burning mugwort to stimulate acupoints on the body. Studies have shown that the mechanism of action of moxibustion mainly includes thermal stimulation and meridian regulation. The former uses the heat released by the moxibustion head to dilate blood vessels in the patient's local area (such as the area where the acupoint is located), thereby promoting blood circulation and accelerating metabolism. The latter stimulates acupoints to activate the activity of meridian qi, regulate the flow of qi and blood in the meridians, and enhance the body's disease resistance.

[0003] While moxibustion therapy (or physiotherapy) is now widely available in traditional Chinese medicine hospitals and health centers, many difficulties remain in extending its use to ordinary households. Firstly, ordinary households struggle to determine a moxibustion plan based on their individual symptoms; secondly, they find it difficult to pinpoint the exact locations of acupoints on the body; and finally, the effectiveness of each moxibustion treatment is difficult to quantify and record, hindering targeted optimization of the treatment plan. All these factors result in significantly reduced effectiveness for ordinary households, and improper operation can even cause harm.

[0004] To address the difficulty in pinpointing acupoint locations, existing technologies include intelligent acupoint location methods based on image recognition. These methods work by using image recognition to identify specific landmarks on the human body, then obtaining the relative positional relationship between these landmarks and the target acupoint based on an anatomical map of meridians and acupoints. Finally, the location of the target acupoint is determined based on this relative positional relationship. However, the accuracy of these methods is highly susceptible to influence by the patient's body shape and posture. Furthermore, regarding the difficulty in quantifying the effects of moxibustion, which hinders the optimization of moxibustion programs, current technologies lack quantifiable solutions. Typically, adjustments are made by physicians based on their professional experience and the patient's clinical presentation.

[0005] In summary, the key to promoting moxibustion therapy to ordinary family users lies in overcoming the problem of poor moxibustion treatment effects caused by inaccurate acupoint location and unsuitable moxibustion plans. Summary of the Invention

[0006] To address the problem of inaccurate acupoint location and unsuitable moxibustion plans leading to poor moxibustion treatment effects in practical applications, the first objective of this application is to provide an intelligent moxibustion control system based on AR acupoint navigation. This system utilizes diverse acupoint location methods and data fusion techniques to reflect the location of target acupoints in a probability distribution manner. It then replaces the existing point heating method with a line heating method, plans the heating path of the moxibustion head according to the moxibustion heating requirements, and finally uses an AR navigation unit to directly overlay and display the heating path on a real-world body image. This ensures that the target acupoint receives a sufficient dose of heat radiation without causing heat diffusion, thus preventing heat damage to the patient. Simultaneously, it can generate personalized moxibustion optimization plans based on relevant feedback data, improving the effectiveness of moxibustion treatment. Based on the above intelligent moxibustion control system, the second objective of this application is to provide an intelligent moxibustion control method based on AR acupoint navigation, the specific scheme of which is as follows:

[0007] An intelligent moxibustion control system based on AR acupoint navigation includes:

[0008] The moxibustion plan generation module has a built-in symptom-plan relationship model, which is configured to generate corresponding basic moxibustion plans based on the acquired user symptom data and determine the target acupoints.

[0009] The moxibustion heating path generation module is configured to generate the heating path of the moxibustion head at each target acupoint during moxibustion.

[0010] The interaction module is configured to acquire user symptom data, receive and output the basic moxibustion plan and heating path;

[0011] The moxibustion heating path generation module includes:

[0012] An image acquisition and recognition unit is configured to acquire images and identify set markers in the images according to a built-in marker recognition algorithm.

[0013] The statistical region generation unit has a built-in benchmark marker-acupoint location relationship model. It is configured to determine the location of the target acupoint based on the location of the identified benchmark marker, enlarge the region by a set ratio using the geometric center of the region as the control point, and then divide the enlarged region into multiple blocks of a set shape and size to form a probability statistical region.

[0014] A probability distribution generation unit with a built-in marker-acupoint location relationship model is configured to determine the location of the target acupoint based on the location of the identified marker and mark it as a region.

[0015] The second-class probability distribution generation unit has a built-in action command-feedback data-acupoint location relationship model, is configured to connect with the interaction module, outputs at least one action command and obtains feedback data after the user executes the above action command, determines the area where the target acupoint is located based on the feedback data, and marks it as the second-class area.

[0016] The probability distribution statistics unit is configured to calculate the projected area of ​​each block in the probability statistics area of ​​the first-class region / second-class region. If the projected area is greater than 1 / 2 of the block area, the block is associated with the name number of the first-class marker / action instruction and stored. The distribution probability value corresponding to each block is calculated based on the set calculation formula.

[0017] The path generation unit generates a closed-loop heating path based on the probability distribution value of each block in the probability statistics region, and uses the geometric center of each block as the heating stopping node. The dwell time of the moxibustion head at each heating stopping node is positively correlated with the magnitude of the probability distribution value of that block.

[0018] The basic moxibustion plan includes: target acupoints, moxibustion sequence, heating time and heating temperature for each target acupoint;

[0019] The set calculation formula is:

[0020]

[0021] In the above calculation formula, P represents the probability distribution, where M and N represent sets containing the names and numbers of a class of markers and the names and numbers of action commands, respectively. i P represents the probability that the current block is a target acupoint region if it contains a marker with name and number i. k This indicates the probability that the current block is the target acupoint area if the current block is associated with an action instruction with name and number k; β1 and β2 are the weight values ​​for calculation.

[0022] Through the above technical solution, based on the symptom-treatment relationship model, users can determine a basic moxibustion treatment plan according to their current symptom data. Then, the moxibustion heating path generation module generates heating paths corresponding to each target acupoint, changing the traditional point heating method to a line heating method, ultimately heating the entire area of ​​the target acupoint. When determining the heating path for the target acupoint, the results obtained through various acupoint location determination methods (acupoint location distribution) are fully considered. The heating path is finally obtained through weighted calculation and statistical distribution probability. Simultaneously, the heating duration of each heating stop node on the heating path is determined based on the above probability distribution, thereby allowing areas with high probability of acupoint location to receive a higher dose of heat radiation, achieving more precise heating. Throughout the entire moxibustion control process, users do not need to know the moxibustion treatment plan corresponding to various symptoms in advance, nor do they need to precisely locate the target acupoint, greatly reducing the difficulty of moxibustion operation, improving the therapeutic effect of moxibustion, and helping to promote moxibustion treatment to ordinary home users.

[0023] Furthermore, the aforementioned type of markers includes anatomical surface landmarks related to the target acupoint, body contours, and specific reference objects used to indicate the distance and positional relationship between the target acupoint and a specific marker;

[0024] The action instructions include stimulating the area where the suspected target acupoint is located or stimulating related acupoints associated with the target acupoint.

[0025] The stimulation methods include one or more of the following: tapping, pressing, kneading, itching, electrical stimulation, and thermal stimulation; the feedback data includes one or more of the following: tenderness index, induration index, temperature change, tactile sensitivity index, thermal sensitivity index, and nerve reaction speed.

[0026] The above technical solutions determine the location of target acupoints through two approaches: visual judgment and tactile stimulation response judgment. Based on image recognition technology, the location of target acupoints can be inferred by selecting different types of markers, which improves the accuracy of acupoint location and reduces the influence of user body shape and posture on the accuracy of acupoint location. Furthermore, by introducing active stimulation and collecting feedback data, the location of target acupoints can be determined from the user's physiological responses, further improving the accuracy of locating the target acupoint area.

[0027] Furthermore, the shape of the block is an equilateral triangle, a square, or a regular hexagon;

[0028] The path generation unit includes:

[0029] The block integration coefficient acquisition subunit is configured to acquire the effective heat radiation area of ​​a fixed point heated by the moxibustion head during moxibustion, calculate the ratio of the effective heat radiation area to the block area, and round the ratio to obtain the block integration coefficient.

[0030] The block integration subunit is configured to integrate the blocks in the probability statistics region according to the above block integration coefficient, divide them into multiple reference regions, calculate the average value of the distribution probability value corresponding to each block in the above reference regions, and store them in association with the reference regions.

[0031] The heating path generation subunit is configured to generate a closed-loop heating path with the geometric center of the reference area as the heating stopping node.

[0032] The heating path can be arranged in the following ways: spiraling inward along the edge of the probability statistics region, spiraling outward from the center of the probability statistics region, or generating the heating path according to the average distribution probability value corresponding to each reference region.

[0033] The above technical solution can determine the heating path based on the single-point heat radiation area of ​​the moxibustion head, avoiding repeated heating of certain areas. At the same time, the use of diverse path planning schemes facilitates the acquisition of the optimal heating path through subsequent analysis, which is beneficial for optimizing the moxibustion heating path.

[0034] Furthermore, the moxibustion heating path generation module also includes:

[0035] The path correction unit is connected to the image acquisition and recognition unit, selects at least one marker in the image as a motion reference, acquires the motion characteristics of the probability statistical region and / or the displacement of the moxibustion head itself during moxibustion, and adjusts the heating path according to the motion characteristics.

[0036] The motion characteristics include the distance and period of the reciprocating displacement of the probability statistics region and / or the moxibustion head in a certain direction during moxibustion.

[0037] Adjusting the arrangement of the heating path according to the motion characteristics includes: adjusting the heating path in the opposite direction of the reciprocating displacement, and the adjustment time being adapted to the period of the reciprocating displacement.

[0038] The above technical solution can overcome the deviation in heating position caused by the patient's body movement or the shaking of the moxibustion head during moxibustion, ensuring that the heating points of the moxibustion head are always arranged along the predetermined heating path, thereby improving the therapeutic effect of moxibustion.

[0039] Furthermore, the interaction module includes a data input terminal and a data output terminal; wherein,

[0040] The data input terminal includes:

[0041] The voice input unit is configured to collect the user's voice information and convert it into input data or commands.

[0042] The gesture recognition unit is configured to acquire and recognize the user's gesture actions, and convert them into input data or instructions based on a set gesture recognition model.

[0043] Click the input unit, which is configured to detect user actions on the interactive interface and convert them into input data or instructions according to the interface settings.

[0044] A data retrieval unit is configured to retrieve target data stored in an external data storage device as input data.

[0045] The data output terminal includes:

[0046] An image display unit is configured to display corresponding text or image information on an interactive interface;

[0047] The AR navigation unit is configured to overlay and display the distribution area of ​​the target acupoint, its heating path, efficacy description, and precautions information onto real-time captured images.

[0048] The voice playback unit is configured to output moxibustion guidance information and feedback data inquiry information via voice.

[0049] The vibration output unit is configured to output vibrations of different frequencies to represent different prompts.

[0050] Through the above technical solutions and with the help of the interactive module, users can interact with the system through various interactive methods to exchange data and information. By introducing an AR navigation unit, the location of the target acupoint and the corresponding heating path can be superimposed on the real-time captured user image, reducing the difficulty of moxibustion operation and making it easier for users to perform moxibustion accurately. By configuring voice playback and vibration output functions, the system can provide comprehensive and multi-faceted guidance for users' moxibustion actions, clarify moxibustion precautions, and improve the effect of moxibustion.

[0051] Furthermore, the intelligent moxibustion control system also includes a heating path optimization module, comprising:

[0052] The physiological characteristic parameter acquisition unit is configured to acquire physiological characteristic parameter data of a set category within a set time period before and after moxibustion.

[0053] The data analysis unit is configured to analyze the natural change trend of the physiological characteristic parameter data of the above-defined category before moxibustion and the response change trend after moxibustion based on the physiological characteristic parameter data, and generate a trend difference value between the response change trend and the natural change trend.

[0054] The data association unit is configured to obtain the distribution probability value of each block along each heating stop node in the heating path, generate a one- or two-dimensional data matrix, and associate and store the above two-dimensional data matrix, physiological characteristic parameter categories and their corresponding trend difference values ​​as training data.

[0055] The correlation analysis unit is configured to acquire a set amount of training data to form a training dataset, and generate the correlation between the numerical distribution of elements and the trend difference values ​​in a two-dimensional data matrix based on neural network analysis or data correlation algorithm analysis.

[0056] The path optimization unit is configured to adjust the element values ​​in the two-dimensional data matrix based on the required trend difference value according to the above correlation, and to adjust the heating path in reverse according to the adjusted two-dimensional data matrix.

[0057] The categories of physiological characteristic parameters include blood glucose, blood pressure, blood lipids, alanine aminotransferase, urine protein, hemoglobin, thyroid-stimulating hormone, and related physiological characteristic parameters that reflect the data of the above categories of physiological characteristic parameters.

[0058] The above technical solution utilizes data correlation analysis to obtain the relationship between the changing trends of physiological characteristic parameters of a set category and the moxibustion heating path. Based on this relationship, the two-dimensional data matrix can be optimized, i.e., the heating path can be optimized. This optimization process does not involve the basic moxibustion protocol; instead, it focuses on the moxibustion heating process of a single target acupoint. Furthermore, the optimization is limited to a probabilistic statistical range, thus avoiding side effects and making it suitable for ordinary home users and those without professional moxibustion knowledge.

[0059] A smart moxibustion control method based on AR acupoint navigation includes:

[0060] Acquire user symptom data and generate corresponding basic moxibustion plans based on the symptom-plan relationship model to determine target acupoints;

[0061] Configure and display the heating path of the moxibustion head for each target acupoint during moxibustion.

[0062] Generating and configuring the heating path includes:

[0063] Acquire and store a marker recognition algorithm for identifying each type of marker from an image, a marker-acupoint positional relationship model for reflecting the relative positional relationship between each type of marker and each acupoint, a reference marker-acupoint positional relationship model for reflecting the relative positional relationship between a reference marker and associated acupoints, and an action instruction-feedback data-acupoint positional relationship model for reflecting the relationship between feedback data after a user executes a specific action instruction and the position of each acupoint;

[0064] Acquire a user's body image and identify markers in the image based on the marker recognition algorithm;

[0065] Based on the identified reference markers, and combined with the reference marker-acupoint location relationship model, the location area of ​​the target acupoint is determined. The area is then magnified by a set ratio using the geometric center of the area as the control point. The magnified area is then divided into multiple blocks of a set shape and size to form a probability statistical area.

[0066] Based on the identified type of markers, and combined with the type of marker-acupoint location relationship model, the location area of ​​the target acupoint is determined, marked as a type of area, and stored in association with the name and number of the type of markers.

[0067] Based on the action instruction-feedback data-acupoint location relationship model, at least one action instruction is output through the interaction module and feedback data after the user executes the above action instruction is obtained. The area where the target acupoint is located is determined based on the feedback data, marked as a second-class area, and stored in association with the action instruction name number.

[0068] The projected area of ​​each type I / type II region on each block in the probability statistics area is counted. If the ratio of the projected area to the block area exceeds a set value, the block is associated with the name number of the type I marker / action instruction and stored. The distribution probability value corresponding to each block is calculated based on the set calculation formula.

[0069] Based on the probability distribution value of each block in the probability statistics region, a closed-loop heating path is generated, with the geometric center of each block as the heating stopping node. The dwell time of the moxibustion head at each heating stopping node is positively correlated with the magnitude of the probability distribution value of that block.

[0070] The basic moxibustion plan includes: target acupoints, moxibustion sequence, heating time and heating temperature for each target acupoint;

[0071] The aforementioned type of markers includes anatomical surface landmarks related to the target acupoint, body outlines, and specific reference objects used to indicate the distance and positional relationship between the target acupoint and a specific marker;

[0072] The action instructions include stimulating the area where the suspected target acupoint is located or stimulating related acupoints associated with the target acupoint.

[0073] The stimulation methods include one or more of the following: tapping, pressing, kneading, tingling, electrical stimulation, and thermal stimulation. The feedback data includes one or more of the following: tenderness index, induration index, temperature change, tactile sensitivity index, thermal sensitivity index, and nerve reaction speed.

[0074] The set calculation formula is:

[0075]

[0076] In the above calculation formula, P represents the probability distribution, where M and N represent sets containing the names and numbers of a class of markers and the names and numbers of action commands, respectively. i P represents the probability that the current block is a target acupoint region if it contains a marker with name and number i. k This indicates the probability that the current block is the target acupoint area if the current block is associated with an action instruction with name and number k; β1 and β2 are the weight values ​​for calculation.

[0077] Furthermore, obtaining user symptom data includes:

[0078] Establish a data connection between the system and an external data storage device to directly retrieve the user's symptom data from the external data storage device; and / or

[0079] Collect and recognize user voice information or gesture images to obtain user symptom data or commands; and / or

[0080] Detect user clicks on the interactive interface and convert them into input data or commands according to the interface settings;

[0081] Configure and display the heating path of the moxibustion head for each target acupoint during moxibustion, including:

[0082] Displaying corresponding text or image information on the interactive interface; and / or

[0083] The distribution area of ​​the target acupoint, its heating path, efficacy description, and precautions are overlaid on real-time captured images; and / or

[0084] The system outputs moxibustion guidance information, feedback data, inquiry information, and prompts via voice or vibration.

[0085] The above technical solution can greatly simplify the operation of moxibustion for users, enabling ordinary household users to accurately locate target acupoints and heat them according to the set heating path and method, thereby improving the therapeutic effect of moxibustion.

[0086] Furthermore, the shape of each block in the probability statistics region is set as an equilateral triangle, a square, or a regular hexagon;

[0087] Generating the closed-loop heating path includes:

[0088] The effective heat radiation area of ​​a fixed point heated by the moxibustion head during moxibustion is obtained. The ratio of the effective heat radiation area to the area of ​​the block is calculated, and the block integration coefficient is obtained by rounding the ratio.

[0089] Based on the above block integration coefficient, the blocks in the probability statistics region are integrated, the probability statistics region is divided into multiple reference regions, the average value of the distribution probability value corresponding to each block in the above reference regions is calculated and stored in association with the reference regions.

[0090] A closed-loop heating path is generated by connecting the geometric center of the reference area as the heating stopping node.

[0091] The arrangement of the heating path includes: spiraling inward along the edge of the probability statistics region through each heating stopping node and then returning; spiraling outward from the center of the probability statistics region through each heating stopping node and then returning; or sequentially connecting the heating paths according to the magnitude of the average distribution probability value corresponding to each reference region.

[0092] The above technical solution can avoid repeated heating of certain skin areas during the movement of the moxibustion head.

[0093] Furthermore, the intelligent moxibustion control method also includes a heating path optimization step, comprising:

[0094] Obtain physiological characteristic parameter data of a set category within a set time period before and after moxibustion for the user;

[0095] Based on the physiological characteristic parameter data, the natural change trend of the physiological characteristic parameter data of the above-mentioned set category before moxibustion and the response change trend of the physiological characteristic parameter data of the above-mentioned set category after moxibustion are analyzed, and the trend difference value between the response change trend and the natural change trend is generated.

[0096] Along each heating stop node in the heating path, the distribution probability value corresponding to each block is obtained, and a one- or two-dimensional data matrix is ​​generated. The above two-dimensional data matrix, physiological characteristic parameter categories and their corresponding trend difference values ​​are associated and stored as training data.

[0097] A set amount of training data is acquired to form the training dataset corresponding to the current user. Based on neural network analysis or data correlation algorithm analysis, the correlation between the distribution of element values ​​in a two-dimensional data matrix and the trend difference value is generated.

[0098] Based on the above correlation and the required trend difference value, adjust the element values ​​in the two-dimensional data matrix, and then adjust the heating path in reverse according to the adjusted two-dimensional data matrix.

[0099] This application includes at least one of the following beneficial effects:

[0100] (1) Based on the disease-plan relationship model, it can help users determine the basic moxibustion plan according to the current disease data, and then generate the heating path corresponding to each target acupoint according to the moxibustion heating path generation module, changing the inherent point heating method to the line heating method, and finally from line to surface, to realize the heating of the entire target acupoint area.

[0101] (2) When determining the heating path of the target acupoint, the results obtained by various acupoint location determination methods (acupoint location distribution) are fully considered. The heating path is finally obtained by weighted calculation and statistical distribution probability. At the same time, the heating duration of each heating stop node on the heating path is determined according to the above probability distribution, so that the area where the acupoint is located receives a higher thermal radiation dose and achieves more precise heating.

[0102] (3) Throughout the entire moxibustion control process, users do not need to know the moxibustion plan corresponding to various diseases in advance, nor do they need to locate the target acupoints very accurately. This greatly reduces the difficulty of moxibustion operation, improves the therapeutic effect of moxibustion, and helps to promote moxibustion treatment to ordinary family users. Attached Figure Description

[0103] Figure 1 This is a schematic diagram of the overall functional modules of the intelligent moxibustion control system;

[0104] Figure 2 A schematic diagram illustrating part of the data mapping relationship in the symptom-treatment relationship model;

[0105] Figure 3 A schematic diagram of the moxibustion heating path generation module;

[0106] Figure 4 This is a schematic diagram of probability statistics regions and blocks;

[0107] Figure 5 This is a structural diagram of the interactive module;

[0108] Figure 6 A schematic diagram of the heating path optimization module;

[0109] Figure 7 This is a schematic diagram of the intelligent moxibustion control method of this application;

[0110] Figure 8 A schematic diagram illustrating the method for generating and configuring heating paths;

[0111] Figure 9 A schematic diagram illustrating the steps for optimizing the heating path.

[0112] Figure labels: 100, Moxibustion plan generation module; 200, Moxibustion heating path generation module; 210, Image acquisition and recognition unit; 220, Statistical region generation unit; 230, Type I probability distribution generation unit; 240, Type II probability distribution generation unit; 250, Probability distribution statistics unit; 260, Path generation unit; 2601, Block integration coefficient acquisition subunit; 2602, Block integration subunit; 2603, Heating path generation subunit; 270, Path correction unit; 300, Interaction module; 310, Data input terminal; 31 01. Voice input unit; 3102. Gesture recognition unit; 3103. Click input unit; 3104. Data retrieval unit; 320. Data output terminal; 3201. Image display unit; 3202. AR navigation unit; 3203. Voice playback unit; 3204. Vibration output unit; 400. Heating path optimization module; 410. Physiological characteristic parameter acquisition unit; 420. Data analysis unit; 430. Data association unit; 440. Correlation analysis unit; 450. Path optimization unit; 5. Probability statistics area; 6. Block. Detailed Implementation

[0113] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0114] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0115] An intelligent moxibustion control system based on AR acupoint navigation, such as Figure 1 As shown, it mainly includes the following functional modules: moxibustion plan generation module 100, moxibustion heating path generation module 200, and interaction module 300.

[0116] The moxibustion plan generation module 100 has a built-in disease-plan relationship model that reflects the correspondence between various symptom data and moxibustion plans, such as... Figure 2The diagram shows a partial data mapping relationship in the aforementioned symptom-treatment relationship model. In this embodiment, the moxibustion treatment plan generation module 100 is configured to generate a corresponding basic moxibustion treatment plan based on the acquired user symptom data, ultimately determining the target acupoints requiring moxibustion. The basic moxibustion treatment plan includes the target acupoints, the moxibustion sequence, the heating time for each target acupoint, and the heating temperature. The user, according to the basic moxibustion treatment plan, manually or with the aid of an automatic moxibustion device, such as a robotic arm, holds the moxibustion head and heats each target acupoint according to the predetermined moxibustion sequence.

[0117] The moxibustion heating path generation module 200 is configured to generate the heating path of the moxibustion head at each target acupoint during moxibustion. Combined with... Figure 3 As shown, the moxibustion heating path generation module 200 mainly includes: an image acquisition and recognition unit 210, a statistical region generation unit 220, a first-class probability distribution generation unit 230, a second-class probability distribution generation unit 240, a probability distribution statistics unit 250, and a path generation unit 260.

[0118] The image acquisition and recognition unit 210 includes a smartphone or tablet computer with image capture function, configured to acquire images of the user's body and identify set markers in the image according to a built-in marker recognition algorithm. The set markers include, but are not limited to, body contours and anatomical landmarks, such as the navel, the tip of the nose, and bony protrusions. The location of relevant acupoints can be determined using the above markers. For example, the location of the Suliao acupoint can be determined by the tip of the nose, and the location of the Tianzong acupoint can be determined by the central depression of the infraspinous fossa of the scapula.

[0119] The statistical region generation unit 220 has a built-in benchmark marker-acupoint location relationship model, which is configured to determine the location of the target acupoint based on the location of the identified benchmark marker, enlarge the region by a set ratio using the geometric center of the region as the control point, and then divide the enlarged region into multiple blocks 6 of a set shape and size to form a probability statistical region 5. In practical applications, the location of an acupoint can usually be determined by using multiple different landmarks as references. The aforementioned benchmark landmarks are the most relevant surface landmarks to the acupoint and are less susceptible to interference. For example, the location of Taichong acupoint on the dorsum of the foot can be determined by referring to the position of the big toe and the second toe, or by the depression below and in front of the junction of the first and second metatarsal bones on the dorsum of the foot, or by inferring from the contours of the sides of the foot. Among the three types of landmarks, the depression below and in front of the junction of the first and second metatarsal bones is the most accurate for determining Taichong acupoint. However, in practice, patients often have lower limb edema or obesity, making it difficult to obtain the above landmarks through image recognition. Furthermore, the contours of the sides of the foot may vary due to individual differences. Therefore, the big toe and the second toe are used as benchmark landmarks for Taichong acupoint.

[0120] The target acupoint area determined based on the reference marker and according to the reference marker-acupoint location relationship model is actually a skin area of ​​0.5-1 cm² (the diameter of the acupoint is usually 1-10 mm). From the above-described target acupoint location process, it can be seen that the acupoint location obtained based on the reference marker may be affected by various factors, and its accuracy is not very precise. Therefore, in this embodiment, the geometric center of the area (usually the center of a circle or the center point of a square) is used as the control point to enlarge the area by a set ratio. To avoid the generated probability statistical region 5 being too large, the enlargement ratio is preferably set to 1-1.2. To seamlessly segment the probability statistical region 5, combined with... Figure 4 As shown, the shape of the block 6 is preferably configured as an equilateral triangle, a square, or a regular hexagon.

[0121] A probability distribution generation unit 230 has a built-in marker-acupoint positional relationship model to reflect the relative positional relationship between each type of marker and the relevant acupoint. This model is configured to determine the location of the target acupoint based on the identified location of the markers and mark it as a type of region. In this embodiment, the markers include, but are not limited to, anatomical surface landmarks related to the target acupoint, body contours, and specific reference objects used to represent the distance and positional relationship between the target acupoint and the specific markers. These specific reference objects typically include specific parts of the fingers, such as taking the width of the thumb joint as one inch (finger-inch acupoint location method). When determining the location of the target acupoint, the thumb is placed next to a surface landmark, and the actual length of "one inch" is obtained by a relevant algorithm (such as the SLAM algorithm). Then, based on the actual length and the actual position of the surface landmark, the region where the target acupoint is located is inferred.

[0122] The Class II probability distribution generation unit 240 has a built-in action command-feedback data-acupoint location relationship model, configured to connect with the interaction module 300. It outputs at least one action command and acquires feedback data after the user executes the command. Based on the feedback data, it determines the location of the target acupoint and marks it as a Class II region. The action commands include, but are not limited to, stimulating the area of ​​the suspected target acupoint or stimulating related acupoints associated with the target acupoint. Stimulation methods include one or more combinations of tapping, pressing, kneading, itching, electrical stimulation, and thermal stimulation. Feedback data includes one or more combinations of tenderness index, induration index, temperature change, tactile sensitivity index, thermal sensitivity index, and nerve reaction speed. For example, when determining the location of the Taichong acupoint, the location is marked on the dorsum of the foot using AR imaging. Simultaneously, the interaction module 300 outputs a voice prompt instructing the user to press the marked area, acquiring the user's tenderness index. If a significant soreness is felt, the marked area is confirmed as the location of the Taichong acupoint, and the area is marked as a Class II region.

[0123] The probability distribution statistics unit 250 is configured to calculate the projected area of ​​each block 6 in the probability statistics area 5 for the first-class / second-class regions. If the ratio of the projected area to the area of ​​block 6 is greater than a set value, such as 0.5, the block 6 is associated with the name number of the first-class marker / action instruction and stored. The probability distribution value corresponding to each block 6 is calculated based on the set calculation formula.

[0124] The set calculation formula is configured as follows:

[0125]

[0126] In the above calculation formula, S(M,N) is the probability distribution value, and M and N represent sets containing the name number of a type of marker and the name number of an action command, respectively. For example, when confirming the location of the Neiguan acupoint, it can be determined by the wrist crease (a distinct transverse crease on the inside of the wrist) or by the flexor carpi radialis tendon (a longitudinal tendon that is adjacent to the palmaris longus tendon and close to the thumb). The name number of the former is set as A, and the name number of the latter is set as B. Similarly, if there is another type of marker, the name number is set as C, and M={A,B,C}.

[0127] P i P represents the probability that the area containing the current block 6 is the target acupoint area if the current block 6 is associated with a marker of type i. k This represents the probability that the area containing block 6 is the target acupoint area if the current block 6 is associated with an action instruction with name and number k. Due to differences in body shape and physique among users, the above P is applied to each user. i and P k The values ​​are also different. β1 and β2 are the calculation weights, and their sum is set to 1. In practical applications, β2 > β1, indicating that the accuracy of acupoint locations determined through feedback after actions such as pressing is higher. In practical applications, P related to each acupoint i and P k The value can be stored in the system memory in advance as a data storage table, and can be directly retrieved when needed.

[0128] like Figure 4 Block 6, numbered 0007 (the area indicated by the arrow in the diagram), is associated with three Class I markers (A, B, and C) and with the names and numbers of two action instructions (D and E). If P... A =0.9, P B =0.8, P C =0.6, P D =0.9, P E=0.6, β1 and β2 are 0.4 and 0.6 respectively, then the probability distribution value corresponding to block 6 with number 0007 is 0.4*(0.9+0.8+0.6)+0.6*(0.9+0.6)=1.82.

[0129] The path generation unit 260 generates a closed-loop heating path based on the probability distribution values ​​corresponding to each block 6 in the probability statistics region 5, using the geometric center of each block 6 as the heating stopping node. In practical applications, if the probability distribution value corresponding to block 6 is less than a certain set value, then the block 6 will not be included in the heating path.

[0130] In the optimized embodiment of this application, the moxibustion head passes through various heating stopping nodes during its movement along the heating path. The duration of the moxibustion head's stay at each heating stopping node is positively correlated with the probability value of the distribution in block 6. Specifically, the heating duration at each heating stopping node is calculated as follows:

[0131] The theoretical heating time corresponding to the current target acupoint is obtained from the basic moxibustion plan. This theoretical heating time is multiplied by a set factor, such as 1.05, increasing the original 20-minute heating time to 21 minutes. The purpose of this time expansion is to compensate for the portion of the heat from the moxibustion head that does not reach the target acupoint due to acupoint positioning deviation. The sum of the probability distribution values ​​of each block 6 along the heating path is calculated. The expanded time is divided by the sum of the probability distribution values ​​to calculate the correspondence between the unit probability distribution value and the heating time. Finally, the heating time is allocated to each block 6 according to the probability distribution value corresponding to each block 6, that is, the dwell time of the moxibustion head at each heating stopping point during the movement. To prevent excessive heating time at a certain block 6 from causing the heat radiation dose to exceed the safety threshold, in practical applications, the product of heating time and temperature is also calculated when generating the heating path (if the moxibustion head temperature is a variable value, then the time-temperature calculation needs to be performed). If the above product corresponding to block 6 exceeds the safety threshold, the heating time corresponding to each current heating stop node is divided by 2. Corresponding to the movement of the moxibustion head, this means that the moxibustion head needs to move one more circle along the heating path. This setting is continued until the above product corresponding to each block 6 does not exceed the safety threshold.

[0132] like Figure 3 As shown, the path generation unit 260 further includes: a block integration coefficient acquisition subunit 2601, a block integration subunit 2602, and a heating path generation subunit 2603.

[0133] The block integration coefficient acquisition subunit 2601 is configured to acquire the effective heat radiation area of ​​a fixed point heated by the moxibustion head during moxibustion, calculate the ratio of the effective heat radiation area to the area of ​​block 6, and round the ratio to obtain the block 6 integration coefficient. In practical applications, the effective heat radiation area is related to the temperature of the moxibustion head and the distance between the moxibustion head and the skin. Effective heat radiation refers to the temperature reaching the acupoint when it reaches a set value. In moxibustion, the expected therapeutic effect can only be achieved when the temperature is higher than a certain set value. Conversely, if the heating temperature is too low, even if the heating time exceeds the set limit, the expected therapeutic effect cannot be achieved.

[0134] The block integration subunit 2602 integrates block 6 in the probability statistics region 5 according to the integration coefficient of block 6 mentioned above, dividing it into multiple reference regions. It calculates the average distribution probability value corresponding to each block 6 in the reference regions and stores it in association with the reference regions. The block integration subunit 2602 is used to integrate multiple blocks 6 into a large reference region, for example, integrating four square blocks 6 into one reference region. The heating path generation subunit 2603 uses the geometric center of the reference region as the heating stopping node, connecting each heating stopping node to generate a closed-loop heating path. The arrangement of the heating path includes: spiraling inward along the edge of the probability statistics region 5, spiraling outward from the center of the probability statistics region 5, or generating an arrangement based on the average distribution probability value corresponding to each reference region. The above technical solution can determine the heating path based on the single-point heat radiation area when the moxibustion head is heated, avoiding repeated heating of certain areas. At the same time, it adopts diverse path planning schemes, facilitating later analysis to obtain the optimal heating path and promoting the optimization of the moxibustion heating path.

[0135] In practical applications, the optimized block 6 is set up in a regular hexagon or square shape, and the area of ​​block 6 is equal to or close to the aforementioned effective thermal radiation area.

[0136] The above-described solution is particularly suitable for automated moxibustion devices equipped with robotic arms, which can precisely heat target acupoints according to a basic moxibustion plan and the corresponding heating path. For ordinary home users, heating can also be performed along a real-time marked heating path in the user's body image using AR imaging.

[0137] To overcome the deviation in heating position caused by patient body movements during moxibustion, such as periodic reciprocating movements of a localized area due to breathing, or the shaking of the moxa head, and to ensure that the heating points of the moxa head are always arranged along a predetermined heating path, thereby improving the therapeutic effect of moxibustion, the aforementioned moxibustion heating path generation module 200 further includes a path correction unit 270. The path correction unit 270 is data-connected to the image acquisition and recognition unit 210, selects at least one landmark in the image as a motion reference, acquires the motion characteristics of the probability statistical region 5 and / or the displacement of the moxa head itself during moxibustion, and adjusts the heating path according to the motion characteristics. The aforementioned motion characteristics include the distance and period of the reciprocating displacement of the probability statistical region 5 and / or the moxa head in a certain direction during moxibustion. Correspondingly, the method of adjusting the heating path arrangement according to the motion characteristics includes: adjusting the heating path in the opposite direction of the reciprocating displacement, with the adjustment duration matching the period of the reciprocating displacement. The aforementioned path correction unit 270 can not only adapt to the movement of the target acupoint during moxibustion but also to the shaking of the moxa cone during moxibustion.

[0138] Combination Figure 3 As shown in the embodiment of this application, the interaction module 300 is mainly used to acquire user symptom data, receive and output display of basic moxibustion plan and heating path.

[0139] Detailed, such as Figure 5 As shown, the interaction module 300 includes a data input terminal 310 and a data output terminal 320. In this embodiment, the interaction module 300 uses a smartphone or tablet computer to implement the relevant functions.

[0140] The data input terminal 310 includes: a voice input unit 3101, a gesture recognition unit 3102, a click input unit 3103, and a data retrieval unit 3104. The voice input unit 3101 is configured to collect the user's voice information and convert it into input data or commands. The gesture recognition unit 3102 is configured to acquire the user's gesture image using the image capturing device in the interaction module 300, then recognize the user's gesture action, and convert it into input data or commands based on a set gesture recognition model. In practical applications, the above-mentioned voice input and gesture recognition are both implemented through an APP configured on a mobile phone that has voice recognition and gesture action recognition capabilities. The click input unit 3103 is configured to detect the user's operation on the interactive interface and convert it into input data or commands according to the interface settings, such as displaying a dialog confirmation box in the mobile APP interface, and then acquiring or confirming relevant data information based on the user's click or input operation. The data retrieval unit 3104 is configured to connect to an external data storage device and retrieve target data (such as heart rate and blood pressure data) stored on the external data storage device, such as a smart bracelet, as input data.

[0141] The data output terminal 320 includes: an image display unit 3201, an AR navigation unit 3202, a voice playback unit 3203, and a vibration output unit 3204. The image display unit 3201 is configured to display corresponding text or image information on the interactive interface, such as detailed efficacy descriptions, personalized moxibustion duration suggestions (e.g., 15-20 minutes for Zusanli acupoint), and operation contraindications (e.g., direct moxibustion is prohibited on facial acupoints). The AR navigation unit 3202 is configured to overlay the distribution area of ​​the target acupoint, its heating path, efficacy descriptions, and precautions onto the real-time captured image. The AR navigation unit 3202 is developed based on the mature SDK of the Baidu AR Open Platform. After the user selects the corresponding ailment through the APP, the system calls the phone's camera to scan the body area, combining computer vision algorithms and a traditional Chinese medicine meridian database to accurately mark the acupoint location in a dynamic animation (error controlled within ≤2mm), and simultaneously displays the heating path, acupoint name, anatomical structure diagram, and related ailment information. The voice playback unit 3203 is configured to output moxibustion guidance information and feedback data inquiry information via voice, and the vibration output unit 3204 is configured to output vibrations of different frequencies to represent different prompts, such as outputting a vibration prompt when the position of the moxibustion head has not moved for a long time.

[0142] Through the interactive module 300 described in the above technical solution, users can interact with the system through various interactive methods to exchange data and information. By introducing the AR navigation unit 3202, the location of the target acupoint and the corresponding heating path can be superimposed on the real-time captured user image, reducing the difficulty of moxibustion operation and making it easier for users to perform moxibustion operations accurately. By configuring voice playback and vibration output functions, the system can provide comprehensive and multi-form guidance for users' moxibustion actions, clarify moxibustion precautions, and improve the effect of moxibustion.

[0143] To continuously optimize the moxibustion treatment plan based on user feedback, the intelligent moxibustion control system described in this application also includes a heating path optimization module 400, combined with... Figure 6 As shown, it specifically includes a physiological characteristic parameter acquisition unit 410, a data analysis unit 420, a data association unit 430, a correlation analysis unit 440, and a path optimization unit 450.

[0144] The physiological characteristic parameter acquisition unit 410 is configured to acquire physiological characteristic parameter data of a set category within a set time period before and after moxibustion. These physiological characteristic parameter categories include blood glucose, blood pressure, blood lipids, alanine aminotransferase (ALT), urine protein, hemoglobin, thyroid-stimulating hormone (TSH), and related physiological characteristic parameters reflecting the above categories of physiological characteristic parameter data. The collection of these physiological characteristic parameter data includes using smart bracelets with health monitoring functions, smart blood pressure / blood glucose meters, smart blood testing devices, etc. In this embodiment, the primary focus is on acquiring blood glucose / blood pressure data within a set time period before and after moxibustion.

[0145] The data analysis unit 420 is configured to analyze the natural trend of physiological characteristic parameter data of the aforementioned preset categories before moxibustion and the response trend after moxibustion, based on the physiological characteristic parameter data, and generate a trend difference value between the response trend and the natural trend. The natural trend refers to the natural trend of a certain physiological characteristic parameter data when the user has not received moxibustion, such as blood sugar gradually decreasing over time after meals. The response trend refers to the trend of a certain physiological characteristic parameter data after moxibustion. The trend difference value is the difference between the natural trend and the response trend of a certain physiological characteristic parameter data, such as the difference in the trend before and after moxibustion at a fixed time after a meal. By analyzing the trend of statistical data rather than the numerical value, data collection errors and interference caused by the body's natural metabolism can be eliminated, facilitating the analysis of the therapeutic effect of moxibustion.

[0146] The data association unit 430 is configured to acquire the probability distribution values ​​corresponding to each block 6 along each heating stop node in the heating path, generate a two-dimensional data matrix, and associate and store the above two-dimensional data matrix, physiological characteristic parameter categories, and their corresponding trend difference values ​​as training data. The correlation analysis unit 440 is configured to acquire a set amount of training data to form a training dataset, and generate the correlation relationship between the numerical distribution of elements in the two-dimensional data matrix and the trend difference values ​​based on neural network analysis or data correlation algorithm analysis. In practical applications, not all selected physiological characteristic parameters will change in response to moxibustion treatment. Therefore, in practice, the physiological characteristic parameters corresponding to the current user can be continuously optimized to reduce unnecessary data collection work.

[0147] The path optimization unit 450 is configured to adjust the element values ​​in the two-dimensional data matrix based on the required trend difference value according to the above-mentioned correlation, and then adjust the heating path in reverse according to the adjusted two-dimensional data matrix. For example, when it is necessary to increase the rate of blood sugar decrease, it is only necessary to know whether the element values ​​in the two-dimensional data matrix should be increased or decreased, and then continuously optimize the two-dimensional data matrix according to the above-mentioned increase or decrease scheme. In practical applications, the element values ​​in the above-mentioned optimized two-dimensional data matrix are probability distribution values. The dwell time of each heating pause node in the heating path can be adjusted by the above-mentioned optimized probability distribution values. In a specific embodiment, the arrangement of the heating path can be adjusted. The above technical solution can use data correlation analysis to obtain the correlation between the changing trend of physiological characteristic parameter data of a set category and the moxibustion heating path. Based on the above correlation, the two-dimensional data matrix can be optimized, that is, the heating path can be optimized. The above-mentioned heating path optimization process does not involve the basic moxibustion scheme, but is carried out for the moxibustion heating process of a single target acupoint, and the optimization scope is limited to the probability statistical region 5, which will not cause side effects. It is suitable for ordinary home users and other users who do not have professional moxibustion knowledge.

[0148] Based on the aforementioned intelligent moxibustion control system, this application also discloses an intelligent moxibustion control method based on AR acupoint navigation, such as... Figure 7 As shown, it includes:

[0149] S100: Obtain user symptom data and generate corresponding basic moxibustion plans based on the symptom-plan relationship model, determining target acupoints, moxibustion sequence, heating duration and heating temperature of each target acupoint;

[0150] S200 configures the heating path of the moxibustion head for each target acupoint and displays the output.

[0151] In a specific implementation, the above-mentioned basic moxibustion plan and the heating path information corresponding to each target acupoint can be directly output to the moxibustion execution mechanism, and the automatic moxibustion device can complete the relevant operations.

[0152] In step S100 above, obtaining user symptom data includes:

[0153] Establish a data connection between the system and an external data storage device to directly retrieve the user's symptom data from the external data storage device; and / or

[0154] Collect and recognize user voice information or gesture images to obtain user symptom data or commands; and / or

[0155] Detect user clicks on the interactive interface and convert them into input data or commands according to the interface settings.

[0156] In step S200 above, configuring the heating path of the moxibustion head for each target acupoint during moxibustion and displaying the output includes:

[0157] Displaying corresponding text or image information on the interactive interface; and / or

[0158] The distribution area of ​​the target acupoint, its heating path, efficacy description, and precautions are overlaid on real-time captured images; and / or

[0159] The system outputs moxibustion guidance information, feedback data, inquiry information, and prompts via voice or vibration.

[0160] In step S200 above, such as Figure 8 As shown, the method for generating and configuring the heating path specifically includes:

[0161] S210, acquire and store a marker recognition algorithm for identifying each type of marker from an image, a marker-acupoint positional relationship model for reflecting the relative positional relationship between each type of marker and each acupoint, a reference marker-acupoint positional relationship model for reflecting the relative positional relationship between a reference marker and associated acupoints, and an action instruction-feedback data-acupoint positional relationship model for reflecting the relationship between feedback data after the user executes a specific action instruction and the position of each acupoint;

[0162] S211, acquire a user's body image and identify markers in the image based on a marker recognition algorithm;

[0163] S212, Based on the identified reference markers and combined with the reference marker-acupoint location relationship model, determine the location area of ​​the target acupoint, and enlarge the area by a set ratio with the geometric center of the area as the control point. Then, divide the enlarged area into multiple blocks of a set shape and size to form a probability statistical area.

[0164] S213. Based on the identified type of markers, combined with the type of marker-acupoint location relationship model, determine the location area of ​​the target acupoint, mark it as a type of area, and store it in association with the name and number of the type of markers.

[0165] S214, based on the action instruction-feedback data-acupoint location relationship model, at least one action instruction is output through the interaction module and feedback data after the user executes the above action instruction is obtained. The area where the target acupoint is located is determined based on the feedback data, marked as a second-class area, and stored in association with the action instruction name number.

[0166] S215, Calculate the projected area of ​​each type I / type II region on each block in the probability statistics area. If the projected area is greater than 1 / 2 of the block area, then associate and store the block with the name number of the type I marker / action instruction, and calculate the distribution probability value corresponding to each block based on the aforementioned calculation formula.

[0167] S216. A closed-loop heating path is generated based on the probability distribution value of each block in the probability statistics region, and the geometric center of each block is used as the heating stopping node. The dwell time of the moxibustion head at each heating stopping node is set to be positively correlated with the magnitude of the probability distribution value of that block.

[0168] The type of markers mentioned in step S213 above includes human anatomical surface landmarks related to the target acupoint, body outlines, and specific reference objects used to indicate the distance and positional relationship between the target acupoint and specific markers;

[0169] The action instructions described in step S214 above include stimulating the area where the suspected target acupoint is located or stimulating the associated acupoints related to the target acupoint. The stimulation methods include one or more of the following: tapping, pressing, kneading, itching, electrical stimulation, and thermal stimulation. The feedback data includes one or more of the following: tenderness index, induration index, temperature change, tactile sensitivity index, thermal sensitivity index, and nerve reaction speed.

[0170] In this embodiment of the application, the shape of each block in the probability statistics region is set as an equilateral triangle, a square, or a regular hexagon.

[0171] In step S216, a closed-loop heating path is generated, including:

[0172] S2160, obtain the effective heat radiation area of ​​a fixed point heated by the moxibustion head during moxibustion, calculate the ratio of the effective heat radiation area to the block area, and round the ratio to obtain the block integration coefficient.

[0173] S2161, Integrate the blocks in the probability statistics region according to the above block integration coefficient, divide the probability statistics region into multiple reference regions, calculate the average value of the distribution probability value corresponding to each block in the above reference regions and store it in association with the reference regions.

[0174] S2162, a closed-loop heating path is generated by connecting the geometric center of the reference area as the heating stopping node. The arrangement of the heating path includes: spiraling inward along the edge of the probability statistics area through each heating stopping node and then returning; spiraling outward from the center of the probability statistics area through each heating stopping node and then returning; or sequentially connecting the heating paths according to the magnitude of the average distribution probability value corresponding to each reference area.

[0175] like Figure 9 As shown, the intelligent moxibustion control method of this application also includes a heating path optimization step, including:

[0176] S310, acquires physiological characteristic parameter data of a set category within a set time period before and after the user's moxibustion;

[0177] S311, Based on physiological characteristic parameter data, analyze the natural change trend of the physiological characteristic parameter data of the above-set category before moxibustion and the response change trend of the physiological characteristic parameter data of the above-set category after moxibustion, and generate the trend difference value between the response change trend and the natural change trend.

[0178] S312, along each heating stop node in the heating path, obtain the distribution probability value corresponding to each block, generate a two-dimensional data matrix, and store the above two-dimensional data matrix, physiological characteristic parameter categories and their corresponding trend difference values ​​as training data.

[0179] S313, acquire a set amount of training data to form the training dataset corresponding to the current user, and generate the correlation between the numerical distribution of elements and the trend difference value in the two-dimensional data matrix based on neural network analysis or data correlation algorithm analysis;

[0180] S314, adjust the element values ​​in the two-dimensional data matrix according to the above correlation and based on the required trend difference value, and adjust the heating path in reverse according to the adjusted two-dimensional data matrix.

[0181] One application scenario of the intelligent moxibustion control method proposed in this application is as follows:

[0182] Users input symptom data, such as blood sugar and blood pressure, into a dialog box on a dedicated app on their smartphone or tablet. The system then automatically generates a basic moxibustion plan and provides voice prompts, instructing the user to point their phone or tablet's camera at a specific body part. The system automatically captures an image of the body and identifies landmarks to determine the location of the target acupoint (Category I region). This region is then marked on the image using AR technology. Voice prompts instruct the user to press this region and inquire about their sensations afterward, further confirming the location of the target acupoint (Category II region). The system then generates a heating path corresponding to the target acupoint, which is overlaid on the captured body image using AR technology. The user can move the moxa stick along the marked heating path until the current acupoint is heated, and then move to the next target acupoint based on prompts. The entire moxibustion process does not require users to have professional acupoint location identification skills, nor does it require users to set their own moxibustion plans. It also provides real-time guidance to users during the operation process, reducing the risk of burns while improving the therapeutic effect of moxibustion. This helps to promote moxibustion treatment and related products to ordinary family users.

[0183] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. An intelligent moxibustion control system based on AR acupoint navigation, characterized in that, The moxibustion plan generation module (100) has a built-in symptom-plan relationship model and is configured to generate corresponding basic moxibustion plans based on the acquired user symptom data and determine the target acupoints. The moxibustion heating path generation module (200) is configured to generate the heating path of the moxibustion head at each target acupoint during moxibustion. The interaction module (300) is configured to acquire user symptom data, receive and output the basic moxibustion plan and heating path; The moxibustion heating path generation module (200) includes: The image acquisition and recognition unit (210) is configured to acquire images and recognize set markers in the images according to a built-in marker recognition algorithm; The statistical region generation unit (220) has a built-in benchmark marker-acupoint location relationship model, which is configured to determine the target acupoint location area based on the identified benchmark marker location, and to enlarge the area by a set ratio using the geometric center of the area as the control point, and then divide the enlarged area into multiple blocks (6) of set shape and size to form a probability statistical region (5). A probability distribution generation unit (230) has a built-in marker-acupoint location relationship model, configured to determine the location of the target acupoint based on the location of the identified marker, and mark it as a region. The second type probability distribution generation unit (240) has a built-in action instruction-feedback data-acupoint location relationship model, is configured to connect with the interaction module (300) for data connection, outputs at least one action instruction and obtains feedback data after the user executes the above action instruction, determines the area where the target acupoint is located based on the feedback data, and marks it as the second type area; The probability distribution statistics unit (250) is configured to calculate the projected area of ​​each block (6) in the probability statistics area (5) of the first-class region / second-class region. If the ratio of the projected area to the area of ​​the block (6) exceeds a set value, the block (6) is associated with the name number of the first-class marker / action instruction and stored, and the distribution probability value corresponding to each block (6) is calculated based on the set calculation formula. The path generation unit (260) generates a closed-loop heating path based on the probability distribution value of each block (6) in the probability statistics region (5) and uses the geometric center of each block (6) as the heating stopping node. The dwell time of the moxibustion head at each heating stopping node is positively correlated with the magnitude of the probability distribution value of that block (6). The basic moxibustion plan includes: target acupoints, moxibustion sequence, heating time and heating temperature for each target acupoint; The set calculation formula is: ； In the above formula, S(M,N) is the probability distribution value, where M and N represent sets containing the names and numbers of a class of markers and the names and numbers of action commands, respectively; P i P represents the probability that the area where the current block (6) is located is the target acupoint area if the current block (6) is associated with a marker of type i. k This indicates the probability that the area where the current block (6) is located is the target acupoint area if the current block (6) is associated with an action instruction with name number k; β1 and β2 are the weight values ​​for calculation.

2. The intelligent moxibustion control system based on AR acupoint navigation according to claim 1, characterized in that, The aforementioned type of markers includes anatomical surface landmarks related to the target acupoint, body contours, and specific reference objects used to indicate the distance and positional relationship between the target acupoint and a specific marker; The action instructions include stimulating the area where the suspected target acupoint is located or stimulating related acupoints associated with the target acupoint. The stimulation methods include one or more of the following: tapping, pressing, kneading, itching, electrical stimulation, and thermal stimulation; the feedback data includes one or more of the following: tenderness index, induration index, temperature change, tactile sensitivity index, thermal sensitivity index, and nerve reaction speed.

3. The intelligent moxibustion control system based on AR acupoint navigation according to claim 1, characterized in that, The shape of the block (6) is an equilateral triangle, a square, or a regular hexagon; The path generation unit (260) includes: The block integration coefficient acquisition subunit (2601) is configured to acquire the effective heat radiation area of ​​a fixed point heated by the moxibustion head during moxibustion, calculate the ratio of the effective heat radiation area to the area of ​​the block (6), and round the ratio to obtain the integration coefficient of the block (6). The block integration subunit (2602) is configured to integrate the blocks (6) in the probability statistics region (5) according to the integration coefficient of the above-mentioned blocks (6), divide them into multiple reference regions, calculate the average value of the distribution probability value corresponding to each block (6) in the above-mentioned reference regions and store them in association with the reference regions. The heating path generation subunit (2603) is configured to generate a closed-loop heating path with the geometric center of the reference area as the heating stopping node; The arrangement of the heating path includes: spiraling inward along the edge of the probability statistics region (5), spiraling outward from the center of the probability statistics region (5), or generating the heating path according to the average distribution probability value corresponding to each reference region.

4. The intelligent moxibustion control system based on AR acupoint navigation according to claim 3, characterized in that, The moxibustion heating path generation module (200) also includes: The path correction unit (270) is connected to the image acquisition and recognition unit (210) for data connection. It selects at least one marker in the image as a motion reference, acquires the motion characteristics of the probability statistics area (5) and / or the displacement of the moxibustion head itself during moxibustion, and adjusts the heating path according to the motion characteristics. The motion characteristics include the probability statistics region (5) and / or the distance and period of the reciprocating displacement of the moxibustion head in a certain direction during moxibustion. Adjusting the arrangement of the heating path according to the motion characteristics includes: adjusting the heating path in the opposite direction of the reciprocating displacement, and the adjustment time being adapted to the period of the reciprocating displacement.

5. The intelligent moxibustion control system based on AR acupoint navigation according to claim 1, characterized in that, The interaction module (300) includes a data input terminal (310) and a data output terminal (320); wherein, The data input terminal (310) includes: The voice input unit (3101) is configured to collect the user's voice information and convert it into input data or instructions; The gesture recognition unit (3102) is configured to acquire and recognize the user's gesture actions and convert them into input data or instructions based on a set gesture recognition model; The click input unit (3103) is configured to detect the user's operation on the interactive interface and convert it into input data or instructions according to the interface settings; The data retrieval unit (3104) is configured to retrieve target data stored in an external data storage device as input data; The data output terminal (320) includes: The image display unit (3201) is configured to display corresponding text or image information on the interactive interface; AR navigation unit (3202) is configured to overlay and display the distribution area of ​​the target acupoint, its heating path, efficacy description and precautions information on a real-time captured image. The voice playback unit (3203) is configured to output moxibustion guidance information and feedback data inquiry information via voice. The vibration output unit (3204) is configured to output vibrations of different frequencies to represent different prompts.

6. The intelligent moxibustion control system based on AR acupoint navigation according to claim 1, characterized in that, The intelligent moxibustion control system also includes a heating path optimization module (400), comprising: The physiological characteristic parameter acquisition unit (410) is configured to acquire physiological characteristic parameter data of a set category within a set time period before and after moxibustion. The data analysis unit (420) is configured to analyze the natural change trend of the physiological characteristic parameter data of the above-set category before moxibustion and the response change trend after moxibustion based on the physiological characteristic parameter data, and generate a trend difference value between the response change trend and the natural change trend. The data association unit (430) is configured to obtain the distribution probability value of each block (6) along each heating stop node in the heating path, generate a two-dimensional data matrix, and associate and store the above two-dimensional data matrix, physiological characteristic parameter categories and their corresponding trend difference values ​​as training data. The correlation analysis unit (440) is configured to acquire a set amount of training data to form a training dataset, and generate the correlation between the numerical distribution of elements and the trend difference values ​​in a two-dimensional data matrix based on neural network analysis or data correlation algorithm analysis. The path optimization unit (450) is configured to adjust the element values ​​in the two-dimensional data matrix based on the required trend difference value according to the above correlation, and to adjust the heating path in reverse according to the adjusted two-dimensional data matrix. The categories of physiological characteristic parameters include blood glucose, blood pressure, blood lipids, alanine aminotransferase, urine protein, hemoglobin, thyroid-stimulating hormone, and related physiological characteristic parameters that reflect the data of the above categories of physiological characteristic parameters.

7. A smart moxibustion control method based on AR acupoint navigation, characterized in that, include: Acquire user symptom data and generate corresponding basic moxibustion plans based on the symptom-plan relationship model to determine target acupoints; Configure and display the heating path of the moxibustion head for each target acupoint during moxibustion. Generating and configuring the heating path includes: Acquire and store a marker recognition algorithm for identifying each type of marker from an image, a marker-acupoint positional relationship model for reflecting the relative positional relationship between each type of marker and each acupoint, a reference marker-acupoint positional relationship model for reflecting the relative positional relationship between a reference marker and associated acupoints, and an action instruction-feedback data-acupoint positional relationship model for reflecting the relationship between feedback data after a user executes a specific action instruction and the position of each acupoint; Acquire a user's body image and identify markers in the image based on the marker recognition algorithm; Based on the identified reference markers, and combined with the reference marker-acupoint location relationship model, the location area of ​​the target acupoint is determined. The area is then magnified by a set ratio using the geometric center of the area as the control point. The magnified area is then divided into multiple blocks of a set shape and size to form a probability statistical area. Based on the identified type of markers, and combined with the type of marker-acupoint location relationship model, the location area of ​​the target acupoint is determined, marked as a type of area, and stored in association with the name and number of the type of markers. Based on the action instruction-feedback data-acupoint location relationship model, at least one action instruction is output through the interaction module and feedback data after the user executes the above action instruction is obtained. The area where the target acupoint is located is determined based on the feedback data, marked as a second-class area, and stored in association with the action instruction name number. The projected area of ​​each type I / type II region on each block in the probability statistics area is counted. If the projected area is greater than 1 / 2 of the block area, the block is associated with the name number of the type I marker / action instruction and stored. The distribution probability value corresponding to each block is calculated based on the set calculation formula. Based on the probability distribution value of each block in the probability statistics region, a closed-loop heating path is generated, with the geometric center of each block as the heating stopping node. The dwell time of the moxibustion head at each heating stopping node is positively correlated with the magnitude of the probability distribution value of that block. The basic moxibustion plan includes: target acupoints, moxibustion sequence, heating time and heating temperature for each target acupoint; The aforementioned type of markers includes anatomical surface landmarks related to the target acupoint, body contours, and specific reference objects used to indicate the distance and positional relationship between the target acupoint and a specific marker; The action instructions include stimulating the area where the suspected target acupoint is located or stimulating related acupoints associated with the target acupoint. The stimulation methods include one or more of the following: tapping, pressing, kneading, tingling, electrical stimulation, and thermal stimulation. The feedback data includes one or more of the following: tenderness index, induration index, temperature change, tactile sensitivity index, thermal sensitivity index, and nerve reaction speed. The set calculation formula is: ； In the above calculation formula, P represents the probability distribution, where M and N represent sets containing the names and numbers of a class of markers and the names and numbers of action commands, respectively. i P represents the probability that the current block is a target acupoint region if it contains a marker with name and number i. k This indicates the probability that the current block is the target acupoint area if the current block is associated with an action instruction with name and number k; β1 and β2 are the weight values ​​for calculation.

8. The intelligent moxibustion control method based on AR acupoint navigation according to claim 7, characterized in that, Obtaining user symptom data includes: Establish a data connection between the system and an external data storage device to directly retrieve the user's symptom data from the external data storage device; and / or Collect and recognize user voice information or gesture images to obtain user symptom data or commands; and / or Detect user clicks on the interactive interface and convert them into input data or commands according to the interface settings; Configure and display the heating path of the moxibustion head for each target acupoint during moxibustion, including: Displaying corresponding text or image information on the interactive interface; and / or The distribution area of ​​the target acupoint, its heating path, efficacy description, and precautions are overlaid on real-time captured images; and / or The system outputs moxibustion guidance information, feedback data, inquiry information, and prompts via voice or vibration.

9. The intelligent moxibustion control method based on AR acupoint navigation according to claim 7, characterized in that, The shape of each block in the probability statistics region is set as an equilateral triangle, a square, or a regular hexagon; Generating the closed-loop heating path includes: The effective heat radiation area of ​​a fixed point heated by the moxibustion head during moxibustion is obtained. The ratio of the effective heat radiation area to the area of ​​the block is calculated, and the block integration coefficient is obtained by rounding the ratio. Based on the above block integration coefficient, the blocks in the probability statistics region are integrated, the probability statistics region is divided into multiple reference regions, the average value of the distribution probability value corresponding to each block in the above reference regions is calculated and stored in association with the reference regions. A closed-loop heating path is generated by connecting the geometric center of the reference area as the heating stopping node. The arrangement of the heating path includes: spiraling inward along the edge of the probability statistics region through each heating stopping node and then returning; spiraling outward from the center of the probability statistics region through each heating stopping node and then returning; or sequentially connecting the heating paths according to the magnitude of the average distribution probability value corresponding to each reference region.

10. The intelligent moxibustion control method based on AR acupoint navigation according to claim 7, characterized in that, The intelligent moxibustion control method also includes a heating path optimization step, including: Obtain physiological characteristic parameter data of a set category within a set time period before and after moxibustion for the user; Based on the physiological characteristic parameter data, the natural change trend of the physiological characteristic parameter data of the above-mentioned set category before moxibustion and the response change trend of the physiological characteristic parameter data of the above-mentioned set category after moxibustion are analyzed, and the trend difference value between the response change trend and the natural change trend is generated. Along each heating stop node in the heating path, the distribution probability value corresponding to each block is obtained, and a one- or two-dimensional data matrix is ​​generated. The above two-dimensional data matrix, physiological characteristic parameter categories and their corresponding trend difference values ​​are associated and stored as training data. A set amount of training data is acquired to form the training dataset corresponding to the current user. Based on neural network analysis or data correlation algorithm analysis, the correlation between the distribution of element values ​​in a two-dimensional data matrix and the trend difference value is generated. Based on the above correlation and the required trend difference value, adjust the element values ​​in the two-dimensional data matrix, and then adjust the heating path in reverse according to the adjusted two-dimensional data matrix.

Images