Intelligent medicine cabinet information interaction method and system based on multimode communication adaptive decision

Abstract

The invention discloses an intelligent medicine cabinet information interaction method and system based on multi-mode communication adaptive decision, and relates to the field of intelligent medical treatment, and the method comprises a collection module which is used for collecting the output medicine correlation information in the medicine output process of an intelligent medicine cabinet; the extraction module is used for traversing the drug associated information and extracting drug feature information from the drug associated information; the interaction module is used for interacting with the intelligent medicine cabinet and obtaining the medicine attributes and parameters of the medicine dispensing task currently executed by the intelligent medicine cabinet; through multi-dimensional information collection and accurate verification and in combination with a dynamically adjusted credibility judgment standard, the drug delivery error rate is greatly reduced, the drug use risk is effectively avoided, the drug use safety of a drug taking user is guaranteed, and the management and control requirements are flexibly adapted for drugs of different risk levels.

Description

Technical Field

[0001] This invention relates to the field of smart healthcare technology, specifically to an intelligent medicine cabinet information interaction method and system based on multi-mode communication adaptive decision-making. Background Technology

[0002] The smart medicine cabinet integrates medicine storage, barcode scanning and identification, and inventory monitoring functions. It supports expiration warnings and usage record traceability. Through intelligent interaction, it simplifies the medicine management process and is suitable for scenarios such as homes and community clinics. It effectively realizes the orderly storage and safe management of medicines and improves the convenience and standardization of medication use.

[0003] However, based on feedback from existing smart medicine cabinet users, the existing smart medicine cabinets do not have a well-designed final verification process before dispensing medicine. As a result, when the internal mechanical mechanism of the smart medicine cabinet malfunctions, the type and quantity of medicine dispensed may be incorrect, causing medication risks to users.

[0004] To address this, we propose an intelligent medicine cabinet information interaction method and system based on multi-mode communication adaptive decision-making. Summary of the Invention

[0005] In view of the above-mentioned shortcomings of the existing technology, the present invention provides an intelligent medicine cabinet information interaction method and system based on multi-mode communication adaptive decision-making, which can effectively solve the problems of the existing technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions; This invention discloses an intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making, comprising: The system comprises the following modules: a data acquisition module for collecting drug-related information during the dispensing process of the smart medicine cabinet; an extraction module for traversing the drug-related information and extracting drug feature information; an interaction module for interacting with the smart medicine cabinet to obtain the drug attributes and parameters currently being dispensed; a verification module for obtaining the output results from the extraction and interaction modules to verify the reliability of the current dispensing; a control module for receiving the reliability result from the verification module, comparing it with a preset threshold, and deciding whether to open the dispensing compartment door; and a feedback module for broadcasting a preset voice prompt to alert the user that the dispensing process has failed if the control module's response is either yes or no. The acquisition module is connected to the extraction module via a wireless network. The extraction module is connected to the interaction module via a wireless network. The interaction module is connected to the verification module via a wireless network. The verification module is connected to the control module and the feedback module via a wireless network. The drug-related information includes drug images, weight, and specifications; drug feature information consists of icons and text information from the drug images; drug attributes include drug name; and drug parameters include drug quantity and weight. During the operation of the feedback module, the source of the incorrect dispensing is traced based on the drug-related information, drug attributes, and parameters.

[0007] Furthermore, the acquisition module integrates a multi-angle high-definition camera, a weighing sensor, and a parameter adaptation unit. The acquisition module executes the process in the following sequence: image acquisition, weight measurement, and parameter correlation. A multi-angle high-definition camera collects image data of different sides of the medicine at preset angles, and the image resolution and acquisition frame rate are dynamically adjusted based on the characteristic size of the medicine. The weighing sensor's measurement accuracy meets the weight identification requirements for the smallest specifications of pharmaceuticals; The parameter adaptation unit performs a preliminary association and matching between the image data, weight data and the specification parameters in the pre-stored drug database; After collecting data, the data acquisition module synchronously performs a data credibility verification: ; In the formula: To assess the overall reliability of the collected data; These are the weighting coefficients; The image sharpness compliance rate; This is the angular deviation attenuation coefficient; This represents the average deviation between the acquisition angle and the optimal acquisition angle. This represents the actual measured weight from the weighing sensor. The standard weight of the corresponding drug is stored in the pre-stored database; like If the data is less than the preset data collection reliability threshold, the data collection process will be re-executed until the threshold is met, or a data collection anomaly warning will be triggered when the number of re-collections exceeds the preset number.

[0008] Furthermore, the feature extraction process of the extraction module is as follows: Icon extraction uses a feature point detection algorithm to extract the geometric parameters of the icon outline in the drug image and performs similarity matching with a pre-stored icon feature library; in the text information extraction stage, OCR recognition is combined with contextual semantic verification to remove invalid text caused by recognition errors. Evaluation of the effectiveness of simultaneous feature extraction: ; In the formula: This is the effectiveness index for feature extraction; Icon feature matching degree; For text information matching degree; This represents the cumulative number of errors during the feature extraction process. , These are the feature weight coefficients; This is the error penalty coefficient; like If the image resolution is less than the preset extraction validity threshold, the image frame rate and the number of acquisition angles are increased, and the extraction process is re-executed until... The value is greater than the preset extraction validity threshold.

[0009] Furthermore, the interaction types between the interaction module and the smart medicine cabinet include short-range communication links and long-range communication links, and the optimal link is dynamically selected based on adaptive decision-making according to communication quality, where communication quality is: ; In the formula: A comprehensive score for communication quality; This is the normalized value of the communication signal strength. For data transmission success rate; This is the normalized value for data transmission delay; , , To evaluate the weighting coefficients; The interactive module calculates the real-time performance of each communication link. ,choose Data exchange is conducted on links that have a maximum quality threshold that is greater than the preset communication quality threshold.

[0010] Furthermore, the credibility verification in the verification module follows the following rules: ; In the formula: To assess the overall credibility of drug delivery; To assess the credibility of image feature matching; For weight consistency reliability; To match the credibility of the quantity; , For dimension weights; To assess the overall reliability of the collected data.

[0011] Furthermore, during the verification module's runtime phase, first... , , The three-dimensional independent credibility verification is used. If the credibility of any dimension is less than the preset threshold of the corresponding dimension, the drug is directly judged to be unreliable. If all individual dimension credibility scores meet the standard, then calculate the overall credibility score. ; in, During the calculation process, , , , The values ​​of are all controlled within the range of (0,1) through normalization.

[0012] Furthermore, the preset threshold of the control module is a dynamic calibration threshold: ; In the formula: The confidence threshold after dynamic calibration; Basic threshold; This is a risk adjustment factor; This refers to the risk level coefficient of a drug. in, Preset by the system user, The information is obtained based on a pre-set drug risk level coefficient lookup table.

[0013] Furthermore, during the feedback module's operation phase, a one-to-one correspondence is established between the association information of all drugs in the erroneous dispensing batches obtained by the collection module and the attributes and parameters of the drugs that should have been dispensed obtained by the interaction module. Then, the degree of deviation between the actual dispensing and the expected dispensing of each drug is quantified through the error association index. ; In the formula: This is the erroneous correlation index; , , These are the feature weight coefficients; The similarity in characteristics between the actual dispensing of medicine and the medicine to be dispensing; This refers to the actual measured weight of the dispensed medication. This refers to the standard weight of the medicine. The degree of matching between the specifications of the actual drugs dispensed and the drugs to be dispensed; During the tracing process, all actual drug deliveries will be... Compared with the preset error judgment threshold, if a certain actual drug is dispensed... If the value is greater than or equal to this threshold, it is directly identified as a drug from an incorrect source. If there is more than one actual drug produced If the standard is exceeded, then proceed as follows: The values ​​are sorted from high to low, and drugs with high deviation are marked first.

[0014] On the other hand, an intelligent medicine cabinet information interaction method based on multi-mode communication adaptive decision-making includes: The system collects drug association information during the dispensing process of the smart medicine cabinet, calculates the overall credibility of the collected data, and re-collects or triggers an anomaly warning if the data does not meet the standard. It extracts icon geometric parameters and text information from the collected drug association information, optimizes feature data by combining similarity matching and semantic verification, and evaluates the data through a validity index. If the data does not meet the standard, the collection conditions are adjusted and the data is re-extracted. The system interacts with the smart medicine cabinet through a multi-mode communication link, dynamically selecting the optimal link or initiating link aggregation based on the comprehensive communication quality score, and encrypts and obtains the drug attributes and parameters corresponding to the dispensing task. It performs single-dimensional credibility verification on image feature matching, weight consistency, and quantity matching. After all dimensions meet the standard, it calculates the overall credibility of the dispensing based on the credibility of the collected data. It dynamically calibrates the credibility judgment threshold based on the drug risk level, compares the overall credibility of the dispensing with this threshold, and controls the opening of the smart medicine cabinet's dispensing compartment door if the value is met. If the dispensing is unreliable, it locks the source drug through an error association index and broadcasts a voice prompt.

[0015] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects: This invention significantly reduces medication error rates and effectively mitigates medication risks by combining multi-dimensional information collection and precise verification with dynamically adjusted credibility judgment standards. It ensures medication safety for users, flexibly adapts to control requirements for drugs with different risk levels, and is compatible with complex data collection environments and differentiated drug characteristics, improving system adaptability and stability. Furthermore, it ensures efficient and stable data interaction by adaptively selecting the optimal communication link or aggregating transmission, reducing transmission delays and interruptions. In case of medication errors, it can quickly trace and locate the problematic drug, accurately broadcast error information, and synchronize it to the management terminal for convenient and timely review and processing. This improves the user's medication experience, reduces management and maintenance costs, and balances safety and convenience, adapting to the stringent requirements of different scenarios such as prescription drugs and drugs with limited supply. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0017] Figure 1 This is a schematic diagram of the structure of an intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making; Figure 2 This is a flowchart illustrating an intelligent medicine cabinet information interaction method based on multi-mode communication adaptive decision-making. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] The present invention will be further described below with reference to embodiments.

[0020] Example 1: This embodiment presents an intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making, such as... Figure 1 As shown, it includes: The data acquisition module is used to collect drug-related information output during the dispensing process of the smart medicine cabinet; The data acquisition module integrates a multi-angle high-definition camera, a weighing sensor, and a parameter adaptation unit. The acquisition module executes the process in the following sequence: image acquisition, weight measurement, and parameter correlation. A multi-angle high-definition camera collects image data of different sides of the medicine at preset angles, and the image resolution and acquisition frame rate are dynamically adjusted based on the characteristic size of the medicine. The weighing sensor's measurement accuracy meets the weight identification requirements for the smallest specifications of pharmaceuticals; The parameter adaptation unit performs a preliminary association and matching between the image data, weight data and the specification parameters in the pre-stored drug database; After collecting data, the data acquisition module synchronously performs a data credibility verification: ; In the formula: To assess the overall reliability of the collected data; These are the weighting coefficients; The image sharpness compliance rate; This is the angular deviation attenuation coefficient; This represents the average deviation between the acquisition angle and the optimal acquisition angle. This represents the actual measured weight from the weighing sensor. The standard weight of the corresponding drug is stored in the pre-stored database; The above formula takes into account the image acquisition clarity, the deviation between the acquisition angle and the optimal angle, and the difference between the measured weight and the standard weight. By reasonably allocating weights, it quantifies the overall reliability of the acquired data. At the same time, it sets up a closed-loop mechanism. If the acquisition reliability does not meet the standard, it will be re-acquired. After multiple failures, an anomaly warning will be triggered to ensure the reliability of subsequent data processing. like If the data is less than the preset data collection confidence threshold, the data collection process will be re-executed until the target is met, or a data collection anomaly warning will be triggered when the number of re-collections exceeds the preset number. in, The values ​​of all values ​​are in the range (0,1), and The sum is 1, and ; The extraction module is used to traverse drug association information and extract drug feature information from the drug association information; The feature extraction process of the extraction module is as follows: Icon extraction uses a feature point detection algorithm to extract the geometric parameters of the icon outline in the drug image, including the outline area, number of corners, and distance between feature points, and performs similarity matching with a pre-stored icon feature library; in the text information extraction stage, based on OCR recognition combined with contextual semantic verification, invalid text caused by recognition errors is removed, including typos and garbled characters in the extracted drug name; Evaluation of the effectiveness of simultaneous feature extraction: ; In the formula: This is the effectiveness index for feature extraction; Icon feature matching degree; For text information matching degree; This represents the cumulative number of errors during the feature extraction process. , These are the feature weight coefficients; This is the error penalty coefficient; The above formula combines the two core extraction results of icon feature matching degree and text information matching degree, introduces an error penalty coefficient to control the cumulative error in the extraction process, highlights the importance of the two types of features through weight allocation, and establishes a dynamic optimization mechanism. When the extraction effectiveness does not meet the standard, the acquisition conditions are optimized by increasing image resolution, increasing acquisition frame rate and number of angles, thereby ensuring the accuracy and effectiveness of feature extraction. like If the image resolution is less than the preset extraction validity threshold, the image frame rate and the number of acquisition angles are increased, and the extraction process is re-executed until... The value exceeds the preset extraction validity threshold; in, , The values ​​of are all in the range of (0,1), and , The sum is 1. The value range is preset to [1,5]. Its value is positively correlated with the degree of influence of text errors on drug feature matching and drug dispensing verification results. That is, the greater the interference of key text information errors such as drug name and specifications on recognition and verification, the larger the value; the smaller the value, the less decorative or auxiliary non-key text errors have a substantial impact on recognition and verification. The interaction module is used to interact with the smart medicine cabinet and obtain the drug attributes and parameters of the medicine cabinet currently performing the dispensing task. The interaction modules and the smart medicine cabinet interact through short-range communication links, including Bluetooth and NFC, and long-range communication links, including Wi-Fi and LoRa. The optimal link is dynamically selected based on adaptive decision-making according to communication quality. The communication quality is defined as follows: ; In the formula: A comprehensive score for communication quality; This is the normalized value of the communication signal strength. For data transmission success rate; This is the normalized value for data transmission delay; , , To evaluate the weighting coefficients; The above formula evaluates communication quality from three key dimensions: communication signal strength, data transmission success rate, and transmission delay. It adapts the priority of communication stability, transmission efficiency, and real-time requirements under different scenarios through weight calibration, dynamically selects the communication link with the best comprehensive score and meets the standard, and starts the link aggregation mode to split the data for parallel transmission if all links fail to meet the standard. At the same time, it uses an encrypted transmission protocol to ensure the security of the transmission process. The interactive module calculates the real-time performance of each communication link. ,choose Data exchange is conducted only on links whose maximum quality exceeds a preset communication quality threshold; otherwise, all links... If all values ​​are less than the threshold, the link aggregation mode is activated, and the data is split and transmitted in parallel through multiple links to ensure the stability of the interaction; during the interaction, an encrypted transmission protocol is used to encrypt the drug attribute and parameter data. in, It is derived from the signal received power; Calculated by the ratio of the amount of data already transmitted to the total amount of data; It is calculated from the ratio of actual delay to maximum allowable delay; , , The value range is (0,1], and is determined based on the priority requirements of communication stability, transmission efficiency, and real-time performance. The verification module is used to obtain the output results of the extraction module and the interaction module to verify the credibility of the current smart medicine cabinet dispensing medicine. In the verification module, the credibility verification follows the following rules: ; In the formula: To assess the overall credibility of drug delivery; To assess the credibility of image feature matching; For weight consistency reliability; To match the credibility of the quantity; , For dimension weights; To assess the overall reliability of the collected data; The above formula integrates four dimensions: data collection credibility, image feature matching credibility, weight consistency credibility, and quantity matching credibility. The weight of each dimension is dynamically adjusted according to the complexity of the collection environment, the level of drug appearance differentiation, weight specification differentiation, and the strictness of quantity control. First, the credibility of each dimension is strictly controlled. If any dimension fails to meet the standard, the drug is judged to be unreliable. After all individual dimensions meet the standard, the overall credibility is calculated, and finally, a comprehensive and accurate drug delivery credibility verification is achieved. in, The preset value range is [0.25, 0.4]. When the collection environment is complex (such as unstable light, vibration interference) and the characteristics of the drug are easily affected by the collection quality, the value is larger, and vice versa. The value range is [0.2, 0.35]. The value is larger when the appearance of the drug is significantly different (such as unique icons or exclusive text labels) and the image features have a higher degree of distinguishability for drug identification, and vice versa. The preset value range is [0.2, 0.35]. When the drug weight specification differentiation is high (such as significant differences in weight between different drugs / specifications) and the weight is less affected by environmental factors (humidity, packaging deformation), the value is larger, and vice versa. The preset value range is [0.15, 0.25]. When the quantity of a drug to be dispensed at one time is strictly required (such as prescription drugs or drugs with limited supply) and the error in quantity will lead to serious medication risks, the value is larger, and vice versa. During the verification module's runtime phase, first , , The three-dimensional independent credibility verification is used. If the credibility of any dimension is less than the preset threshold of the corresponding dimension, the drug is directly judged to be unreliable. If all individual dimension credibility scores meet the standard, then calculate the overall credibility score. ; in, During the calculation process, , , , The values ​​of all are controlled within the range of (0,1) through normalization. The control module receives the credibility result output by the verification module, and decides whether to open the medicine dispensing compartment door of the smart medicine cabinet based on the comparison of the preset threshold and the credibility result. The preset threshold of the control module is the dynamic calibration threshold: ; In the formula: The confidence threshold after dynamic calibration; Basic threshold; This is a risk adjustment factor; This refers to the risk level coefficient of a drug. The above formula uses the system's preset basic threshold as a benchmark, and dynamically calibrates the credibility judgment threshold by combining the drug risk level coefficient and the risk adjustment coefficient. The risk adjustment coefficient increases as the drug use risk level increases and the safety control requirements become stricter, so that the judgment standard can be adapted to the safety control needs of drugs with different risk levels. High-risk drugs correspond to higher judgment thresholds to further reduce the risk of medication. in, Preset by the system user, Retrieved based on a pre-set drug risk level coefficient lookup table; The preset value range is [0.1, 0.5]. Its value increases as the risk level of drug use increases and the safety control requirements become stricter, and decreases as the risk level decreases. The feedback module is used to play a preset voice message when the control module's control result is yes or no, to remind the user that there is an error in the current medication dispensing process. During the feedback module's operation phase, a one-to-one correspondence is established between the association information of all drugs in the erroneous dispensing batches obtained by the data collection module and the attributes and parameters of the drugs that should have been dispensed obtained by the interaction module. Then, the degree of deviation between the actual dispensing and the expected dispensing of each drug is quantified through the error association index. ; In the formula: This is the erroneous correlation index; , , These are the feature weight coefficients; The similarity in characteristics between the actual dispensing of medicine and the medicine to be dispensing; This refers to the actual measured weight of the dispensed medication. This refers to the standard weight of the medicine. The degree of matching between the specifications of the actual drugs dispensed and the drugs to be dispensed; The above formula quantifies the degree of deviation between the actual and the expected drugs from three dimensions: similarity of characteristics, degree of weight deviation, and degree of specification matching. It establishes a drug locking mechanism for the source of errors. If the deviation index of a single drug meets the standard, it is directly locked. If the deviation index of multiple drugs exceeds the standard, the drugs with high deviation are marked first according to the degree of deviation. At the same time, relevant characteristic information, deviation dimensions and traceability results are recorded simultaneously. Targeted voice broadcasts are made and the information is synchronized to the management terminal to provide a basis for error review and processing. During the tracing process, all actual drug deliveries will be... Compared with the preset error judgment threshold, if a certain actual drug is dispensed... If the value is greater than or equal to this threshold, it is directly identified as a drug from an incorrect source. If there is more than one actual drug produced If the standard is exceeded, then proceed as follows: The values ​​are sorted from highest to lowest, and drugs with high deviations are marked first.

[0021] After identifying the source of the erroneous drug, the system synchronously records its characteristic information, deviation dimension, and traceability results. The voice broadcast content specifically prompts the key information of the erroneous drug that is inconsistent, such as "Drug dispensing error: The actual drug dispensed is XX (drug name), which is different from the XX (drug name that should be dispensed) you need. Please check or contact staff." At the same time, the erroneous drug information is synchronized to the management terminal for subsequent review and processing. in, , , The values ​​of all values ​​are in the range (0,1), and , , The sum is 1. It is calculated by weighting the icon outline matching degree and the text information overlap degree in the image. The dimensions and volume of the drug are comprehensively determined, and ∈ (0,1], ∈ (0,1]; Among them, drug-related information includes drug images, weight, and specifications; drug feature information is icons and text information derived from drug images; drug attributes include drug name; and drug parameters include drug quantity and weight. During the operation of the feedback module, the source of the incorrect dispensing is traced based on the drug association information, drug attributes, and parameters. The acquisition module is connected to the extraction module via a wireless network. The extraction module is connected to the interaction module via a wireless network. The interaction module is connected to the verification module via a wireless network. The verification module is connected to the control module and the feedback module via a wireless network.

[0022] In this embodiment, the acquisition module collects drug-related information during the dispensing process of the smart medicine cabinet. The extraction module then runs to traverse the drug-related information and extracts drug feature information from it. The interaction module interacts synchronously with the smart medicine cabinet to obtain the drug attributes and parameters currently being dispensed. The verification module further obtains the output results from the extraction and interaction modules to verify the credibility of the current dispensing. The control module receives the credibility result from the verification module, compares it with a preset threshold, and decides whether to open the dispensing compartment door of the smart medicine cabinet. Finally, the feedback module, when the control module determines whether the result is yes or no, broadcasts a preset voice message to alert the user that there is a dispensing error and traces the source of the error.

[0023] In the above embodiments, when the system is configured in a smart medicine cabinet, it can accurately verify the relevant information of the medicine dispensing, adapt to different communication environments to ensure stable data transmission, dynamically adjust the judgment criteria according to the risk of the medicine, effectively reduce the error of dispensing medicine, and at the same time, it can quickly locate the problematic medicine and promptly remind the user of key discrepancies, which not only improves the safety of medication use, but also facilitates subsequent management and review, making the medicine retrieval process more reliable and efficient.

[0024] Example 2: At the implementation level, based on Example 1, this example refers to... Figure 2 A further detailed description of the intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making in Example 1 is provided below: A smart medicine cabinet information interaction method based on multi-mode communication adaptive decision-making includes: Collect drug-related information during the dispensing process of the smart medicine cabinet, calculate the overall reliability of the collected data, and if the data does not meet the standard, re-collect the data or trigger an anomaly warning. The geometric parameters and text information of icons are extracted from the collected drug association information. The feature data is optimized by combining similarity matching and semantic verification. The effectiveness index is evaluated. If the standard is not met, the collection conditions are adjusted and the data is extracted again. Interact with the smart medicine cabinet through a multi-mode communication link, dynamically select the optimal link or start link aggregation based on the comprehensive communication quality score, and encrypt and obtain the drug attributes and parameters corresponding to the dispensing task. The credibility of the drug was verified by one dimension, namely image feature matching, weight consistency and quantity matching. After all dimensions met the standards, the overall credibility of the drug was calculated by combining the credibility of the collected data. Based on the dynamic calibration of the credibility judgment threshold of the drug risk level, the overall credibility of the drug dispensing is compared with the threshold. If the threshold is met, the opening of the smart medicine cabinet dispensing door is controlled. If the medication is unreliable, the source drug will be identified through the error association index, and a voice prompt will be broadcast.

[0025] Application Example 3: Ms. Li, a resident of the community, went to the community's smart medicine cabinet to pick up her prescription for "XX Blood Sugar Lowering Capsules" using her electronic prescription. The system then initiated the information interaction process. The data acquisition module captured images of different sides of the medicine using multi-angle high-definition cameras, and the weighing sensor measured the total weight of the medicine. The parameter adaptation unit correlated and matched the acquired images and weight data with the pre-stored medicine database. Finally, the overall reliability of the acquired data was calculated to be 0.91, which is higher than the preset threshold of 0.85, indicating that the data acquisition process was effectively completed.

[0026] The extraction module performs feature extraction on the acquired drug images: it extracts the geometric parameters of the icon outline through a feature point detection algorithm, and matches them with the pre-stored icon feature library to obtain a similarity of 0.94; with the help of OCR recognition combined with semantic verification, it accurately extracts the text information of "XX hypoglycemic capsules 0.5g / capsule", with a text matching degree of 0.96 and a feature extraction effectiveness index of 0.95, which meets the preset threshold requirements and the extraction result is effective.

[0027] The interaction module simultaneously detects four communication links: Bluetooth, NFC, Wi-Fi, and LoRa, and calculates the comprehensive communication quality score for each link. The Wi-Fi link, with a score of 0.92, is the best link and is higher than the preset communication quality threshold of 0.8. The system obtains the medication dispensing task information of the smart medicine cabinet through the Wi-Fi encrypted link and confirms that the medicine to be dispensed is "XX hypoglycemic capsules", with a quantity of 14 capsules and a standard total weight of 7g.

[0028] The verification module first performs single-dimensional credibility verification: image feature matching credibility is 0.93, weight consistency credibility is 0.91, and quantity matching credibility is 0.94, all of which are higher than the preset threshold of 0.8 for the corresponding dimension. Then, combined with the collected data credibility of 0.91, a weighted calculation yields a comprehensive credibility of 0.92 for dispensing the medication. Because the medication is a prescription drug with high safety control requirements, the credibility judgment threshold after dynamic calibration by the control module is 0.89. The comprehensive credibility of 0.92 is higher than this threshold, so the system decides to open the dispensing compartment door of the smart medicine cabinet, and Ms. Li successfully retrieves her medication.

[0029] If the system collects the actual medication dispensed as "YY Blood Sugar Lowering Capsules," 12 capsules in total, weighing 5.8g, and the feature extraction effectiveness index meets the standard after secondary optimization of the collection parameters by the extraction module, and the interaction module obtains the required medication information, the verification module detects an image feature matching confidence level of 0.71 (below the threshold of 0.8), directly determining that the medication dispensing is unreliable. The control module does not initiate the door unlocking, and the feedback module calculates an error correlation index of 0.83 (above the preset threshold of 0.7), locking the source of the error as "YY Blood Sugar Lowering Capsules." A simultaneous voice announcement is made: "Dispensing Error: The actual medication dispensed is YY Blood Sugar Lowering Capsules, which does not match your required XX Blood Sugar Lowering Capsules. Please verify or contact staff." The error information is then uploaded to the management terminal for subsequent on-site handling by staff.

[0030] In summary, the systems and methods described above, through multi-dimensional information collection and precise verification, combined with dynamically adjusted credibility judgment standards, significantly reduce the medication error rate, effectively mitigate medication risks, and ensure medication safety for users. They flexibly adapt to control requirements for drugs with different risk levels, accommodate complex data collection environments and differentiated drug characteristics, improving system adaptability and stability. Furthermore, by adaptively selecting the optimal communication link or aggregating transmission, they ensure efficient and stable data interaction, reducing transmission delays and interruptions. In the event of a medication error, they can quickly trace and locate the problematic drug, accurately broadcast error information, and synchronize it to the management terminal for convenient and timely review and processing. This improves the user's medication experience, reduces management and maintenance costs, and balances safety and convenience, adapting to the stringent requirements of different scenarios such as prescription drugs and drugs with limited supply.

[0031] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making, characterized in that, include: The data acquisition module is used to collect drug-related information output during the dispensing process of the smart medicine cabinet; The extraction module is used to traverse drug association information and extract drug feature information from the drug association information; The interaction module is used to interact with the smart medicine cabinet and obtain the drug attributes and parameters of the medicine cabinet currently performing the dispensing task. The verification module is used to obtain the output results of the extraction module and the interaction module to verify the credibility of the current smart medicine cabinet dispensing medicine. The control module receives the credibility result output by the verification module, and decides whether to open the medicine dispensing compartment door of the smart medicine cabinet based on the comparison of the preset threshold and the credibility result. The feedback module is used to play a preset voice message when the control module's control result is yes or no, to remind the user that there is an error in the current medication dispensing process. Among them, drug-related information includes drug images, weight, and specifications; drug feature information is icons and text information derived from drug images; drug attributes include drug name; and drug parameters include drug quantity and weight. During the operation phase of the feedback module, the source of the incorrect dispensing drug is traced based on the drug association information, drug attributes, and parameters.

2. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 1, characterized in that, The acquisition module integrates a multi-angle high-definition camera, a weighing sensor, and a parameter adaptation unit. The acquisition module executes the process in the following sequence: image acquisition, weight measurement, and parameter association. A multi-angle high-definition camera collects image data of different sides of the medicine at preset angles, and the image resolution and acquisition frame rate are dynamically adjusted based on the characteristic size of the medicine. The weighing sensor's measurement accuracy meets the weight identification requirements for the smallest specifications of pharmaceuticals; The parameter adaptation unit performs a preliminary association and matching between the image data, weight data and the specification parameters in the pre-stored drug database; After collecting data, the data acquisition module simultaneously performs a data credibility verification: ； In the formula: To assess the overall reliability of the collected data; These are the weighting coefficients; The image sharpness compliance rate; This is the angular deviation attenuation coefficient; This represents the average deviation between the acquisition angle and the optimal acquisition angle. This represents the actual measured weight from the weighing sensor. The standard weight of the corresponding drug is stored in the pre-stored database; like If the data is less than the preset data collection reliability threshold, the data collection process will be re-executed until the threshold is met, or a data collection anomaly warning will be triggered when the number of re-collections exceeds the preset number.

3. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 1, characterized in that, The feature extraction process of the extraction module is as follows: Icon extraction uses a feature point detection algorithm to extract the geometric parameters of the icon outline in the drug image and performs similarity matching with a pre-stored icon feature library; in the text information extraction stage, OCR recognition is combined with contextual semantic verification to remove invalid text caused by recognition errors. Evaluation of the effectiveness of simultaneous feature extraction: ； In the formula: This is the effectiveness index for feature extraction; Icon feature matching degree; For text information matching degree; This represents the cumulative number of errors during the feature extraction process. , These are the feature weight coefficients; This is the error penalty coefficient; like If the image resolution is less than the preset extraction validity threshold, the image frame rate and the number of acquisition angles are increased, and the extraction process is re-executed until... The value is greater than the preset extraction validity threshold.

4. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 1, characterized in that, The interaction module interacts with the smart medicine cabinet through both short-range and long-range communication links, and dynamically selects the optimal link based on adaptive decision-making according to communication quality. The communication quality is defined as follows: ； In the formula: A comprehensive score for communication quality; This is the normalized value of the communication signal strength. For data transmission success rate; This is the normalized value for data transmission delay; , , To evaluate the weighting coefficients; The interactive module calculates the information of each communication link in real time. ,choose Data exchange is conducted on links that have a maximum quality threshold that is greater than the preset communication quality threshold.

5. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 1, characterized in that, The credibility verification in the verification module follows the following rules: ； In the formula: To assess the overall credibility of drug delivery; To assess the credibility of image feature matching; For weight consistency reliability; To match the credibility of the quantity; , For dimension weights; To assess the overall reliability of the collected data.

6. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 5, characterized in that, During the operation phase of the verification module, first , , The three-dimensional independent credibility verification is used. If the credibility of any dimension is less than the preset threshold of the corresponding dimension, the drug is directly judged to be unreliable. If all individual dimension credibility scores meet the standard, then calculate the overall credibility score. ; in, During the calculation process, , , , The values ​​of are all controlled within the range of (0,1) through normalization.

7. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 1, characterized in that, The preset threshold of the control module is a dynamic calibration threshold: ； In the formula: The confidence threshold after dynamic calibration; Basic threshold; This is a risk adjustment factor; This refers to the risk level coefficient of a drug. in, Preset by the system user, The information is obtained based on a pre-set drug risk level coefficient lookup table.

8. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 1, characterized in that, During the operation phase of the feedback module, a one-to-one correspondence is established between the association information of all drugs in the erroneous dispensing batch obtained by the acquisition module and the attributes and parameters of the drugs that should be dispensed obtained by the interaction module. Then, the degree of deviation between the actual dispensing and the drugs that should be dispensed is quantified by the error association index. ； In the formula: This is the erroneous correlation index; , , These are the feature weight coefficients; The similarity in characteristics between the actual dispensing of medicine and the medicine to be dispensing; This refers to the actual measured weight of the dispensed medication. This refers to the standard weight of the medicine. The degree of matching between the specifications of the actual drugs dispensed and the drugs to be dispensed; During the tracing process, all actual drug deliveries will be... Compared with the preset error judgment threshold, if a certain actual drug is dispensed... If the value is greater than or equal to this threshold, it is directly identified as a drug from an incorrect source. If there is more than one actual drug produced If the standard is exceeded, then proceed as follows: The values ​​are sorted from high to low, and drugs with high deviation are marked first.

9. The intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making according to claim 1, characterized in that, The acquisition module is interconnected with the extraction module via a wireless network. The extraction module is interconnected with the interaction module via a wireless network. The interaction module is interconnected with the verification module via a wireless network. The verification module is interconnected with the control module and the feedback module via a wireless network.

10. A method for information interaction in an intelligent medicine cabinet based on multi-mode communication adaptive decision-making, wherein the method is an implementation method of an intelligent medicine cabinet information interaction system based on multi-mode communication adaptive decision-making as described in any one of claims 1-9, characterized in that, include: Collect drug-related information during the dispensing process of the smart medicine cabinet, calculate the overall reliability of the collected data, and if the data does not meet the standard, re-collect the data or trigger an anomaly warning. The geometric parameters and text information of icons are extracted from the collected drug association information. The feature data is optimized by combining similarity matching and semantic verification. The effectiveness index is evaluated. If the standard is not met, the collection conditions are adjusted and the data is extracted again. Interact with the smart medicine cabinet through a multi-mode communication link, dynamically select the optimal link or start link aggregation based on the comprehensive communication quality score, and encrypt and obtain the drug attributes and parameters corresponding to the dispensing task. The credibility of the drug was verified by one dimension, namely image feature matching, weight consistency and quantity matching. After all dimensions met the standards, the overall credibility of the drug was calculated by combining the credibility of the collected data. Based on the dynamic calibration of the credibility judgment threshold of the drug risk level, the overall credibility of the drug dispensing is compared with the threshold. If the threshold is met, the opening of the smart medicine cabinet dispensing door is controlled. If the medication is unreliable, the source medication will be identified through an error correlation index, and a voice prompt will be broadcast.

Images