Hospital infection monitoring method and system

Abstract

The invention relates to the technical field of intelligent medical treatment, particularly provides a hospital infection monitoring method and system, and aims to solve the technical problems that at present, effective monitoring of standardized diagnosis and treatment operation and hospital infection control steps and measures is lacked, and hospital infection monitoring is generally low in digitization degree. In order to achieve the purpose, the method comprises the steps that hand health monitoring data of medical staff and / or disinfection monitoring data of medical equipment are / is acquired; processing the hand health monitoring data and / or disinfection monitoring data to obtain effective monitoring data; analyzing the effective monitoring data to obtain an analysis result; and fusing data in the analysis result to obtain a hospital infection monitoring result. Through the implementation mode, real-time monitoring and management of hand hygiene of medical staff and disinfection conditions of medical equipment can be achieved, the intelligence and the digitization degree of hospital infection monitoring are improved, the hospital infection control level is improved, the hospital infection occurrence rate is reduced, and the safety of patients and the medical staff is guaranteed.

Description

Technical Field

[0001] The present application relates to the field of intelligent medical technology, and specifically to a hospital infection monitoring method and system. Background Art

[0002] Hospital infection is hospital infection, and the occurrence of hospital infection will have many impacts on patients, hospitals and national public health. On the one hand, hospital infection will seriously affect the quality of life of patients; on the other hand, it will also increase the economic burden on patients, hospitals and national public health. In the daily work of hospitals, standardized operating procedures are an important part of hospital infection control.

[0003] At present, there are many technical problems in hospital infection control. First, there is a lack of effective monitoring of standardized diagnosis and treatment operations and hospital infection control steps and measures, which has led to a certain degree of nosocomial infection in major hospitals. Secondly, the degree of digitization of hospital infection monitoring is generally low, and intelligent monitoring has not yet been achieved, which is inconsistent with the development trend of the digital and intelligent era.

[0004] Accordingly, the art needs a new technical solution to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects, the present application is proposed to provide a hospital infection monitoring method and system that solves or at least partially solves the current technical problems of lack of effective monitoring of standardized diagnosis and treatment operations, hospital infection control steps and measures, and the generally low degree of digitization of hospital infection monitoring.

[0006] In a first aspect, a hospital infection monitoring method is provided, the method comprising:

[0007] Obtaining hand hygiene monitoring data of medical staff and / or disinfection monitoring data of medical equipment;

[0008] Processing the hand hygiene monitoring data and / or the disinfection monitoring data to obtain effective monitoring data;

[0009] Analyze the effective monitoring data to obtain analysis results; wherein the analysis results include hand images, personnel identification information and / or equipment images, equipment identification information;

[0010] The data in the analysis results are integrated to obtain hospital infection monitoring results.

[0011] In a technical solution of the above-mentioned hospital infection monitoring method, obtaining the hand hygiene monitoring data of medical staff and / or the disinfection monitoring data of medical equipment includes:

[0012] In response to a medical staff and / or medical equipment entering a preset monitoring area, obtaining hand hygiene monitoring data of the medical staff and / or disinfection monitoring data of the medical equipment;

[0013] The hand hygiene monitoring data includes at least one of first image data, pressure data and radar data; and the disinfection monitoring data includes at least one of second image data and temperature data.

[0014] In a technical solution of the above-mentioned hospital infection monitoring method, the hand hygiene monitoring data and / or the disinfection monitoring data are processed to obtain effective monitoring data including:

[0015] Preprocessing the hand hygiene monitoring data and / or the disinfection monitoring data to obtain standard monitoring data;

[0016] The standard monitoring data is input into a convolutional neural network, and the standard monitoring data is screened and identified based on the convolutional neural network to obtain the effective monitoring data.

[0017] In a technical solution of the above-mentioned hospital infection monitoring method, the effective monitoring data is analyzed to obtain analysis results including:

[0018] Inputting the valid monitoring data into an artificial intelligence model, identifying the valid monitoring data based on the artificial intelligence model, and obtaining an identification result;

[0019] The recognition result is evaluated based on a decision tree or random forest algorithm to obtain the analysis result.

[0020] In a technical solution of the above-mentioned hospital infection monitoring method, the hospital infection monitoring result includes compliance of monitoring data or non-compliance of monitoring data; the method further includes:

[0021] When the hospital infection monitoring result shows that the monitoring data is non-compliant, an alarm signal is triggered;

[0022] Wherein, the alarm signal includes at least one of an audio alarm signal, a text alarm signal and a light alarm signal.

[0023] In a technical solution of the above-mentioned hospital infection monitoring method, the preprocessing of the hand hygiene monitoring data and / or the disinfection monitoring data includes:

[0024] At least one of grayscale conversion processing, noise removal processing and edge detection processing is performed on the first image data and / or the second image data.

[0025] In a technical solution of the above-mentioned hospital infection monitoring method, the identifying the effective monitoring data based on the artificial intelligence model includes:

[0026] Identify the first image data to obtain the hand cleaning method and cleaning duration of the medical staff;

[0027] and / or, identifying the pressure data to obtain water usage data and disinfectant usage data of medical staff;

[0028] and / or, identifying the radar data to obtain hand movement data of the medical staff;

[0029] and / or, identifying the second image data to obtain a disinfection method and disinfection duration of the medical device;

[0030] And / or, identifying the temperature data to obtain the surface temperature of the medical device.

[0031] In a technical solution of the above-mentioned hospital infection monitoring method, the evaluating the recognition result includes:

[0032] Determine whether the medical staff's hand cleaning method and cleaning duration, the water usage data and disinfectant usage data, and the hand movement data meet the specifications;

[0033] and / or,

[0034] Determine whether the disinfection method and disinfection duration of the medical device and the surface temperature meet preset conditions.

[0035] In a second aspect, the present application provides a hospital infection monitoring system, the system comprising:

[0036] A sensor module configured to obtain hand hygiene monitoring data of medical personnel and / or disinfection monitoring data of medical equipment;

[0037] a processing module, configured to process the hand hygiene monitoring data and / or the disinfection monitoring data to obtain effective monitoring data;

[0038] An analysis module, configured to analyze the effective monitoring data to obtain analysis results; wherein the analysis results include hand images, personnel identification information and / or device images, device identification information;

[0039] A fusion module is configured to fuse the data in the analysis results to obtain hospital infection monitoring results.

[0040] In a technical solution of the above hospital infection monitoring system, the hospital infection monitoring system further includes an interaction module;

[0041] The interactive module is provided with an interactive interface, and the interactive interface is configured to visually display the hospital infection monitoring results.

[0042] The above one or more technical solutions of this application have at least one or more of the following Beneficial effects:

[0043] In the implementation of the technical solution of this application, first obtain the hand hygiene monitoring data of medical staff and / or the disinfection monitoring data of medical equipment, then process the hand hygiene monitoring data and / or the disinfection monitoring data to obtain effective monitoring data, and analyze the effective monitoring data to obtain analysis results; wherein the analysis results include hand images, personnel identification information and / or equipment images, equipment identification information; finally, fuse the data in the analysis results to obtain hospital infection monitoring results. Through the above implementation, it is possible to digitally process and intelligently analyze the hand hygiene monitoring data of medical staff and the disinfection monitoring data of medical equipment, integrate data from different sensors, obtain hospital infection monitoring results, realize real-time monitoring and management of medical staff hand hygiene and medical equipment disinfection, improve the intelligence and digitization of hospital infection monitoring, improve the level of hospital infection control, reduce the incidence of hospital infection, and ensure the safety of patients and medical staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The disclosure of the present application will become easier to understand with reference to the accompanying drawings. It is easy for those skilled in the art to understand that these drawings are only for illustrative purposes and are not intended to limit the scope of protection of the present application. Among them:

[0045] Figure 1 It is a schematic flow chart of the main steps of a hospital infection monitoring method according to an embodiment of the present application;

[0046] Figure 2 It is a flowchart of the main steps of processing hand hygiene monitoring data and / or disinfection monitoring data to obtain effective monitoring data according to an embodiment of the present application;

[0047] Figure 3 This is a flowchart of the main steps of analyzing effective monitoring data and obtaining analysis results according to an embodiment of the present application;

[0048] Figure 4 It is a main structural block diagram of a hospital infection monitoring system according to an embodiment of the present application.

[0049] List of reference numerals:

[0050] 401: sensor module; 402: processing module; 403: analysis module; 404: fusion module. DETAILED DESCRIPTION

[0051] Some embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application and are not intended to limit the protection scope of the present application.

[0052] In the description of this application, "module" may include hardware, software or a combination of the two. A module may include hardware circuits, various suitable sensors, communication ports, memories, and may also include software parts, such as program codes, or a combination of software and hardware. The term "A and / or B" represents all possible combinations of A and B, such as just A, just B, or A and B. The term "at least one A or B" or "at least one of A and B" has a similar meaning to "A and / or B" and may include just A, just B, or A and B. The singular terms "a" and "the" may also include plural forms.

[0053] Here we first explain some terms involved in this application.

[0054] Nosocomial infection: also known as hospital infection, refers to the infection acquired by inpatients in the hospital, including infections occurring during hospitalization and infections occurring after discharge from the hospital, but excluding infections that have started before admission or are already in the incubation period at admission. Infections acquired by hospital staff in the hospital are also considered nosocomial infections.

[0055] As mentioned in the background technology, the occurrence of hospital infection will have many impacts on patients, hospitals and national public health. On the one hand, hospital infection will seriously affect the quality of life of patients; on the other hand, it will also increase the economic burden on patients, hospitals and national public health. In the daily work of hospitals, standardized operating procedures are an important part of hospital infection control.

[0056] At present, there are many technical problems in hospital infection control. First, there is a lack of effective monitoring of standardized diagnosis and treatment operations and hospital infection control steps and measures, which has led to a certain degree of nosocomial infection in major hospitals. Secondly, the degree of digitization of hospital infection monitoring is generally low, and intelligent monitoring has not yet been achieved, which is inconsistent with the development trend of the digital and intelligent era.

[0057] In order to solve the above problems, the present application provides a hospital infection monitoring method and system.

[0058] See attached Figure 1 , Figure 1 FIG. 1 is a flow chart of the main steps of a hospital infection monitoring method according to an embodiment of the present application. Figure 1 As shown, the hospital infection monitoring method in the embodiment of the present application mainly includes the following steps S101 to S104.

[0059] Step S101: Acquire hand hygiene monitoring data of medical staff and / or disinfection monitoring data of medical equipment;

[0060] Step S102: Processing the hand hygiene monitoring data and / or the disinfection monitoring data to obtain effective monitoring data;

[0061] Step S103: Analyze the effective monitoring data to obtain analysis results;

[0062] The analysis results include hand images, personnel identification information and / or device images, device identification information.

[0063] Step S104: Fuse the data in the analysis results to obtain hospital infection monitoring results.

[0064] Based on the method described in steps S101 to S104 above, the hand hygiene monitoring data of medical staff and the disinfection monitoring data of medical equipment can be digitally processed and intelligently analyzed, and data from different sensors can be integrated to obtain hospital infection monitoring results, thereby realizing real-time monitoring and management of medical staff's hand hygiene and medical equipment disinfection, improving the intelligence and digitization of hospital infection monitoring, improving the level of hospital infection control, reducing the incidence of hospital infection, and ensuring the safety of patients and medical staff.

[0065] The above steps S101 to S104 are further explained below.

[0066] In some embodiments of step S101, in response to the medical staff and / or medical equipment entering a preset monitoring area, the medical staff's hand hygiene monitoring data and / or the medical equipment's disinfection monitoring data may be acquired.

[0067] Specifically, corresponding infrared sensors can be set up in different areas of the hospital where hospital infection monitoring is required, such as general ward areas, intensive care units, nurse stations, treatment rooms and toilets. The infrared sensors can be used to detect whether medical staff or medical equipment have entered the corresponding monitoring area.

[0068] Among them, infrared sensors are devices that use the physical properties of infrared rays for detection. Infrared rays are invisible light with thermal radiation properties. When medical staff enter the preset monitoring area, the infrared sensor will detect changes in infrared radiation because the human body emits infrared radiation of a specific intensity and wavelength. This change will be captured by the sensor and converted into an electrical signal. Similarly, when medical equipment enters the preset monitoring area, the infrared radiation it emits will change the infrared radiation distribution in the area. The infrared sensor will detect this change and convert it into an electrical signal for processing.

[0069] Furthermore, the circuit inside the infrared sensor will analyze and process the electrical signal. If the amplitude of the signal change exceeds the preset threshold, it will be determined that a medical staff or medical equipment has entered the area. At this time, the infrared sensor can send a data collection signal to other sensors to start acquiring the medical staff's hand hygiene monitoring data and / or the disinfection monitoring data of the medical equipment.

[0070] In some embodiments, other sensors may include cameras, millimeter wave radars, pressure sensors, etc.

[0071] The acquired hand hygiene monitoring data includes at least one of first image data, pressure data and radar data; the disinfection monitoring data includes at least one of second image data and temperature data.

[0072] Specifically, pictures and video images of medical staff (i.e., first image data) and pictures and video images of medical equipment (i.e., second image data) can be obtained through cameras; pressure data of hand sanitizer, disinfectant and water flow can be obtained through pressure sensors; radar data of medical staff's hand movements can be obtained through millimeter-wave radars; temperature data of medical equipment can be obtained through infrared sensors, etc.

[0073] Among them, medical equipment may include diagnostic equipment (such as X-ray machines, computer tomography scanners, magnetic resonance imaging equipment, electrocardiographs, electroencephalographs, endoscopes, etc.), therapeutic equipment (such as ventilators, anesthesia machines, dialysis machines, laser therapy devices, hyperbaric oxygen chambers, etc.), monitoring equipment (such as electrocardiographs, fetal monitors, multi-parameter monitors, etc.), and testing equipment (such as biochemical analyzers, blood cell analyzers, immunoassay analyzers, etc.).

[0074] It should be pointed out that the above examples of sensors, medical equipment, and the acquired hand hygiene monitoring data of medical staff and disinfection monitoring data of medical equipment are only schematic illustrations. In actual applications, technical personnel in this field can make settings according to specific scenarios and needs, and no limitation is made here.

[0075] The above is a further description of step S101. The following is a further description of step S102.

[0076] In some implementations of the above step S102, see the attached Figure 2 , Figure 2 This is a flow chart of the main steps of processing hand hygiene monitoring data and / or disinfection monitoring data to obtain effective monitoring data according to an embodiment of the present application. Figure 2 As shown, it mainly includes the following steps S1021 to S1022.

[0077] Step S1021: pre-processing the hand hygiene monitoring data and / or disinfection monitoring data to obtain standard monitoring data;

[0078] In some implementations of step S1021, at least one of grayscale conversion processing, noise removal processing, and edge detection processing may be performed on the first image data and / or the second image data.

[0079] Among them, grayscale conversion processing can convert color images or video frames into grayscale images, reducing the amount of data while highlighting the brightness information of the image, making the image features more obvious and convenient for subsequent processing. Noise removal can eliminate interference factors in the data, such as electronic noise and uneven lighting, improve the accuracy and reliability of the data, and make subsequent analysis and processing more accurate. Edge detection can highlight the edge and contour information of objects in the image. The edge usually contains important feature information in the image. Through edge detection, the shape, size and position of the object can be better identified, providing a basis for subsequent image analysis and recognition tasks.

[0080] Specifically, when performing grayscale conversion processing, the following formula (1) can be used:

[0081] I(x,y)=0.299R+0.587G+0.114B (1)

[0082] Among them, I(x,y) represents the grayscale value of the pixel with coordinates (x,y); (x,y) represents the pixel coordinates of a specific position in the image.

[0083] Each pixel of a color image consists of three color channels (R, G, B), representing red, green, and blue, respectively, while a grayscale image has only one channel, representing brightness information. The above formula (1) can perform a weighted calculation of the color values ​​of the three red, RGB channels of the pixel at that position in the color image to obtain a single grayscale value, thereby realizing the conversion of a color image to a grayscale image.

[0084] Among them, the weight of the red channel R is 0.299, indicating that red has little effect on the final grayscale value, the weight of the green channel is 0.587, indicating that green has the greatest impact on the grayscale value, and the weight of the blue channel is 0.114, indicating that blue has the least impact on the grayscale value.

[0085] It should be noted that the above weighting coefficients 0.299, 0.587 and 0.114 are designed based on the sensitivity of the human eye to the three colors. The human eye is most sensitive to green, so green has the highest weight. In practical applications, those skilled in the art can set the weighting coefficients of the three color channels according to specific scenarios, which is not limited here.

[0086] In some implementations, when performing noise removal processing, methods such as Gaussian filtering, median filtering, and bilateral filtering may be used.

[0087] Taking Gaussian filtering as an example, the formula for noise removal is shown in the following formula (2):

[0088]

[0089] Where G(x,y) represents the value of the Gaussian filter at the coordinate (x,y); σ is the standard deviation of the Gaussian function; is the normalization coefficient; e is the base of the natural logarithm, which is approximately equal to 2.7183.

[0090] In the above Gaussian filter formula, σ determines the width of the Gaussian kernel. A larger σ value will make the filter smoother, but may blur image details. A smaller σ value will keep the filter sharper edges. When calculating weights, the farther (x, y) is from the center, the larger the σ value is, and the exponential part will change the decay speed with the change of . When σ is large, the exponential part decays slower, and pixels farther from the center will have relatively large weights; when σ is small, the exponential part decays faster, and only pixels closer to the center will have larger weights.

[0091] Specifically, the Gaussian filtering process generally includes the following steps ab:

[0092] a. Define Gaussian kernel: Calculate the size of the Gaussian kernel (usually an odd number, such as 3x3, 5x5, etc.) based on the given standard deviation σ. The value of the Gaussian kernel is obtained by sampling the Gaussian function in two-dimensional space.

[0093] b. Apply the filter: that is, apply the Gaussian kernel to each pixel of the image. Specifically, for each pixel (x, y) in the image, multiply the pixel values ​​in its surrounding neighborhood by the value of the corresponding position of the Gaussian kernel, add up the results, and finally divide by the sum of all elements of the Gaussian kernel. In this way, the new pixel value after filtering is obtained.

[0094] Compared with other noise removal methods, the advantage of Gaussian filtering is that it can effectively remove Gaussian noise. When smoothing images, it will not blur the edges like average filtering. In practical applications, Gaussian filtering is often used in the image preprocessing stage to provide a better foundation for subsequent edge detection, feature extraction and other operations.

[0095] In some implementations, when edge detection processing is performed, common edge detection algorithms include Robert operator, Sobel operator, Prewitt operator, Canny edge detection algorithm, and the like.

[0096] Among them, the Robert operator is an edge detection operator based on local difference. It detects edges by calculating the gradient between adjacent pixels in the image. Its advantages are simple calculation and fast speed, but it is sensitive to noise, the detected edges are relatively thin, and sometimes discontinuity may occur.

[0097] The Sobel operator is also a gradient-based edge detection operator. It determines the direction and strength of the edge by calculating the gradients in the horizontal and vertical directions. Compared with the Robert operator, it is less sensitive to noise and the detected edges are more stable. At the same time, the accuracy of edge detection can be controlled by adjusting the size of the filter.

[0098] The Prewitt operator is similar to the Sobel operator in that it also detects edges by calculating gradients. It uses different templates to calculate gradients in the horizontal and vertical directions, and has a certain suppressive effect on noise, but its detection effect is slightly inferior to that of the Sobel operator.

[0099] The Canny edge detection algorithm is considered to be a relatively excellent edge detection algorithm. It includes the steps of Gaussian filtering to smooth the image, calculating the gradient amplitude and direction, non-maximum suppression, double threshold processing and edge connection. It has a good edge detection effect and can detect relatively accurate and continuous edges. At the same time, it has a certain suppression effect on noise.

[0100] In practical applications, those skilled in the art can select corresponding algorithms for edge detection processing according to specific scenarios, which is not limited here.

[0101] Through the above-mentioned grayscale conversion processing, noise removal processing and edge detection processing, standard first image data and / or second image data can be obtained.

[0102] In some implementations, the pressure data may be processed by removing outliers, smoothing the data, and converting units.

[0103] Specifically, it is possible to check whether there are values ​​in the pressure data that are significantly deviated from the normal range. These abnormal values ​​may be caused by sensor failure, measurement error or other interference factors. The abnormal values ​​can be identified and removed by setting thresholds or using statistical methods (such as the standard deviation method). For example, if the value range of the pressure data is usually within a certain interval, then the values ​​outside this interval can be regarded as abnormal values ​​and removed.

[0104] There may be some noise in the pressure data. In order to reduce the impact of noise, data smoothing can be performed. Common methods include moving average method, median filtering, etc. For example, the moving average method takes the average value of data within a certain window size as the smoothed value, and the average value of 5 consecutive data points can be taken as the smoothed value of the middle data point.

[0105] Furthermore, if the units of the pressure data are not uniform or do not meet the requirements of subsequent analysis, unit conversion is required. The purpose of unit conversion is to ensure that all pressure data are in the same unit for comparison and analysis.

[0106] In some implementations, the radar data may be processed by noise removal, coordinate conversion, target recognition and tracking, etc.

[0107] Specifically, radar data may be interfered by various noises, such as environmental noise, electromagnetic interference, etc., so filtering methods can be used to remove noise, such as Gaussian filtering, median filtering, etc.

[0108] If the coordinate system of the radar data is inconsistent with the coordinate system required for subsequent analysis, coordinate conversion is required, for example, converting the radar data from polar coordinates to rectangular coordinates, or converting the coordinates of different radar devices uniformly.

[0109] Furthermore, targets related to hand hygiene or disinfection monitoring can be identified from radar data, such as the hands of medical staff, disinfection equipment, etc. Target detection algorithms, such as threshold-based methods and clustering algorithms, can be used to track the identified targets in order to obtain their motion trajectories and state changes. Specifically, tracking algorithms such as Kalman filtering and particle filtering can be used, which are not limited here.

[0110] In some implementations, the temperature data may be processed for outlier processing, calibration, and data normalization.

[0111] Similar to pressure data, outliers in temperature data, that is, temperature values ​​that are too high or too low, may be erroneous measurement results. A reasonable temperature range can be set according to actual conditions to remove outliers that exceed the range.

[0112] In addition, the temperature sensor may have certain errors and need to be calibrated. Specifically, the sensor can be calibrated using a standard temperature source, or the temperature data can be adjusted by comparing it with a device with a known accurate temperature.

[0113] Data normalization is to facilitate the comparison and analysis of different temperature data. Data normalization can be performed to map the temperature data to a specific range. Commonly used normalization methods include minimum-maximum normalization, Z-score normalization, and other methods.

[0114] Through the above preprocessing method, the hand hygiene monitoring data of medical staff and / or the disinfection monitoring data of medical equipment can be converted into standard monitoring data, providing a reliable basis for subsequent data analysis and processing.

[0115] The above is a further explanation of step S1021.

[0116] Step S1022: input the standard monitoring data into the convolutional neural network, screen and identify the standard monitoring data based on the convolutional neural network, and obtain effective monitoring data.

[0117] Convolutional neural network is a deep learning model specially used for processing data. It consists of convolutional layer, pooling layer, fully connected layer, etc. Among them, the convolutional layer performs convolution operation with the input data through the convolution kernel to extract local features in the data. The pooling layer is used to reduce the dimension of the data and reduce the amount of calculation while retaining the main features. The fully connected layer integrates the extracted features and outputs the final classification or prediction results.

[0118] When standard monitoring data is input into the convolutional neural network, the convolutional neural network will automatically learn the characteristic patterns in the data. For hand hygiene monitoring and disinfection monitoring data, the convolutional neural network can learn the association between different data patterns and effective monitoring results. Therefore, after the standard monitoring data is input into the convolutional neural network, the convolutional neural network can screen the standard monitoring data, remove those data that are irrelevant to the effective monitoring results or have high noise, and only retain data that is valuable for effective monitoring. On the basis of screening, the convolutional neural network can also identify the standard monitoring data, determine which data corresponds to effective hand hygiene operations or disinfection operations, and divide the input data into effective monitoring data and invalid monitoring data by analyzing and classifying the data features, and retain the effective monitoring data for output for subsequent analysis and decision-making.

[0119] Among them, commonly used convolutional neural network models include YOLO (You Only Look Once, a target detection algorithm based on deep neural network), Faster R-CNN (Faster Region-based Convolutional Neural Networks, a deep learning model for target detection), etc.

[0120] Take the YOLO algorithm as an example. The YOLO algorithm is a real-time target detection method. Its core idea is to complete feature extraction and bounding box prediction in a single network structure. The loss function of the YOLO algorithm consists of three parts: location loss, confidence loss, and classification loss. These loss functions jointly guide the network to learn how to accurately predict the location, confidence, and category of the target.

[0121] Specifically, the positioning loss measures the error between the predicted bounding box center coordinates and the true bounding box center coordinates. The program code of the formula part is as follows:

[0122] [\lambda_{\text{coord}}\sum_{i=0}^{S^2}\sum_{j=0}^{B}\mathbb{1}_{ij}^{\text{obj}}\left[(x_i-\hat{x}_i)^2+(y_i-\hat{y}_i)^2\right]]

[0123] where “\lambda_{\text{coord}}” is the weight of the localization loss; “S” is the side length of the grid (e.g., a 640x640 image might be divided into an 8x8 grid); “B” is the maximum number of bounding boxes that can be predicted for each grid cell; “\mathbb{1}_{ij}^{\text{obj}}” is an indicator variable that is 1 if the grid cell “(i,j)” contains an object and 0 otherwise; “x_i” and “y_i” are the center coordinates of the ground-truth bounding box in the i-th grid cell; “\hat{x}_i” and “\hat{y}_i” are the center coordinates of the bounding box predicted by the model in the i-th grid cell.

[0124] Confidence loss measures the difference between the confidence of whether the predicted bounding box contains an object and the actual situation. The program code of the formula part is as follows:

[0125] [\lambda_{\text{conf}}\sum_{i=0}^{S^2}\sum_{j=0}^{B}\ma thbb{1}_{ij}^{\text{obj}}(\hat{C}_i-C_i)^2]

[0126] Among them, “\lambda_{\text{conf}}” is the weight of the confidence loss; “\hat{C}_i” is the confidence of the bounding box in the i-th grid cell predicted by the model; “C_i” is the confidence of the true bounding box (if the grid cell contains an object, then C_i=1, otherwise C_i=0).

[0127] The classification loss measures the difference between the probability of the predicted bounding box category and the true category. The program code of the formula is as follows:

[0128]

[0129] Among them, “\lambda_{\text{class}}” is the weight of classification loss; is the probability predicted by the model that the bounding box in the i-th grid cell belongs to category c; is the probability that the true bounding box belongs to category c (if the grid cell contains an object and the true category is c, then \ ",otherwise ).

[0130] The loss function of the YOLO algorithm takes into account the positioning accuracy, the confidence of the target existence, and the accuracy of the predicted category. By minimizing this loss function, the network can gradually optimize its prediction ability during the training process, thereby achieving more accurate target detection.

[0131] Through the above steps S1021 to S1022, standard monitoring data can be screened and identified to obtain effective monitoring data.

[0132] The above is a further description of step S102. The following is a further description of step S103.

[0133] In some implementations of the above step S103, see the attached Figure 3 , Figure 3 FIG. 1 is a flow chart of the main steps of analyzing effective monitoring data and obtaining analysis results according to an embodiment of the present application. Figure 3 As shown, it mainly includes the following steps S1031 to S1032.

[0134] Step S1031: input the effective monitoring data into the artificial intelligence model, identify the effective monitoring data based on the artificial intelligence model, and obtain an identification result;

[0135] The artificial intelligence model is a tool built using machine learning algorithms that can analyze and predict data. It can automatically learn patterns and rules in the data based on various input sensor data combined with the professional behavior of medical staff, and use this knowledge to process and predict new data.

[0136] In some embodiments, the artificial intelligence model can be a timing model, including a long short-term memory network (LSTM), a gated recurrent unit (GRU), etc. Through the timing model, the hand hygiene behavior of medical staff and the disinfection behavior of medical equipment can be identified.

[0137] Taking LSTM as an example, the program code of the formula part is as follows:

[0138]

[0139] LSTM can effectively capture dependencies in long time series by introducing "memory unit" and "gate" structures. The basic structure of LSTM includes input gate, forget gate and output gate. The following is a step-by-step explanation of the meaning of the above program code.

[0140] 1. Input Gate: [i_t=\sigma(W_{ix}x_t+W_{ih}h_{t-1}+b_i)]

[0141] Among them, "i_t" is the input gate value of the tth step; "W_{ix}" and "W_{ih}" are the weight matrices of the input gate for the current input "x_t" and the hidden state "h_{t-1}" at the previous moment respectively; "b_i" is the bias vector of the input gate; "\sigma" is a nonlinear activation function, usually a sigmoid function, whose range is [0,1].

[0142] 2. Output Gate: [o_t=\sigma(W_{ox}x_t+W_{oh}h_{t-1}+b_o)]

[0143] Among them, "o_t" is the output gate value of the tth step; "W_{ox}" and "W_{oh}" are the weight matrices of the output gate for the current input "x_t" and the hidden state "h_{t-1}" at the previous moment respectively; "b_o" is the bias vector of the output gate.

[0144] 3. Forget Gate: [f_t=\sigma(W_{fx}x_t+W_{fh}h_{t-1}+b_f)]

[0145] Among them, "f_t" is the forget gate value of the tth step; "W_{fx}" and "W_{fh}" are the weight matrices of the forget gate for the current input "x_t" and the hidden state "h_{t-1}" at the previous moment respectively; "b_f" is the bias vector of the forget gate.

[0146] 4. New state (Cell State): The new state consists of two parts, one is the part that the forget gate decides to discard, and the other is the new information that the input gate decides to add. The program code of the formula part is: [\tilde{C}t=\tanh(W{cx}x_t+W_{ch}h_{t-1}+b_c)]

[0147] Among them, "\tilde{C}_t" is the new state of the tth step; "W_{cx}" and "W_{ch}" are the weight matrices of the new state for the current input "x_t" and the hidden state "h_{t-1}" at the previous moment respectively; "b_c" is the bias vector of the new state; "\tanh" is the hyperbolic tangent activation function, whose range is [-1,1].

[0148] 5. Current state (Hidden State): The current state is the result of multiplying the new state by the output gate. The program code for the formula is: [h_t=o_t\odot\tanh(C_t)]

[0149] Among them, “h_t” is the hidden state at the tth step; “C_t” is the cell state at the tth step; “\odot” represents the element-by-element multiplication operation.

[0150] The calculation of the above LSTM mainly includes the steps of calculating the input gate, calculating the forget gate, calculating the output gate, calculating the new state, updating the cell state, and calculating the current state. These steps ensure that LSTM can effectively learn long-term dependencies and can better process time series data.

[0151] Furthermore, in some implementations of step S1031, the effective monitoring data may be identified by LSTM using one or more of the following:

[0152] The first image data is identified to obtain the medical staff's hand cleaning method and cleaning time; the pressure data is identified to obtain the medical staff's water usage data and disinfectant usage data; the radar data is identified to obtain the medical staff's hand movement data; the second image data is identified to obtain the disinfection method and disinfection time of the medical equipment; the temperature data is identified to obtain the surface temperature of the medical equipment.

[0153] The above is a further explanation of step S1031.

[0154] Step S1032: Evaluate the recognition result based on the decision tree or random forest algorithm to obtain an analysis result.

[0155] Among them, the decision tree is a machine learning algorithm based on a tree structure. It judges a series of features of the data, gradually branches downward from the root node, and finally reaches the leaf node to obtain classification or prediction results.

[0156] Random forest is an ensemble learning algorithm composed of multiple decision trees. It builds multiple different decision trees and integrates their prediction results to improve the accuracy and stability of predictions. Each decision tree randomly selects a part of features and samples for training during the construction process, which can increase the diversity of the model and reduce the risk of overfitting. When evaluating the recognition results, random forest can integrate the opinions of multiple decision trees to give more reliable evaluation results.

[0157] Taking the decision tree as an example, the process of evaluating the recognition results includes the steps of feature selection, building a decision tree, and evaluating the recognition results.

[0158] When performing feature selection, appropriate features should be selected from the recognition results and other related data as the input of the decision tree. These features can include specific attributes in the recognition results, relevant statistics of the original monitoring data, etc.

[0159] Building a decision tree is to use labeled training data to build a decision tree. The training data includes known correct results and corresponding feature values. The decision tree construction process is a recursive process, starting from the root node, selecting the best features for branching according to criteria such as information gain and Gini coefficient until the stopping condition is met, such as reaching the maximum depth or the number of samples at the leaf node is too small.

[0160] When evaluating the recognition results, the recognition results to be evaluated and their corresponding feature values ​​are input into the constructed decision tree. The decision tree starts from the root node, branches according to the feature values, and finally reaches a leaf node to obtain an evaluation result. For example, if the recognition result is that hand hygiene may be qualified, but the decision tree judges it as unqualified based on the input feature value, then this recognition result needs to be further reviewed or adjusted.

[0161] Specifically, in decision trees, entropy is a measure of the degree of confusion or uncertainty in a data set. It is used in decision trees to select the best features to divide data. Decision trees will select features that minimize entropy in order to classify data more clearly.

[0162] For example: If all data belong to the same category (e.g., all medical device disinfections are qualified), the entropy is 0, indicating complete order; if the data are evenly distributed among multiple categories (e.g., the ratio of clean to unclean hands of medical staff is equal), the entropy is large, indicating that the data are relatively chaotic.

[0163] For a dataset S containing n different categories, the formula for calculating its entropy is the following formula (3):

[0164]

[0165] where p i is the probability that a sample of the i-th category appears in the dataset, which can be calculated by dividing the number of samples belonging to that category by the total number of samples in the dataset.

[0166] Furthermore, the program code for the entropy formula of the decision tree is as follows:

[0167]

[0168] where "H(S)", i.e., "entropy", represents the uncertainty or chaos of set S. The higher the entropy, the more difficult it is to classify the data; the lower the entropy, the more concentrated or easier it is to classify the data. "n" is the number of categories in the dataset. For a classification problem, there may be multiple different categories (e.g., two categories of "clean" and "unclean"). "pi" is the probability of the i-th category, i.e., the proportion of category i in the entire dataset. If pi = 1, the data completely belong to a certain category and the entropy is 0 (the most ordered state); if the probabilities of all categories are equal, the entropy is the largest (the most chaotic state). represents the logarithm operation with base 2 of the probability pi of category i, indicating the amount of information of that category. When pi is smaller, the amount of information is larger because rare events contain more information. The role of the logarithm is to make small-probability events produce a larger information increment. The negative sign "-" is used to take the negative value of the sum of the entropies because for 0 < pi < 1, it is negative, while the value of entropy needs to be positive. The sum "∑" is to perform a weighted sum of the probabilities pi of all categories and the corresponding amounts of information to obtain the entropy of the entire dataset.

[0169] Furthermore, in some embodiments of step S1032, it is possible to determine whether the hand cleaning methods, cleaning durations, water usage data, disinfectant usage data, and hand movement data of medical staff comply with the specifications in the above manner.

[0170] Specifically, when medical staff perform hand cleaning, they usually have a specific standard process, such as the seven-step hand washing method. Therefore, a decision tree or random forest algorithm can be used to determine whether the cleaning method actually used by the medical staff is the standard seven-step hand washing method, and whether the cleaning time of each step meets the standard.

[0171] When medical staff clean their hands, the amount of water used, the water flow rate, the use of disinfectant, etc. may affect the cleaning effect. Therefore, a decision tree or random forest algorithm can be used to determine whether medical staff use water in a standardized manner and whether disinfectant is used correctly, including whether the concentration of the disinfectant meets the requirements and whether the amount used is sufficient.

[0172] When medical staff clean their hands, the movement and amplitude of their hands will also affect the cleaning effect. For example, the correct hand washing method requires sufficient rubbing of all parts of the hands, including fingers, palms, backs of hands, etc. Therefore, a decision tree or random forest algorithm can be used to determine whether the hand movements of medical staff meet the standards and whether they can effectively clean all parts of the hands.

[0173] In some implementations of step S1032, the above-mentioned method can also be used to determine whether the disinfection method, disinfection duration, and surface temperature of the medical device meet preset conditions.

[0174] Specifically, different medical devices may require different disinfection methods, such as high-temperature steam disinfection, chemical disinfectant disinfection, ultraviolet disinfection, etc. Therefore, it is necessary to use a decision tree or random forest algorithm to determine whether the actual disinfection method is consistent with the requirements and specifications of the device. In addition, each disinfection method usually has a recommended disinfection time length to ensure that bacteria, viruses and other pathogens on the surface of the device can be completely killed. Therefore, a decision tree or random forest algorithm can be used to determine whether the disinfection time of medical equipment is sufficient.

[0175] For some medical devices that use high-temperature disinfection, surface temperature is an important indicator. Decision trees or random forest algorithms can be used to determine whether the surface temperature of the device during the disinfection process has reached the specified temperature requirements to ensure the effectiveness of disinfection. At the same time, excessively high surface temperatures may also damage the device, so it is necessary to ensure that the temperature is within a safe range.

[0176] By analyzing the hand cleaning of medical staff and the disinfection process of medical equipment, we can obtain analysis results including medical staff hand images, personnel identification information (i.e. which medical staff), equipment images, equipment identification information (i.e. which medical equipment), etc. These results can provide hospital managers with detailed information on the hand hygiene of medical staff and the disinfection of medical equipment, helping them take corresponding measures to improve the hospital's hygiene level and medical quality.

[0177] The above is a further description of step S103. The following is a further description of step S104.

[0178] In some implementations of the above step S104, the monitoring data of different sensors can be integrated to form a unified evaluation report to obtain the final hospital infection monitoring result.

[0179] Specifically, after obtaining the analysis results of each monitoring data, the data from different sources can be fused in a weighted manner to obtain more comprehensive and accurate information.

[0180] In some implementations, the program code of the monitoring data fusion formula part is as follows:

[0181] Dfused=∑i=1nwiDi∑i=1nwiD_{\text{fused}}=\frac{\sum_{i=1}^{n}w_i D_i}{\sum_{i=1}^{n}w_i}Dfused=∑i=1n wi∑i=1n wi Di

[0182] Wherein, "wi" is the weight of the ith data source, indicating the relative importance of each data source in the fusion process. The weight determines the impact of the data source on the fusion result. The larger the weight value, the higher the credibility or importance of the data source in the evaluation. "Di" is the measurement value of the ith data source, that is, the actual observation value from the ith sensor. The Di of each data source may measure different aspects of hand hygiene monitoring data and disinfection monitoring data, such as temperature data provided by infrared sensors, image data provided by cameras, etc. "DfusedD_{\text{fused}}Dfused" represents the unified evaluation result obtained by integrating multiple data sources through weighted average. This value is calculated comprehensively from the monitoring data Di of multiple sensors to form a more comprehensive and accurate evaluation report. "n" represents the number of data sources involved in the fusion, which may be multiple sensors, each of which provides different measurement values ​​related to hand hygiene or medical device disinfection. "∑i=1nwiDi\sum_{i=1}^{n}w_i D_i∑i=1n "wiDi" means the weighted sum of the measurement values ​​Di of all data sources, which reflects the measurement information under the comprehensive effect of all data sources. Each measurement value Di is multiplied by the corresponding weight wi, which can reflect the importance of different data sources; "∑i=1nwi\sum_{i=1}^{n}w_i∑i=1n wi" means the sum of the weights wi of all data sources, which is used for standardized weighted averaging, so as to ensure that the final fusion result Dfused is within a reasonable range.

[0183] Data fusion integrates measurements from different data sources through a weighted average method to generate a unified evaluation result. It can adjust the impact of each data source on the result according to the credibility or importance of each data source, thereby improving the accuracy and reliability of hand hygiene monitoring and medical equipment disinfection evaluation.

[0184] In some embodiments of step S104, the hospital infection monitoring results in the evaluation report include compliance of the monitoring data or non-compliance of the monitoring data.

[0185] When the hospital infection monitoring result is that the monitoring data is compliant, it means that after analyzing the monitoring data collected by different sensors, it is determined that all indicators of the hospital in terms of infection control meet the established specifications and standards. That is, the medical staff's hand cleaning methods, cleaning time, water use data, disinfectant usage data, and hand movement data all meet the regulatory requirements. At the same time, the disinfection methods, disinfection time, and surface temperature of medical equipment are also within the preset conditions. This shows that the hospital's infection control measures are well implemented during the current monitoring cycle, and the risk of infection is relatively low.

[0186] On the contrary, if the hospital infection monitoring result shows that the monitoring data is not compliant, it means that non-compliance with norms and standards was found during the monitoring process. It may be that there are certain problems in the hand cleaning process of medical staff, such as incorrect cleaning methods, insufficient cleaning time, improper water use, or insufficient use of disinfectant, etc.; it may also be that the disinfection of medical equipment does not meet the requirements, such as incorrect disinfection methods, insufficient disinfection time, or surface temperature does not meet the preset conditions. In this case, it means that there is a certain risk of infection in the hospital, and timely measures need to be taken to correct and improve.

[0187] Furthermore, when the hospital infection monitoring result shows that the monitoring data is non-compliant, an alarm signal can be triggered.

[0188] The logic for triggering the alarm signal may be:

[0189] if Hand Hygiene Status=False or Device Disinfection Status=F alsethen Trigger Alarm\text{if}\text{Hand Hygiene Status}=\text{False}\text{or}\text{Device Disinfection Status}=\text{False}\text{then Trigger

[0190] Alarm}ifHand Hygiene Status=False or Device Disinfection Status=False then Trigger Alarm

[0191] Among them, "Hand Hygiene Status" indicates the status of hand hygiene. When the status is False, it means that the hand hygiene is unqualified and proper hand washing or disinfection has not been performed.

[0192] “Device Disinfection Status” indicates the status of medical device disinfection. When the status is False, it means that the medical device has not been disinfected or the disinfection is unqualified.

[0193] The "or" operator represents a logical "or" condition, that is, the condition is true when either side (hand hygiene or medical device disinfection) fails (i.e., the status is False).

[0194] "TriggerAlarm" means that if the conditions are met, that is, hand hygiene is not up to standard or medical equipment disinfection is not up to standard, an alarm will be triggered. Triggering an alarm means that the system will immediately issue a prompt or notification to remind relevant personnel to make corrections.

[0195] The above program code is the core logic of the real-time monitoring system. When it is detected that the hand hygiene or medical equipment disinfection status is False (that is, either one fails to pass the standard), an alarm will be triggered to remind you to take immediate action.

[0196] The alarm signal includes at least one of an audio alarm signal, a text alarm signal and a light alarm signal.

[0197] Specifically, the audio alarm signal can attract the attention of people around by emitting specific sounds, such as alarm sounds, buzzers, etc. This alarm method can also play a good reminder role in a noisy environment.

[0198] Text alarm signals can display text information to clearly indicate the specific content and location of non-compliance. For example, it can display "Hand cleaning in a certain department is not compliant" or "Disinfection of a certain equipment does not meet the standards", etc., so that relevant personnel can quickly locate the problem and take measures.

[0199] Light alarm signals can use flashing lights of different colors to indicate non-compliance. For example, flashing red lights indicate serious non-compliance, flashing yellow lights indicate minor non-compliance, etc. Light alarm signals can also be easily detected at a long distance or in a dim environment.

[0200] By clarifying the classification of hospital infection monitoring results and triggering various types of alarm signals in non-compliance situations, hospital staff can be reminded in a timely and effective manner to pay attention to infection control issues, take appropriate measures to reduce infection risks, and help ensure that the hygiene standards in the environment meet the requirements.

[0201] The above is an explanation of the hospital infection monitoring method provided in this application.

[0202] Through the above implementation methods, the hand hygiene monitoring data of medical staff and the disinfection monitoring data of medical equipment can be digitally processed and intelligently analyzed, and data from different sensors can be integrated to obtain hospital infection monitoring results, thereby realizing real-time monitoring and management of medical staff's hand hygiene and medical equipment disinfection, improving the intelligence and digitization of hospital infection monitoring, improving the level of hospital infection control, reducing the incidence of hospital infection, and reducing overall treatment costs. At the same time, it also reduces the risk of hospital infection outbreaks and ensures the safety of patients and medical staff.

[0203] It should be pointed out that although the various steps in the above embodiments are described in a specific order, those skilled in the art can understand that in order to achieve the effect of the present application, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders. These changes are within the scope of protection of the present application.

[0204] Furthermore, the present application also provides a hospital infection monitoring system.

[0205] See attached Figure 4 , Figure 4 FIG. 1 is a main structural block diagram of a hospital infection monitoring system according to an embodiment of the present application. Figure 4 As shown, the hospital infection monitoring system in the embodiment of the present application mainly includes a sensor module 401, a processing module 402, an analysis module 403 and a fusion module 404.

[0206] In some embodiments, the sensor module 401 can be configured to process hand hygiene monitoring data and / or disinfection monitoring data to obtain effective monitoring data. The processing module 402 can be configured to process hand hygiene monitoring data and / or disinfection monitoring data to obtain effective monitoring data. The analysis module 403 can be configured to analyze the effective monitoring data to obtain an analysis result; wherein the analysis result includes a hand image, personnel identification information and / or equipment image, equipment identification information. The fusion module 404 can be configured to fuse the data in the analysis result to obtain a hospital infection monitoring result. In one embodiment, the description of the specific implementation function can be found in steps S101 to S104.

[0207] In some embodiments, the hospital infection monitoring system may further include an interactive module; the interactive module is provided with an interactive interface, and the interactive interface is configured to visualize the hospital infection monitoring results. That is, the interactive interface provides real-time monitoring data and reports, so that medical staff can timely understand their own health conditions and equipment disinfection conditions.

[0208] Furthermore, it should be understood that since the setting of each module is only for illustrating the functional units of the hospital infection monitoring system of the present application, the physical devices corresponding to these modules can be the processor itself, or a part of the software in the processor, a part of the hardware, or a part of the combination of software and hardware. Therefore, the number of each module and device in the figure is only schematic.

[0209] It is understood by those skilled in the art that the modules in the hospital infection monitoring system can be adaptively split or merged. Such splitting or merging of specific modules will not cause the technical solution to deviate from the principle of this application. Therefore, the technical solutions after splitting or merging will fall within the protection scope of this application.

[0210] The above hospital infection monitoring system is used to implement Figure 1 The hospital infection monitoring method embodiment shown in the figure has similar technical principles, technical problems solved and technical effects produced. Technical personnel in this technical field can clearly understand that for the convenience and conciseness of description, the specific working process and related instructions of the hospital infection monitoring system can refer to the contents described in the embodiment of the hospital infection monitoring method, which will not be repeated here.

[0211] It should be noted that the relevant user personal information that may be involved in the various embodiments of the present application is strictly in accordance with the requirements of laws and regulations, following the principles of legality, legitimacy and necessity, based on the reasonable purposes of business scenarios, to process the personal information that users actively provide during the use of products / services or generated due to the use of products / services, as well as the personal information obtained with the user's authorization.

[0212] This application attaches great importance to the security of user personal information and has taken reasonable and feasible security protection measures that meet industry standards to protect user information and prevent personal information from being accessed, disclosed, used, modified, damaged or lost without authorization.

[0213] So far, the technical solution of the present application has been described in conjunction with an embodiment shown in the accompanying drawings, but it is easy for those skilled in the art to understand that the protection scope of the present application is obviously not limited to these specific embodiments. Without departing from the principles of the present application, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present application.

Claims

1. A hospital infection monitoring method, characterized in that: The method comprises: Obtaining hand hygiene monitoring data of medical staff and / or disinfection monitoring data of medical equipment; Processing the hand hygiene monitoring data and / or the disinfection monitoring data to obtain effective monitoring data; Analyze the effective monitoring data to obtain analysis results; wherein the analysis results include hand images, personnel identification information and / or equipment images, equipment identification information; The data in the analysis results are integrated to obtain hospital infection monitoring results.

2. The hospital infection monitoring method according to claim 1, characterized in that: The obtaining of the medical staff's hand hygiene monitoring data and / or the medical equipment's disinfection monitoring data includes: In response to a medical staff and / or medical equipment entering a preset monitoring area, obtaining hand hygiene monitoring data of the medical staff and / or disinfection monitoring data of the medical equipment; The hand hygiene monitoring data includes at least one of first image data, pressure data and radar data; and the disinfection monitoring data includes at least one of second image data and temperature data.

3. The hospital infection monitoring method according to claim 2, characterized in that: The processing of the hand hygiene monitoring data and / or the disinfection monitoring data to obtain effective monitoring data includes: Preprocessing the hand hygiene monitoring data and / or the disinfection monitoring data to obtain standard monitoring data; The standard monitoring data is input into a convolutional neural network, and the standard monitoring data is screened and identified based on the convolutional neural network to obtain the effective monitoring data.

4. The hospital infection monitoring method according to claim 3, characterized in that: The analysis of the effective monitoring data to obtain the analysis results includes: Inputting the valid monitoring data into an artificial intelligence model, identifying the valid monitoring data based on the artificial intelligence model, and obtaining an identification result; The recognition result is evaluated based on a decision tree or random forest algorithm to obtain the analysis result.

5. The hospital infection monitoring method according to any one of claims 1 to 4, characterized in that: The hospital infection monitoring result includes compliance of monitoring data or non-compliance of monitoring data; the method further includes: When the hospital infection monitoring result shows that the monitoring data is non-compliant, an alarm signal is triggered; Wherein, the alarm signal includes at least one of an audio alarm signal, a text alarm signal and a light alarm signal.

6. The hospital infection monitoring method according to claim 3, characterized in that: The preprocessing of the hand hygiene monitoring data and / or the disinfection monitoring data includes: At least one of grayscale conversion processing, noise removal processing and edge detection processing is performed on the first image data and / or the second image data.

7. The hospital infection monitoring method according to claim 4, characterized in that: The identifying the valid monitoring data based on the artificial intelligence model includes: Identify the first image data to obtain the hand cleaning method and cleaning duration of the medical staff; and / or, identifying the pressure data to obtain water usage data and disinfectant usage data of medical staff; and / or, identifying the radar data to obtain hand movement data of the medical staff; and / or, identifying the second image data to obtain a disinfection method and disinfection duration of the medical device; And / or, identifying the temperature data to obtain the surface temperature of the medical device.

8. The hospital infection monitoring method according to claim 7, characterized in that: The evaluating the recognition result comprises: Determine whether the medical staff's hand cleaning method and cleaning duration, the water usage data and disinfectant usage data, and the hand movement data meet the specifications; and / or, Determine whether the disinfection method and disinfection duration of the medical device and the surface temperature meet preset conditions.

9. A hospital infection monitoring system, characterized in that: The system comprises: A sensor module configured to obtain hand hygiene monitoring data of medical personnel and / or disinfection monitoring data of medical equipment; a processing module, configured to process the hand hygiene monitoring data and / or the disinfection monitoring data to obtain effective monitoring data; An analysis module, configured to analyze the effective monitoring data to obtain analysis results; wherein the analysis results include hand images, personnel identification information and / or device images, device identification information; A fusion module is configured to fuse the data in the analysis results to obtain hospital infection monitoring results.

10. The hospital infection monitoring system according to claim 9, characterized in that: The hospital infection monitoring system also includes an interaction module; The interactive module is provided with an interactive interface, and the interactive interface is configured to visually display the hospital infection monitoring results.

Images