Medical instrument disinfection dosage control method and system

Abstract

The invention discloses a medical instrument disinfection dosage control method and system, and belongs to the technical field of medical instrument disinfection dosage control. According to the method, image data are collected, the instrument type is determined according to instrument contour feature points, the pollutant type and pollutant concentration of the medical instrument are determined by extracting color depth features and color features, and the optimal disinfectant combination is determined through a particle swarm algorithm; by training the disinfection dosage evaluation model, the optimal disinfection dosage evaluation model is determined, so that the disinfection dosage of the medical instrument can be accurately controlled. According to the method, the characteristics in the image data are extracted, the disinfection information required by the medical instrument is accurately represented, the disinfection dosage of the medical instrument is efficiently and accurately obtained through the trained disinfection dosage evaluation model, it is ensured that the medical instrument can be accurately disinfected, low efficiency and insufficient accuracy caused by manual disinfection are avoided, and the medical instrument disinfection efficiency is improved. The disinfection safety and efficiency of the medical instruments are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical device disinfection dosage control, and in particular relates to a medical device disinfection dosage control method and system. Background Art

[0002] As a special kind of equipment, medical devices need to be disinfected after each use or before use to ensure the safety of the equipment. In the current classification of medical devices, the accuracy of disinfection for medical devices with different uses is different, and different medical devices have certain requirements for the selection of disinfectants.

[0003] At present, the disinfection process of medical devices mainly relies on manual operation. Although this traditional method can ensure the sanitation and safety of the equipment to a certain extent, for some high-precision and high-sensitivity precision instruments, if the degree of disinfection is too large, such as using too strong disinfectants or too much disinfection dosage, it is likely to cause irreversible damage to the material, performance and even internal structure of these precision instruments, thereby affecting their accuracy and reliability in subsequent medical processes. On the contrary, if the degree of disinfection is not enough, it may not be possible to completely kill harmful microorganisms attached to the equipment, reducing the safety of medical device disinfection. In addition, the disinfection efficiency of medical devices is further reduced by manual disinfection. Therefore, there is an urgent need for a method and system for controlling the amount of disinfection of medical devices to solve the defects of the prior art. Summary of the invention

[0004] The present invention aims to provide a method and system for controlling the amount of disinfection of medical devices to solve the above-mentioned technical problems, by identifying the medical devices to be disinfected and evaluating the required disinfection amount, thereby improving the safety and efficiency of medical device disinfection.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a method for controlling the amount of medical device disinfection, comprising: Collecting image data of the medical device to be sterilized, extracting device contour feature points of the image data, constructing a three-dimensional space model of the medical device based on the device contour feature points, and determining the device type of the medical device; Extracting color depth features and color features of the image data based on the three-dimensional space model, and clustering the color depth features and color features to obtain the type and concentration of pollutants of the medical device; Obtain historical medical device disinfection data, input the pollutant type and device type of the medical device into a preset particle swarm model, and construct a fitness function of the particle swarm model in combination with the historical medical device disinfection data, iteratively solve the particle swarm model based on the fitness function, and obtain the optimal disinfectant combination for the medical device; Decomposing the historical medical device disinfection data according to a preset decomposition algorithm to obtain modal components of the historical medical device disinfection data, inputting the modal components into a preset disinfection dosage evaluation model, training the disinfection dosage evaluation model according to a preset model training algorithm, and obtaining an optimal disinfection dosage evaluation model; The optimal disinfectant combination and pollutant concentration of the medical device are input into the optimal disinfection dosage evaluation model to obtain the optimal disinfection dosage of each disinfectant in the optimal disinfectant combination to control the disinfection dosage of the medical device.

[0006] It can be understood that, compared with the prior art, the present invention collects image data of medical devices to be disinfected, determines the type of device by extracting the feature points of the device contour, and determines the type and concentration of pollutants of the medical devices by extracting the color depth feature and the color feature, and then determines the optimal disinfectant combination by the particle swarm algorithm; the optimal disinfection dosage evaluation model is determined by training the disinfection dosage evaluation model, so that the disinfection dosage of the medical devices can be accurately evaluated and controlled. The present invention obtains the characteristic information of the medical devices by extracting the features in the image data, accurately characterizes the required disinfection information of the medical devices, and efficiently and accurately obtains the disinfection dosage of the medical devices through the trained disinfection dosage evaluation model, ensuring that the medical devices can be accurately disinfected, avoiding the inefficiency and insufficient accuracy caused by manual disinfection, and improving the safety and efficiency of medical device disinfection.

[0007] As a preferred solution, the collecting of image data of the medical device to be sterilized, extracting device contour feature points of the image data, constructing a three-dimensional space model of the medical device based on the device contour feature points, and determining the device type of the medical device specifically includes: Collecting image data of the medical device to be sterilized, and removing high-frequency noise from the image data through a preset Gaussian filter; Decomposing the image data after high-frequency noise is removed by layers to obtain a plurality of first layers; Calculate a first gradient value set of pixel points in the first layer in the X-axis direction according to a preset Sobel edge detection operator; Calculate a second gradient value set of pixel points in the first layer in the Y-axis direction according to a preset Sobel edge detection operator; Determine a layer gradient value set of the first layer based on the first gradient value set and the second gradient value set, and select gradient values ​​greater than a preset gradient threshold from the layer gradient value set to determine a first gradient value set of the first layer; Determine edge points of the first layer based on pixel points corresponding to gradient values ​​in the first gradient value set, and determine feature points of the device contour of the first layer based on the edge points of the first layer; Reconstructing the first layer based on the instrument contour feature points to obtain a second layer; Converting the second layer into a voxel layer based on a preset voxel algorithm, and interpolating the voxel layer according to the feature points of the device contour to construct a three-dimensional spatial model of the medical device; The three-dimensional spatial model of the medical device is graphic matched with a preset medical device database according to a preset graphic matching algorithm to determine the device type of the medical device.

[0008] This preferred solution filters the image through a Gaussian filter to improve the clarity and quality of the image, maintain the clarity of the image edges, and provide a more accurate image basis for subsequent layer decomposition and gradient calculation; through layer decomposition and gradient value calculation, the contour of the graphic can be accurately characterized, thereby extracting accurate instrument contour feature points, and constructing a three-dimensional space model through the instrument contour feature points and performing graphic matching with the medical instrument database, which can efficiently and accurately determine the instrument type of the medical instrument, and provide an accurate data basis for the subsequent control of the amount of medical instrument disinfection.

[0009] As a preferred solution, the three-dimensional spatial model of the medical device is pattern matched with a preset medical device database according to a preset pattern matching algorithm to determine the device type of the medical device, specifically including: Acquiring feature points of a three-dimensional space model of the medical device based on the device contour feature points; A Gaussian pyramid is constructed based on the feature points, and the neighborhood gradient direction histogram of the feature points is calculated based on the Gaussian pyramid to determine the direction of the feature points; Taking the feature point as the center, slide the Gaussian pyramid through the preset window, calculate the gradient amplitude of the Gaussian pyramid in the sliding window, and determine the eigenvalue of the feature point; Construct feature descriptors of medical devices based on the directions and feature values ​​of feature points; According to a preset feature matching algorithm, the feature descriptor is matched with the device image in a preset medical device database to determine the device type of the medical device.

[0010] This preferred solution accurately captures the feature points of medical devices through device contour feature points, thereby better describing the key morphological information of medical devices, and improves the robustness and noise resistance of feature points through Gaussian pyramids; constructing feature descriptors through feature points can more accurately describe the contour shape of medical devices, thereby improving the accuracy of identifying the device type of medical devices, and providing an accurate data basis for the subsequent control of the disinfection dosage of medical devices.

[0011] As a preferred solution, the extracting of the color depth features and color features of the image data based on the three-dimensional space model, and clustering the color depth features and color features to obtain the pollutant type and pollutant concentration of the medical device specifically includes: Performing dimension conversion and color space conversion on the three-dimensional space model to determine a first color image; Extracting the number of bits of color information in the first color image to determine a color depth feature; Calculating the first-order moment, the second-order moment and the third-order moment of each color channel in the first color image, and constructing the color moment of the first color image based on the first-order moment, the second-order moment and the third-order moment to determine the color feature; Constructing color key-value pairs according to the color depth features and the color features, and clustering the color key-value pairs based on a preset clustering algorithm to obtain a number of clusters; Calculate the silhouette coefficient of each cluster, and screen the clusters based on the silhouette coefficient and the preset number to obtain several characteristic clusters; Based on the feature cluster, a feature color key-value pair is obtained, and the feature color key-value pair is matched with a preset pollutant database, the pollutant type of the medical device is determined according to the color feature, and the pollutant concentration of the medical device is determined according to the color depth feature.

[0012] This preferred solution can more accurately capture the color information of the surface of medical devices through dimensional conversion and color space conversion, and improve the accuracy of extracting color depth features and color features; by constructing color key-value pairs and performing clustering, it can discover the intrinsic relationship between color depth features and color features, thereby more accurately obtaining the type and concentration of pollutants in medical devices, and providing an accurate data basis for the subsequent control of the amount of medical device disinfection.

[0013] As a preferred solution, the method of obtaining historical medical device disinfection data, inputting the pollutant type and device type of the medical device into a preset particle swarm model, and constructing a fitness function of the particle swarm model in combination with the historical medical device disinfection data, iteratively solving the particle swarm model based on the fitness function, and obtaining the optimal disinfectant combination for the medical device specifically includes: The historical medical device disinfection data include: the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the device; Input the pollutant type and device type of the medical device into the preset particle swarm model, and initialize the initial particles of the particle swarm model in combination with the disinfectant types in the preset disinfectant database; According to the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the equipment, the fitness function of the particle swarm model is constructed, where the fitness function of the particle swarm model is specifically: ; in, is the fitness function of the particle swarm model, Types of disinfectants, is the type of device, is a collection of pollutants, Used to indicate the Disinfectants, For the Disinfectant for the Compensation factor for the loss rate of each device type; For the Disinfectant for the The attrition rate of each device type; For the Disinfectant for the The success rate of disinfection of pollutants; Used to indicate the Type of device; Used to indicate the pollutants; Iterate the initial particles of the initialized particle swarm model according to the fitness function of the particle swarm model and the preset number of iterations to obtain a particle swarm; Calculating the fitness of each particle in the particle group according to the fitness function, and selecting the particle with the smallest fitness as the optimal particle; Obtain the optimal disinfectant combination for medical devices based on optimal particles.

[0014] This preferred solution introduces historical medical device disinfection data, improves the accuracy of constructing the fitness function of the particle swarm model based on the disinfection success rate of disinfectants on pollutants and the loss rate of disinfectants on devices, and iteratively optimizes the particle swarm model based on the disinfection success rate of disinfectants on pollutants and the loss rate of disinfectants on devices. It can automatically search for the optimal disinfectant combination, avoids inaccurate and inefficient disinfection of medical devices caused by manual selection of disinfectants, and improves the safety and accuracy of medical device disinfection dosage control.

[0015] As a preferred solution, the historical medical device disinfection data is decomposed according to a preset decomposition algorithm to obtain the modal components of the historical medical device disinfection data, and the modal components are input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain the optimal disinfection dosage evaluation model, which specifically includes: The historical medical device disinfection data includes: historical disinfectant usage data and historical pollutant concentration data; Respectively extract the time series features of the historical disinfectant usage data and the historical pollutant concentration data, and construct a mapping relationship between the historical disinfectant usage data and the historical pollutant concentration data in the time dimension according to the extracted time series features to obtain a disinfection effect change sequence; Performing variational modal decomposition on the disinfection effect change sequence according to a preset variational modal decomposition algorithm to obtain modal components of the historical medical device disinfection data; The modal components are input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model.

[0016] This preferred solution reveals the dynamic changes between the amount of disinfectant used and the pollutant concentration during the disinfection process by extracting time series features and constructing a mapping relationship in the time dimension, providing an accurate data basis for the training of the disinfection dosage evaluation model; through modal decomposition, it can fully obtain the changes in different frequency components of historical medical device disinfection data, thereby improving the training efficiency and accuracy of the disinfection dosage evaluation model.

[0017] As a preferred solution, the modal component is input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model, specifically including: The preset disinfection dosage evaluation model includes: a convolutional neural network layer and a long short-term memory network layer; Inputting the modal component into the convolutional neural network layer of the preset disinfection dosage assessment model, performing convolution and maximum pooling operations on the modal component, and extracting the feature vector in the modal component; Input the feature vector into the long short-term memory network layer of the preset disinfection dosage evaluation model to extract the time series features in the feature vector; Constructing a disinfection dosage prediction value of each modal component according to the time series characteristics, and determining a dosage prediction value of each disinfectant based on the disinfection dosage prediction value of each modal component; Matching the predicted usage value of each disinfectant with the historical disinfectant usage data to evaluate the prediction accuracy of the disinfection usage evaluation model; The disinfection dosage evaluation model is iteratively trained according to the prediction accuracy until the prediction accuracy of the disinfection dosage evaluation model meets the preset requirements, the training of the disinfection dosage evaluation model is completed, and the optimal disinfection dosage evaluation model is obtained.

[0018] This preferred solution can effectively extract feature vectors from modal components through convolution and maximum pooling operations, characterize the relationship between disinfectant dosage and pollutant concentration with feature vectors, and mine the changing trend of disinfectant dosage with long and short-term memory network layers, thereby improving the training efficiency and accuracy of the disinfection dosage evaluation model and improving the safety and efficiency of medical device disinfection dosage control.

[0019] Accordingly, an embodiment of the present invention provides a medical device disinfection dosage control system, comprising: an instrument type identification module, a pollutant identification module, a disinfectant combination acquisition module, a disinfection dosage evaluation model training module and a disinfection dosage control module; The device type recognition module is used to collect image data of the medical device to be sterilized, extract device contour feature points from the image data, construct a three-dimensional space model of the medical device based on the device contour feature points, and determine the device type of the medical device; The pollutant identification module is used to extract the color depth features and color features of the image data based on the three-dimensional space model, and cluster the color depth features and color features to obtain the pollutant type and pollutant concentration of the medical device; The disinfectant combination acquisition module is used to obtain historical medical device disinfection data, input the pollutant type and device type of the medical device into a preset particle swarm model, and construct a fitness function of the particle swarm model in combination with the historical medical device disinfection data, iteratively solve the particle swarm model based on the fitness function, and obtain the optimal disinfectant combination of the medical device; The disinfection dosage evaluation model training module is used to decompose the historical medical device disinfection data according to a preset decomposition algorithm to obtain the modal components of the historical medical device disinfection data, and input the modal components into a preset disinfection dosage evaluation model, and train the disinfection dosage evaluation model according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model; The disinfection dosage control module is used to input the optimal disinfectant combination and pollutant concentration of the medical device into the optimal disinfection dosage evaluation model to obtain the optimal disinfection dosage of each disinfectant in the optimal disinfectant combination to control the disinfection dosage of the medical device.

[0020] As a preferred solution, the device type identification module includes: a device type identification unit; The device type identification unit is used to collect image data of the medical device to be sterilized, and remove high-frequency noise from the image data through a preset Gaussian filter; Decomposing the image data after high-frequency noise is removed by layers to obtain a plurality of first layers; Calculate a first gradient value set of pixel points in the first layer in the X-axis direction according to a preset Sobel edge detection operator; Calculate a second gradient value set of pixel points in the first layer in the Y-axis direction according to a preset Sobel edge detection operator; Determine a layer gradient value set of the first layer based on the first gradient value set and the second gradient value set, and select gradient values ​​greater than a preset gradient threshold from the layer gradient value set to determine a first gradient value set of the first layer; Determine edge points of the first layer based on pixel points corresponding to gradient values ​​in the first gradient value set, and determine feature points of the device contour of the first layer based on the edge points of the first layer; Reconstructing the first layer based on the instrument contour feature points to obtain a second layer; Converting the second layer into a voxel layer based on a preset voxel algorithm, and interpolating the voxel layer according to the feature points of the device contour to construct a three-dimensional spatial model of the medical device; The three-dimensional spatial model of the medical device is graphic matched with a preset medical device database according to a preset graphic matching algorithm to determine the device type of the medical device.

[0021] As a preferred solution, the pollutant identification module includes: a pollutant identification unit; The pollutant identification unit is used to perform dimension conversion and color space conversion on the three-dimensional space model to determine a first color image; Extracting the number of bits of color information in the first color image to determine a color depth feature; Calculating the first-order moment, the second-order moment and the third-order moment of each color channel in the first color image, and constructing the color moment of the first color image based on the first-order moment, the second-order moment and the third-order moment to determine the color feature; Constructing color key-value pairs according to the color depth features and the color features, and clustering the color key-value pairs based on a preset clustering algorithm to obtain a number of clusters; Calculate the silhouette coefficient of each cluster, and screen the clusters based on the silhouette coefficient and the preset number to obtain several characteristic clusters; Based on the feature cluster, a feature color key-value pair is obtained, and the feature color key-value pair is matched with a preset pollutant database, the pollutant type of the medical device is determined according to the color feature, and the pollutant concentration of the medical device is determined according to the color depth feature.

[0022] It can be understood that, compared with the prior art, this system collects image data of medical devices to be disinfected, determines the type of device by extracting the feature points of the device contour, and determines the type and concentration of pollutants in the toxic medical devices by extracting color depth features and color features, and then determines the optimal disinfectant combination through the particle swarm algorithm; by training the disinfection dosage evaluation model, the optimal disinfection dosage evaluation model is determined, so that the disinfection dosage of medical devices can be accurately evaluated and controlled. This system obtains the characteristic information of medical devices by extracting features from image data, accurately characterizes the required disinfection information of medical devices, and efficiently and accurately obtains the disinfection dosage of medical devices through the trained disinfection dosage evaluation model, ensuring that medical devices can be accurately disinfected, avoiding the inefficiency and lack of accuracy caused by manual disinfection, and improving the safety and efficiency of medical device disinfection. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 : A flow chart of steps of a method for controlling the amount of medical equipment disinfection provided by an embodiment of the present invention; Figure 2 : A schematic structural diagram of a medical device disinfection dosage control system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] Embodiment 1 Please refer to Figure 1 , is a step flow chart of a method for controlling the amount of medical device disinfection provided by an embodiment of the present invention, including steps S101 to S105.

[0026] Step S101: collecting image data of a medical device to be sterilized, extracting device contour feature points from the image data, constructing a three-dimensional space model of the medical device based on the device contour feature points, and determining the device type of the medical device.

[0027] In this embodiment, the collecting of image data of the medical device to be sterilized, extracting device contour feature points of the image data, constructing a three-dimensional space model of the medical device based on the device contour feature points, and determining the device type of the medical device specifically includes: Collecting image data of the medical device to be sterilized, and removing high-frequency noise from the image data through a preset Gaussian filter; Decomposing the image data after high-frequency noise is removed by layers to obtain a plurality of first layers; Calculate a first gradient value set of pixel points in the first layer in the X-axis direction according to a preset Sobel edge detection operator; Calculate a second gradient value set of pixel points in the first layer in the Y-axis direction according to a preset Sobel edge detection operator; Determine a layer gradient value set of the first layer based on the first gradient value set and the second gradient value set, and select gradient values ​​greater than a preset gradient threshold from the layer gradient value set to determine a first gradient value set of the first layer; Determine edge points of the first layer based on pixel points corresponding to gradient values ​​in the first gradient value set, and determine feature points of the device contour of the first layer based on the edge points of the first layer; Reconstructing the first layer based on the instrument contour feature points to obtain a second layer; Converting the second layer into a voxel layer based on a preset voxel algorithm, and interpolating the voxel layer according to the feature points of the device contour to construct a three-dimensional spatial model of the medical device; The three-dimensional spatial model of the medical device is graphic matched with a preset medical device database according to a preset graphic matching algorithm to determine the device type of the medical device.

[0028] It should be noted that the Gaussian filter is a low-pass filter, which is used to perform weighted averaging on each pixel in the image and the pixels in its neighborhood, thereby reducing the high-frequency noise of the image and achieving the effect of smoothing the image; in addition to the Gaussian filter described in this embodiment, the image can also be filtered by low-pass filters such as a mean filter, a Butterworth filter and a Chebyshev filter.

[0029] In an optional embodiment, the Sobel edge detection operator is a discrete differential operator for image edge detection. The first layer is converted into a two-dimensional coordinate system. , where Represents the pixel points in the first layer The gray value of .

[0030] The pixel The pixel to its right Subtract the gray value of the pixel point The gradient value in the X-axis direction, Right now , in, Pixel The gradient value in the X-axis direction, Pixel The gray value of Pixel Gray value of The gradient value of each pixel in the X-axis direction is counted, and the gradient matrix of the first layer in the X-axis direction is constructed based on the two-dimensional coordinate system. Then, the gradient matrix is ​​convolved through the Sobel edge detection operator (Sobel operator), and all elements of the gradient matrix after the convolution operation are extracted to obtain the first gradient value set of the pixel points in the first layer in the X-axis direction.

[0031] The pixel The pixel above it Subtract the gray value of the pixel point The gradient value in the Y-axis direction, Right now , in, Pixel The gradient value in the Y-axis direction, Pixel The gray value of Pixel Gray value of The gradient value of each pixel in the Y-axis direction is counted, and the gradient matrix of the first layer in the Y-axis direction is constructed based on the two-dimensional coordinate system. Then, the gradient matrix is ​​convolved through the Sobel edge detection operator (Sobel operator), and all elements of the gradient matrix after the convolution operation are extracted to obtain the second gradient value set of the pixel points in the first layer in the Y-axis direction.

[0032] In an optional embodiment, the preset gradient threshold includes a gradient threshold in the X-axis direction and a gradient threshold in the Y-axis direction, and the gradient values ​​in the layer gradient value set that simultaneously satisfy the gradient threshold greater than the gradient threshold in the X-axis direction and the gradient threshold in the Y-axis direction are screened to determine the first gradient value set of the first layer.

[0033] It should be noted that the voxel algorithm is a three-dimensional data processing method, which involves three-dimensional space data processing. It is based on voxels, the smallest volume unit in three-dimensional space, to represent, analyze and process three-dimensional objects or scenes.

[0034] This embodiment performs filtering processing on the image through a Gaussian filter to improve the clarity and quality of the image, maintain the clarity of the image edge, and provide a more accurate image basis for subsequent layer decomposition and gradient calculation; through layer decomposition and gradient value calculation, the contour of the graphic can be accurately characterized, thereby extracting accurate instrument contour feature points, and constructing a three-dimensional space model through the instrument contour feature points and performing graphic matching with a medical instrument database, which can efficiently and accurately determine the instrument type of the medical instrument, and provide an accurate data basis for the subsequent control of the amount of medical instrument disinfection.

[0035] In this embodiment, the three-dimensional spatial model of the medical device is pattern matched with a preset medical device database according to a preset pattern matching algorithm to determine the device type of the medical device, specifically including: Acquiring feature points of a three-dimensional space model of the medical device based on the device contour feature points; A Gaussian pyramid is constructed based on the feature points, and the neighborhood gradient direction histogram of the feature points is calculated based on the Gaussian pyramid to determine the direction of the feature points; Taking the feature point as the center, slide the Gaussian pyramid through the preset window, calculate the gradient amplitude of the Gaussian pyramid in the sliding window, and determine the eigenvalue of the feature point; Construct feature descriptors of medical devices based on the directions and feature values ​​of feature points; According to a preset feature matching algorithm, the feature descriptor is matched with the device image in a preset medical device database to determine the device type of the medical device.

[0036] In an optional embodiment, Gaussian Pyramid is a commonly used technology in image processing. With the feature point as the center, the neighborhood is set as a neighborhood window of size 3×4, and the gradient amplitude (i.e., gradient amplitude) and gradient direction of each pixel point in the neighborhood window are calculated. The gradient amplitude can be determined by the gray value of the pixel point, and the gradient direction can be determined by the direction of the largest gray value change in the neighborhood window; the neighborhood gradient direction histogram is constructed by the gradient amplitude and gradient direction, and the peak value in the neighborhood gradient direction histogram is extracted. The direction of the feature point is determined based on the gradient direction corresponding to the peak value, and the characteristic value of the feature point is determined based on the gradient amplitude corresponding to the peak value.

[0037] This embodiment accurately captures the feature points of medical devices through device contour feature points, thereby better describing the key morphological information of the medical devices, and improves the robustness and noise resistance of the feature points through the Gaussian pyramid; by constructing feature descriptors through feature points, it is possible to more accurately describe the contour shape of the medical device, thereby improving the accuracy of identifying the device type of the medical device, and providing an accurate data basis for the subsequent control of the disinfection dosage of the medical device.

[0038] Step S102: extracting the color depth features and color features of the image data based on the three-dimensional space model, and clustering the color depth features and color features to obtain the pollutant type and pollutant concentration of the medical device.

[0039] In this embodiment, the color depth features and color features of the image data are extracted based on the three-dimensional space model, and the color depth features and color features are clustered to obtain the pollutant type and pollutant concentration of the medical device, specifically including: Performing dimension conversion and color space conversion on the three-dimensional space model to determine a first color image; Extracting the number of bits of color information in the first color image to determine a color depth feature; Calculating the first-order moment, the second-order moment and the third-order moment of each color channel in the first color image, and constructing the color moment of the first color image based on the first-order moment, the second-order moment and the third-order moment to determine the color feature; Constructing color key-value pairs according to the color depth features and the color features, and clustering the color key-value pairs based on a preset clustering algorithm to obtain a number of clusters; Calculate the silhouette coefficient of each cluster, and screen the clusters based on the silhouette coefficient and the preset number to obtain several characteristic clusters; Based on the feature cluster, a feature color key-value pair is obtained, and the feature color key-value pair is matched with a preset pollutant database, the pollutant type of the medical device is determined according to the color feature, and the pollutant concentration of the medical device is determined according to the color depth feature.

[0040] In an optional embodiment, the three-dimensional space model is mapped to a two-dimensional image, and the two-dimensional graphic is converted from the RGB color space to the Lab color space to determine a first color image; the number of bits of color information in the first color image is extracted to determine a color distribution histogram of the first color image, and statistical features are extracted from the color distribution histogram, the statistical features including: mean, standard deviation, kurtosis and skewness, to determine color depth features.

[0041] The first-order moment (mean), second-order moment (variance), and third-order moment (skewness) of each color channel in the first color image are calculated, and the color moment of the first color image is constructed based on the first-order moment, second-order moment, and third-order moment, a color feature vector is determined based on the color moment, and a color feature is obtained through the color feature vector.

[0042] It should be noted that the clustering effect of each cluster is determined by calculating the Silhouette Coefficient. Ten clusters are screened out after being sorted from large to small according to the Silhouette Coefficient, and 10 feature clusters are obtained. Then, the feature color key-value pairs (i.e., color key-value pairs) of the feature clusters are matched with the preset pollutant database. The pollutant type of the medical device is determined according to the color characteristics, and the pollutant concentration of the medical device is determined according to the color depth characteristics.

[0043] This embodiment can more accurately capture the color information of the surface of medical devices through dimensional conversion and color space conversion, and improve the accuracy of extracting color depth features and color features; by constructing color key-value pairs and performing clustering, it is possible to discover the intrinsic relationship between color depth features and color features, thereby more accurately obtaining the type and concentration of pollutants in medical devices, and providing an accurate data basis for the subsequent control of the amount of medical device disinfection.

[0044] Step S103: Obtain historical medical device disinfection data, input the pollutant type and device type of the medical device into a preset particle swarm model, and construct a fitness function of the particle swarm model in combination with the historical medical device disinfection data, iteratively solve the particle swarm model based on the fitness function, and obtain the optimal disinfectant combination for the medical device.

[0045] In this embodiment, the historical medical device disinfection data is obtained, the pollutant type and device type of the medical device are input into a preset particle swarm model, and the fitness function of the particle swarm model is constructed in combination with the historical medical device disinfection data. The particle swarm model is iteratively solved based on the fitness function to obtain the optimal disinfectant combination for the medical device, specifically including: The historical medical device disinfection data include: the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the device; Input the pollutant type and device type of the medical device into the preset particle swarm model, and initialize the initial particles of the particle swarm model in combination with the disinfectant types in the preset disinfectant database; According to the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the equipment, the fitness function of the particle swarm model is constructed, where the fitness function of the particle swarm model is specifically: ; in, is the fitness function of the particle swarm model, Types of disinfectants, is the type of device, is a collection of pollutants, Used to indicate the Disinfectants, For the Disinfectant for the Compensation factor for the loss rate of each device type; For the Disinfectant for the The attrition rate of each device type; For the Disinfectant for the The success rate of disinfection of pollutants; Used to indicate the Type of device; Used to indicate the pollutants; Iterate the initial particles of the initialized particle swarm model according to the fitness function of the particle swarm model and the preset number of iterations to obtain a particle swarm; Calculating the fitness of each particle in the particle group according to the fitness function, and selecting the particle with the smallest fitness as the optimal particle; Obtain the optimal disinfectant combination for medical devices based on optimal particles.

[0046] It should be noted that the particle swarm model (also known as particle swarm optimization, PSO for short) is a model that simulates the group behavior of nature to find the optimal solution to the problem; in the particle swarm model, the fitness function is used to evaluate the performance of each particle in the solution space, and the actual solution effect of each particle can be obtained through the fitness function. Each particle represents a disinfectant combination, and a disinfectant combination includes several disinfectants.

[0047] This embodiment introduces historical medical device disinfection data, improves the accuracy of constructing the fitness function of the particle swarm model based on the disinfection success rate of the disinfectant to the pollutants and the loss rate of the disinfectant to the equipment, and iteratively optimizes the particle swarm model based on the disinfection success rate of the disinfectant to the pollutants and the loss rate of the disinfectant to the equipment, so as to automatically search for the optimal disinfectant combination, avoid inaccurate and inefficient disinfection of medical devices caused by manual selection of disinfectants, and improve the safety and accuracy of medical device disinfection dosage control.

[0048] Step S104: Decompose the historical medical device disinfection data according to a preset decomposition algorithm to obtain modal components of the historical medical device disinfection data, input the modal components into a preset disinfection dosage evaluation model, train the disinfection dosage evaluation model according to a preset model training algorithm, and obtain an optimal disinfection dosage evaluation model.

[0049] In this embodiment, the historical medical device disinfection data is decomposed according to a preset decomposition algorithm to obtain modal components of the historical medical device disinfection data, and the modal components are input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model, specifically including: The historical medical device disinfection data includes: historical disinfectant usage data and historical pollutant concentration data; Respectively extract the time series features of the historical disinfectant usage data and the historical pollutant concentration data, and construct a mapping relationship between the historical disinfectant usage data and the historical pollutant concentration data in the time dimension according to the extracted time series features to obtain a disinfection effect change sequence; Performing variational modal decomposition on the disinfection effect change sequence according to a preset variational modal decomposition algorithm to obtain modal components of the historical medical device disinfection data; The modal components are input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model.

[0050] It should be noted that the variational mode decomposition algorithm (VMD) is an algorithm used to decompose time series data into a series of intrinsic mode functions (IMF) with limited bandwidth.

[0051] This embodiment reveals the dynamic change law between the amount of disinfectant used and the pollutant concentration during the disinfection process by extracting time series features and constructing a mapping relationship in the time dimension, providing an accurate data basis for the training of the disinfection dosage evaluation model; through modal decomposition, it is possible to fully obtain the changes in different frequency components of historical medical device disinfection data, thereby improving the training efficiency and accuracy of the disinfection dosage evaluation model.

[0052] In this embodiment, the modal component is input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model, specifically including: The preset disinfection dosage evaluation model includes: a convolutional neural network layer and a long short-term memory network layer; Inputting the modal component into the convolutional neural network layer of the preset disinfection dosage assessment model, performing convolution and maximum pooling operations on the modal component, and extracting the feature vector in the modal component; Input the feature vector into the long short-term memory network layer of the preset disinfection dosage evaluation model to extract the time series features in the feature vector; Constructing a disinfection dosage prediction value of each modal component according to the time series characteristics, and determining a dosage prediction value of each disinfectant based on the disinfection dosage prediction value of each modal component; Matching the predicted usage value of each disinfectant with the historical disinfectant usage data to evaluate the prediction accuracy of the disinfection usage evaluation model; The disinfection dosage evaluation model is iteratively trained according to the prediction accuracy until the prediction accuracy of the disinfection dosage evaluation model meets the preset requirements, the training of the disinfection dosage evaluation model is completed, and the optimal disinfection dosage evaluation model is obtained.

[0053] It should be noted that the convolutional neural network layer (CNN Layer) described in this embodiment includes two convolution units, two pooling units and a fully connected unit, and the long short-term memory network layer (LSTM Layer) includes an input gate, a forget gate and an output gate. At present, there are many studies on the fusion of the convolutional neural network layer and the long short-term memory network layer into an evaluation model, and the present invention will not go into details here.

[0054] Through convolution and maximum pooling operations, this embodiment can effectively extract feature vectors from modal components, characterize the relationship between disinfectant dosage and pollutant concentration with feature vectors, and mine the changing trend of disinfectant dosage with long and short-term memory network layers, thereby improving the training efficiency and accuracy of the disinfection dosage evaluation model and improving the safety and efficiency of medical device disinfection dosage control.

[0055] Step S105: Input the optimal disinfectant combination and pollutant concentration of the medical device into the optimal disinfection dosage evaluation model to obtain the optimal disinfection dosage of each disinfectant in the optimal disinfectant combination to control the disinfection dosage of the medical device.

[0056] This embodiment collects image data of medical devices to be disinfected, determines the type of device by extracting the feature points of the device contour, and determines the type and concentration of pollutants of the medical device by extracting the color depth feature and the color feature, and then determines the optimal disinfectant combination through the particle swarm algorithm; the optimal disinfection dosage evaluation model is determined by training the disinfection dosage evaluation model, so that the disinfection dosage of the medical device can be accurately evaluated and controlled. This embodiment obtains the characteristic information of the medical device by extracting the features in the image data, accurately characterizes the required disinfection information of the medical device, and efficiently and accurately obtains the disinfection dosage of the medical device through the trained disinfection dosage evaluation model, ensuring that the medical device can be accurately disinfected, avoiding the inefficiency and insufficient accuracy caused by manual disinfection, and improving the safety and efficiency of medical device disinfection.

[0057] Embodiment 2 Please refer to Figure 2, which is a structural schematic diagram of a medical device disinfection dosage control system provided in an embodiment of the present invention, including: an instrument type identification module 201, a pollutant identification module 202, a disinfectant combination acquisition module 203, a disinfection dosage evaluation model training module 204 and a disinfection dosage control module 205.

[0058] The device type identification module 201 is used to collect image data of the medical device to be sterilized, extract device contour feature points from the image data, construct a three-dimensional space model of the medical device based on the device contour feature points, and determine the device type of the medical device.

[0059] In this embodiment, the device type identification module 201 includes: a device type identification unit; The device type identification unit is used to collect image data of the medical device to be sterilized, and remove high-frequency noise from the image data through a preset Gaussian filter; Decomposing the image data after high-frequency noise is removed by layers to obtain a plurality of first layers; Calculate a first gradient value set of pixel points in the first layer in the X-axis direction according to a preset Sobel edge detection operator; Calculate a second gradient value set of pixel points in the first layer in the Y-axis direction according to a preset Sobel edge detection operator; Determine a layer gradient value set of the first layer based on the first gradient value set and the second gradient value set, and select gradient values ​​greater than a preset gradient threshold from the layer gradient value set to determine a first gradient value set of the first layer; Determine edge points of the first layer based on pixel points corresponding to gradient values ​​in the first gradient value set, and determine feature points of the device contour of the first layer based on the edge points of the first layer; Reconstructing the first layer based on the instrument contour feature points to obtain a second layer; Converting the second layer into a voxel layer based on a preset voxel algorithm, and interpolating the voxel layer according to the feature points of the device contour to construct a three-dimensional spatial model of the medical device; The three-dimensional spatial model of the medical device is graphic matched with a preset medical device database according to a preset graphic matching algorithm to determine the device type of the medical device.

[0060] In this embodiment, the device type identification unit includes: a graphic matching subunit; The graphic matching subunit is used to obtain feature points of the three-dimensional space model of the medical device based on the device contour feature points; A Gaussian pyramid is constructed based on the feature points, and the neighborhood gradient direction histogram of the feature points is calculated based on the Gaussian pyramid to determine the direction of the feature points; Taking the feature point as the center, slide the Gaussian pyramid through the preset window, calculate the gradient amplitude of the Gaussian pyramid in the sliding window, and determine the eigenvalue of the feature point; Construct feature descriptors of medical devices based on the directions and feature values ​​of feature points; According to a preset feature matching algorithm, the feature descriptor is matched with the device image in a preset medical device database to determine the device type of the medical device.

[0061] The pollutant identification module 202 is used to extract the color depth features and color features of the image data based on the three-dimensional space model, and cluster the color depth features and color features to obtain the pollutant type and pollutant concentration of the medical device.

[0062] In this embodiment, the pollutant identification module 202 includes: a pollutant identification unit; The pollutant identification unit is used to perform dimension conversion and color space conversion on the three-dimensional space model to determine a first color image; Extracting the number of bits of color information in the first color image to determine a color depth feature; Calculating the first-order moment, the second-order moment and the third-order moment of each color channel in the first color image, and constructing the color moment of the first color image based on the first-order moment, the second-order moment and the third-order moment to determine the color feature; Constructing color key-value pairs according to the color depth features and the color features, and clustering the color key-value pairs based on a preset clustering algorithm to obtain a number of clusters; Calculate the silhouette coefficient of each cluster, and screen the clusters based on the silhouette coefficient and the preset number to obtain several characteristic clusters; Based on the feature cluster, a feature color key-value pair is obtained, and the feature color key-value pair is matched with a preset pollutant database, the pollutant type of the medical device is determined according to the color feature, and the pollutant concentration of the medical device is determined according to the color depth feature.

[0063] The disinfectant combination acquisition module 203 is used to obtain historical medical device disinfection data, input the pollutant type and device type of the medical device into a preset particle swarm model, and construct a fitness function of the particle swarm model in combination with the historical medical device disinfection data, and iteratively solve the particle swarm model based on the fitness function to obtain the optimal disinfectant combination for the medical device.

[0064] In this embodiment, the disinfectant combination acquisition module 203 includes: a disinfectant combination acquisition unit; The disinfectant combination obtains the historical medical device disinfection data including: the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the device; Input the pollutant type and device type of the medical device into the preset particle swarm model, and initialize the initial particles of the particle swarm model in combination with the disinfectant types in the preset disinfectant database; According to the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the equipment, the fitness function of the particle swarm model is constructed, where the fitness function of the particle swarm model is specifically: , in, is the fitness function of the particle swarm model, Types of disinfectants, is the type of device, is a collection of pollutants, Used to indicate the Disinfectants, For the Disinfectant for the Compensation factor for the loss rate of each device type; For the Disinfectant for the The attrition rate of each device type; For the Disinfectant for the The success rate of disinfection of pollutants; Used to indicate the Type of device; Used to indicate the pollutants; Iterate the initial particles of the initialized particle swarm model according to the fitness function of the particle swarm model and the preset number of iterations to obtain a particle swarm; Calculating the fitness of each particle in the particle group according to the fitness function, and selecting the particle with the smallest fitness as the optimal particle; Obtain the optimal disinfectant combination for medical devices based on optimal particles.

[0065] The disinfection dosage assessment model training module 204 is used to decompose the historical medical device disinfection data according to a preset decomposition algorithm to obtain the modal components of the historical medical device disinfection data, and input the modal components into a preset disinfection dosage assessment model, and train the disinfection dosage assessment model according to a preset model training algorithm to obtain an optimal disinfection dosage assessment model.

[0066] In this embodiment, the disinfection dosage assessment model training module 204 includes: a disinfection dosage assessment model training unit; The disinfection dosage assessment model training unit is used for the historical medical device disinfection data including: historical disinfectant dosage data and historical pollutant concentration data; Respectively extract the time series features of the historical disinfectant usage data and the historical pollutant concentration data, and construct a mapping relationship between the historical disinfectant usage data and the historical pollutant concentration data in the time dimension according to the extracted time series features to obtain a disinfection effect change sequence; Performing variational modal decomposition on the disinfection effect change sequence according to a preset variational modal decomposition algorithm to obtain modal components of the historical medical device disinfection data; The modal components are input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model.

[0067] In this embodiment, the disinfection dosage assessment model training unit includes: a disinfection dosage assessment model training subunit; The disinfection dosage evaluation model training subunit is used for the preset disinfection dosage evaluation model and includes: a convolutional neural network layer and a long short-term memory network layer; Inputting the modal component into the convolutional neural network layer of the preset disinfection dosage assessment model, performing convolution and maximum pooling operations on the modal component, and extracting the feature vector in the modal component; Input the feature vector into the long short-term memory network layer of the preset disinfection dosage evaluation model to extract the time series features in the feature vector; Constructing a disinfection dosage prediction value of each modal component according to the time series characteristics, and determining a dosage prediction value of each disinfectant based on the disinfection dosage prediction value of each modal component; Matching the predicted usage value of each disinfectant with the historical disinfectant usage data to evaluate the prediction accuracy of the disinfection usage evaluation model; The disinfection dosage evaluation model is iteratively trained according to the prediction accuracy until the prediction accuracy of the disinfection dosage evaluation model meets the preset requirements, the training of the disinfection dosage evaluation model is completed, and the optimal disinfection dosage evaluation model is obtained.

[0068] The disinfection dosage control module 205 is used to input the optimal disinfectant combination and pollutant concentration of the medical device into the optimal disinfection dosage evaluation model to obtain the optimal disinfection dosage of each disinfectant in the optimal disinfectant combination to control the disinfection dosage of the medical device.

[0069] This embodiment collects image data of medical devices to be disinfected, determines the type of device by extracting the feature points of the device contour, and determines the type and concentration of pollutants of the medical device by extracting the color depth feature and the color feature, and then determines the optimal disinfectant combination through the particle swarm algorithm; the optimal disinfection dosage evaluation model is determined by training the disinfection dosage evaluation model, so that the disinfection dosage of the medical device can be accurately evaluated and controlled. This embodiment obtains the characteristic information of the medical device by extracting the features in the image data, accurately characterizes the required disinfection information of the medical device, and efficiently and accurately obtains the disinfection dosage of the medical device through the trained disinfection dosage evaluation model, ensuring that the medical device can be accurately disinfected, avoiding the inefficiency and insufficient accuracy caused by manual disinfection, and improving the safety and efficiency of medical device disinfection.

[0070] In summary, the embodiment of the present invention collects image data of medical devices to be disinfected, determines the type of device by extracting the feature points of the device contour, and determines the type and concentration of pollutants of the medical device by extracting color depth features and color features, and then determines the optimal disinfectant combination through the particle swarm algorithm; the optimal disinfection dosage evaluation model is determined by training the disinfection dosage evaluation model, so that the disinfection dosage of the medical device can be accurately evaluated and controlled. The embodiment of the present invention obtains the characteristic information of the medical device by extracting features from the image data, accurately characterizes the required disinfection information of the medical device, and efficiently and accurately obtains the disinfection dosage of the medical device through the trained disinfection dosage evaluation model, ensuring that the medical device can be accurately disinfected, avoiding the inefficiency and insufficient accuracy caused by manual disinfection, and improving the safety and efficiency of medical device disinfection.

[0071] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for controlling the amount of medical equipment disinfection, characterized in that: include: Collecting image data of the medical device to be sterilized, extracting device contour feature points of the image data, constructing a three-dimensional space model of the medical device based on the device contour feature points, and determining the device type of the medical device; Extracting color depth features and color features of the image data based on the three-dimensional space model, and clustering the color depth features and color features to obtain the type and concentration of pollutants of the medical device; Obtain historical medical device disinfection data, input the pollutant type and device type of the medical device into a preset particle swarm model, and construct a fitness function of the particle swarm model in combination with the historical medical device disinfection data, iteratively solve the particle swarm model based on the fitness function, and obtain the optimal disinfectant combination for the medical device; Decomposing the historical medical device disinfection data according to a preset decomposition algorithm to obtain modal components of the historical medical device disinfection data, inputting the modal components into a preset disinfection dosage evaluation model, training the disinfection dosage evaluation model according to a preset model training algorithm, and obtaining an optimal disinfection dosage evaluation model; The optimal disinfectant combination and pollutant concentration of the medical device are input into the optimal disinfection dosage evaluation model to obtain the optimal disinfection dosage of each disinfectant in the optimal disinfectant combination to control the disinfection dosage of the medical device.

2. A method for controlling the amount of medical equipment disinfection as claimed in claim 1, characterized in that: The collecting of image data of the medical device to be sterilized, extracting device contour feature points of the image data, constructing a three-dimensional space model of the medical device based on the device contour feature points, and determining the device type of the medical device specifically includes: Collecting image data of the medical device to be sterilized, and removing high-frequency noise from the image data through a preset Gaussian filter; Decomposing the image data after high-frequency noise is removed by layers to obtain a plurality of first layers; Calculate a first gradient value set of pixel points in the first layer in the X-axis direction according to a preset Sobel edge detection operator; Calculate a second gradient value set of pixel points in the first layer in the Y-axis direction according to a preset Sobel edge detection operator; Determine a layer gradient value set of the first layer based on the first gradient value set and the second gradient value set, and select gradient values ​​greater than a preset gradient threshold from the layer gradient value set to determine a first gradient value set of the first layer; Determine edge points of the first layer based on pixel points corresponding to gradient values ​​in the first gradient value set, and determine feature points of the device contour of the first layer based on the edge points of the first layer; Reconstructing the first layer based on the instrument contour feature points to obtain a second layer; Converting the second layer into a voxel layer based on a preset voxel algorithm, and interpolating the voxel layer according to the feature points of the device contour to construct a three-dimensional spatial model of the medical device; The three-dimensional spatial model of the medical device is graphic matched with a preset medical device database according to a preset graphic matching algorithm to determine the device type of the medical device.

3. A method for controlling the amount of medical equipment disinfection as claimed in claim 2, characterized in that: The method of performing graphic matching on the three-dimensional space model of the medical device and a preset medical device database according to a preset graphic matching algorithm to determine the device type of the medical device specifically includes: Acquiring feature points of a three-dimensional space model of the medical device based on the device contour feature points; A Gaussian pyramid is constructed based on the feature points, and the neighborhood gradient direction histogram of the feature points is calculated based on the Gaussian pyramid to determine the direction of the feature points; Taking the feature point as the center, slide the Gaussian pyramid through the preset window, calculate the gradient amplitude of the Gaussian pyramid in the sliding window, and determine the eigenvalue of the feature point; Construct feature descriptors of medical devices based on the directions and feature values ​​of feature points; According to a preset feature matching algorithm, the feature descriptor is matched with the device image in a preset medical device database to determine the device type of the medical device.

4. A method for controlling the amount of medical equipment disinfection as claimed in claim 1, characterized in that: The extracting the color depth features and color features of the image data based on the three-dimensional space model, and clustering the color depth features and color features to obtain the pollutant type and pollutant concentration of the medical device specifically includes: Performing dimension conversion and color space conversion on the three-dimensional space model to determine a first color image; Extracting the number of bits of color information in the first color image to determine a color depth feature; Calculating the first-order moment, the second-order moment and the third-order moment of each color channel in the first color image, and constructing the color moment of the first color image based on the first-order moment, the second-order moment and the third-order moment to determine the color feature; Constructing color key-value pairs according to the color depth features and the color features, and clustering the color key-value pairs based on a preset clustering algorithm to obtain a number of clusters; Calculate the silhouette coefficient of each cluster, and screen the clusters based on the silhouette coefficient and the preset number to obtain several characteristic clusters; Based on the feature cluster, a feature color key-value pair is obtained, and the feature color key-value pair is matched with a preset pollutant database, the pollutant type of the medical device is determined according to the color feature, and the pollutant concentration of the medical device is determined according to the color depth feature.

5. A method for controlling the amount of medical equipment disinfection as claimed in claim 1, characterized in that: The step of obtaining historical medical device disinfection data, inputting the pollutant type and device type of the medical device into a preset particle swarm model, and constructing a fitness function of the particle swarm model in combination with the historical medical device disinfection data, and iteratively solving the particle swarm model based on the fitness function to obtain the optimal disinfectant combination for the medical device specifically includes: The historical medical device disinfection data include: the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the device; Input the pollutant type and device type of the medical device into the preset particle swarm model, and initialize the initial particles of the particle swarm model in combination with the disinfectant types in the preset disinfectant database; According to the disinfection success rate of the disinfectant on the pollutants and the loss rate of the disinfectant on the equipment, the fitness function of the particle swarm model is constructed, where the fitness function of the particle swarm model is specifically: ； in, is the fitness function of the particle swarm model, Types of disinfectants, is the type of device, is a collection of pollutants, Used to indicate the Disinfectants, For the Disinfectant for the Compensation factor for the loss rate of each device type; For the Disinfectant for the The attrition rate of each device type; For the Disinfectant for the The success rate of disinfection of pollutants; Used to indicate the Type of device; Used to indicate the pollutants; Iterate the initial particles of the initialized particle swarm model according to the fitness function of the particle swarm model and the preset number of iterations to obtain a particle swarm; Calculating the fitness of each particle in the particle group according to the fitness function, and selecting the particle with the smallest fitness as the optimal particle; Obtain the optimal disinfectant combination for medical devices based on optimal particles.

6. A method for controlling the amount of medical equipment disinfection as claimed in claim 1, characterized in that: The historical medical device disinfection data is decomposed according to a preset decomposition algorithm to obtain modal components of the historical medical device disinfection data, and the modal components are input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model, specifically including: The historical medical device disinfection data includes: historical disinfectant usage data and historical pollutant concentration data; Respectively extract the time series features of the historical disinfectant usage data and the historical pollutant concentration data, and construct a mapping relationship between the historical disinfectant usage data and the historical pollutant concentration data in the time dimension according to the extracted time series features to obtain a disinfection effect change sequence; Performing variational modal decomposition on the disinfection effect change sequence according to a preset variational modal decomposition algorithm to obtain modal components of the historical medical device disinfection data; The modal components are input into a preset disinfection dosage evaluation model, and the disinfection dosage evaluation model is trained according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model.

7. A method for controlling the amount of medical equipment disinfection as claimed in claim 6, characterized in that: The step of inputting the modal component into a preset disinfection dosage evaluation model, training the disinfection dosage evaluation model according to a preset model training algorithm, and obtaining an optimal disinfection dosage evaluation model specifically includes: The preset disinfection dosage evaluation model includes: a convolutional neural network layer and a long short-term memory network layer; Inputting the modal component into the convolutional neural network layer of the preset disinfection dosage assessment model, performing convolution and maximum pooling operations on the modal component, and extracting the feature vector in the modal component; Input the feature vector into the long short-term memory network layer of the preset disinfection dosage evaluation model to extract the time series features in the feature vector; Constructing a disinfection dosage prediction value of each modal component according to the time series characteristics, and determining a dosage prediction value of each disinfectant based on the disinfection dosage prediction value of each modal component; Matching the predicted usage value of each disinfectant with the historical disinfectant usage data to evaluate the prediction accuracy of the disinfection usage evaluation model; The disinfection dosage evaluation model is iteratively trained according to the prediction accuracy until the prediction accuracy of the disinfection dosage evaluation model meets the preset requirements, the training of the disinfection dosage evaluation model is completed, and the optimal disinfection dosage evaluation model is obtained.

8. A medical device disinfection dosage control system, characterized in that: include: Equipment type identification module, contaminant identification module, disinfectant combination acquisition module, disinfection dosage assessment model training module and disinfection dosage control module; The device type recognition module is used to collect image data of the medical device to be sterilized, extract device contour feature points from the image data, construct a three-dimensional space model of the medical device based on the device contour feature points, and determine the device type of the medical device; The pollutant identification module is used to extract the color depth features and color features of the image data based on the three-dimensional space model, and cluster the color depth features and color features to obtain the pollutant type and pollutant concentration of the medical device; The disinfectant combination acquisition module is used to obtain historical medical device disinfection data, input the pollutant type and device type of the medical device into a preset particle swarm model, and construct a fitness function of the particle swarm model in combination with the historical medical device disinfection data, iteratively solve the particle swarm model based on the fitness function, and obtain the optimal disinfectant combination of the medical device; The disinfection dosage evaluation model training module is used to decompose the historical medical device disinfection data according to a preset decomposition algorithm to obtain the modal components of the historical medical device disinfection data, and input the modal components into a preset disinfection dosage evaluation model, and train the disinfection dosage evaluation model according to a preset model training algorithm to obtain an optimal disinfection dosage evaluation model; The disinfection dosage control module is used to input the optimal disinfectant combination and pollutant concentration of the medical device into the optimal disinfection dosage evaluation model to obtain the optimal disinfection dosage of each disinfectant in the optimal disinfectant combination to control the disinfection dosage of the medical device.

9. A medical device disinfection dosage control system as claimed in claim 8, characterized in that: The device type identification module comprises: a device type identification unit; The device type identification unit is used to collect image data of the medical device to be sterilized, and remove high-frequency noise from the image data through a preset Gaussian filter; Decomposing the image data after high-frequency noise is removed by layers to obtain a plurality of first layers; Calculate a first gradient value set of pixel points in the first layer in the X-axis direction according to a preset Sobel edge detection operator; Calculate a second gradient value set of pixel points in the first layer in the Y-axis direction according to a preset Sobel edge detection operator; Determine a layer gradient value set of the first layer based on the first gradient value set and the second gradient value set, and select gradient values ​​greater than a preset gradient threshold from the layer gradient value set to determine a first gradient value set of the first layer; Determine edge points of the first layer based on pixel points corresponding to gradient values ​​in the first gradient value set, and determine feature points of the device contour of the first layer based on the edge points of the first layer; Reconstructing the first layer based on the instrument contour feature points to obtain a second layer; Converting the second layer into a voxel layer based on a preset voxel algorithm, and interpolating the voxel layer according to the feature points of the device contour to construct a three-dimensional spatial model of the medical device; The three-dimensional spatial model of the medical device is graphic matched with a preset medical device database according to a preset graphic matching algorithm to determine the device type of the medical device.

10. A medical device disinfection dosage control system as claimed in claim 8, characterized in that: The pollutant identification module includes: a pollutant identification unit; The pollutant identification unit is used to perform dimension conversion and color space conversion on the three-dimensional space model to determine a first color image; Extracting the number of bits of color information in the first color image to determine a color depth feature; Calculating the first-order moment, the second-order moment and the third-order moment of each color channel in the first color image, and constructing the color moment of the first color image based on the first-order moment, the second-order moment and the third-order moment to determine the color feature; Constructing color key-value pairs according to the color depth features and the color features, and clustering the color key-value pairs based on a preset clustering algorithm to obtain a number of clusters; Calculate the silhouette coefficient of each cluster, and screen the clusters based on the silhouette coefficient and the preset number to obtain several characteristic clusters; Based on the feature cluster, a feature color key-value pair is obtained, and the feature color key-value pair is matched with a preset pollutant database, the pollutant type of the medical device is determined according to the color feature, and the pollutant concentration of the medical device is determined according to the color depth feature.

Images