A single-soldier intelligent linkage standby device management method and system

Abstract

The present invention belongs to the technical field of fire fighting equipment, and discloses a single-soldier intelligent linkage standby device management method and system, including collecting equipment parameter data, task parameter data and physiological measurement image data; preprocessing the equipment parameter data and task parameter data to obtain equipment feature data set and task feature data set, and fusing the equipment feature data set and task feature data set to obtain a comprehensive feature data set; training an equipment state prediction model, inputting the comprehensive feature data set into the equipment state prediction model, and predicting a protective equipment performance index; comparing the predicted protective equipment performance index with a preset protective equipment performance index threshold to determine whether the protective equipment meets the alarm standard; being able to quickly evaluate the performance status of the protective equipment, ensuring that the equipment reaches the best state before the alarm, thereby avoiding rescue delays or safety accidents caused by equipment problems.

Description

Technical Field

[0001] The present invention relates to the technical field of fire fighting equipment, and more specifically, to a management method and system for a single-soldier intelligent linkage standby device. Background Art

[0002] In the process of modern urbanization, firefighting and rescue work is facing increasingly complex environmental challenges. As the first responder, the intelligence level of firefighters' equipment is directly related to the efficiency and safety of emergency rescue.

[0003] Traditional firefighter equipment mostly adopts single-function design and manual control mode, which has the following main problems:

[0004] Relying on manual segmentation or simple threshold segmentation results in inaccurate segmentation results, and it is impossible to accurately locate the individual soldier's body area; using only two-dimensional images to measure and calculate the individual soldier's physiological characteristics may lead to information loss due to issues such as viewing angle and occlusion, and it is impossible to fully reflect the individual soldier's physical characteristics; it is impossible to accurately measure and calculate the individual soldier's physical characteristics; it is impossible to accurately estimate key indicators such as the individual soldier's volume and weight, thus affecting the accurate assessment of the individual soldier's physical condition and protective equipment needs; decisions made based on inaccurate assessment results may lead to problems such as improper selection of protective equipment and unreasonable training plans, which in turn affect the individual soldier's safety and training results;

[0005] Failure to consider the use of feature vectors and statistical methods for matching may lead to a lack of personalization in the size selection of protective equipment, and the inability to accurately meet the specific needs of each individual soldier; the lack of an accurate matching mechanism may lead to the selection of sizes that are too large or too small, thereby affecting the wearing comfort and protective effect of the equipment; failure to adopt an automated matching method may require reliance on manual experience and judgment to select equipment sizes, which is not only time-consuming and labor-intensive, but also prone to errors; in an emergency, improper or time-consuming equipment size selection may delay the dispatch time and affect the execution of the mission.

[0006] In view of this, the present invention proposes a single-soldier intelligent linkage standby device management method and system to solve the above problems. Summary of the invention

[0007] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for managing a single-soldier intelligent linkage standby device, comprising:

[0008] S1, collecting equipment parameter data, task parameter data and physiological measurement image data;

[0009] S2. Preprocess the equipment parameter data and the task parameter data to obtain an equipment feature data set and a task feature data set, and fuse the equipment feature data set and the task feature data set to obtain a comprehensive feature data set;

[0010] S3, training the equipment status prediction model, inputting the comprehensive feature data set into the equipment status prediction model, and predicting the performance index of the protective equipment;

[0011] S4. Compare the predicted performance index of the protective equipment with the preset performance index threshold of the protective equipment to determine whether the protective equipment meets the alarm standard;

[0012] S5. If the police dispatch standard is met, the physiological measurement image data is processed to obtain a physiological characteristic data set, and the physiological characteristic data set is matched and analyzed with the size of the protective equipment to determine whether the size of the protective equipment meets the needs of individual police dispatch; if the police dispatch standard is not met, the protective equipment is promptly repaired and replaced;

[0013] S6. If the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen will issue different police dispatch instructions according to the characteristics of the task characteristic data, and the individual soldier will select protective equipment suitable for the task characteristics according to the police dispatch instructions; if the size of the protective equipment does not meet the needs of individual soldiers for police dispatch, the device will automatically rotate to protective equipment of matching size.

[0014] Furthermore, the equipment parameter data includes the storage location, storage time, usage frequency and maintenance status of the equipment; the task parameter data includes the task target type, task target location and task target priority; the physiological measurement image data includes the three-dimensional body image data and key part size image data of the soldier; the three-dimensional body image data of the soldier includes the volume, height and weight of the soldier; the key part size image data includes the chest circumference, waist circumference and hip circumference of the soldier.

[0015] Furthermore, the method of preprocessing the equipment parameter data and the task parameter data to obtain the equipment feature data set and the task feature data set, and fusing the equipment feature data set and the task feature data set to obtain the comprehensive feature data set includes:

[0016] The outliers in the equipment parameter data and the task parameter data are identified by the Laida criterion method, and the mean of each data in the equipment parameter data and the task parameter data is calculated respectively. The mean of each data is: Wherein, O(i) is each data in the equipment parameter data and the task parameter data; c′ is the total number of each data;

[0017] Calculate the standard deviation of each data Among them, μ is the mean of each data; the region range of the outlier is determined to be (μ-3σ, μ+3σ), and the values ​​beyond the outlier region are determined as outliers;

[0018] The outliers identified in the equipment parameter data and the task parameter data are removed, and the missing values ​​in the equipment parameter data and the task parameter data are filled by the forward filling method to obtain the processed equipment feature data set and the task feature data set;

[0019] The processed equipment feature data set and task feature data set are normalized by standard deviation and converted into standard normal distribution, and finally the normalized equipment feature data set and task feature data set are obtained;

[0020] The equipment feature data set and the task feature data set are fused through a weighted model to form a comprehensive feature data set. The equipment feature data set is denoted as E1, and the task feature data set is denoted as E2. The weighted model is: ST = E1·δ1+E2·δ2; ​​wherein δ1 is the weight coefficient of the equipment feature data set; δ2 is the weight coefficient of the task feature data set.

[0021] Furthermore, the training method of the equipment status prediction model includes:

[0022] Divide the data set into training set, validation set and test set, train the model and evaluate the model performance; build an equipment status prediction model, which includes input layer, hidden layer and output layer; the hidden layer uses ReLU activation function;

[0023] The input layer of the model is a historical comprehensive feature data set, and the number of neurons in the input layer matches the number of features in the historical comprehensive feature data set; the output layer of the model is a protective equipment performance indicator, and the number of neurons in the output layer corresponds to the number of predicted targets, and the predicted value is output through one neuron; the equipment status prediction model is a fully connected neural network model;

[0024] The mean absolute error is used as the loss function to measure the error between the model's predicted value and the actual value; the mean absolute error loss function is: Among them, MAE is the mean absolute error; n is the number of data sets; y i′ is the true value of the performance index of the i′th protective equipment; is the predicted value of the performance index of the i′th protective equipment;

[0025] Use the training set data to train the model, minimize the loss function through the Adam optimizer, and tune the model's hyperparameters;

[0026] The model is evaluated through the test set. When the performance of the model in the prediction task reaches the preset performance threshold, the test is stopped to obtain the equipment status prediction model; the current comprehensive feature data set is input into the trained equipment status prediction model to obtain the protective equipment performance index.

[0027] Furthermore, the method for tuning the hyperparameters of the model includes:

[0028] A hyperparameter space is preset, which contains all possible hyperparameter combinations. The hyperparameter space is: θ = {θ1, θ2, ..., θ n′}, where n′ is the number of hyperparameter combinations; the hyperparameter space is divided into m subsets, each of which contains a hyperparameter combination θ i″ ;

[0029] Define R as the set of available computing resources, R = {r1, r2, ..., r k}; where k is the number of available computing resources; each computing resource r j is assigned a hyperparameter combination θ i″ , for each hyperparameter combination θ i″ , in computing resources r j Perform training and evaluation on

[0030] The mean absolute error of the loss function in the model training process is used as the evaluation function, and S is preset to record the performance score set of the evaluation function. The performance score set is: S = {S1, S2, ..., S v}, where v is the number of performance scores;

[0031] Select the one with the best performance score S i′ The hyperparameter combination θ * , minimize the MAE value of the mean absolute error of the evaluation function; Among them, θ * is the best hyperparameter combination; argmin is the variable value operator that minimizes the evaluation function, using the best hyperparameter combination θ * The equipment status prediction model is trained on the training set, and the performance of the equipment status prediction model is evaluated using the test set.

[0032] Furthermore, the method of comparing the predicted protective equipment performance index with the preset protective equipment performance index threshold to determine whether the protective equipment meets the alarm standard includes:

[0033] If the predicted performance index of the protective equipment is greater than or equal to the preset performance index threshold of the protective equipment, it is determined that the protective equipment has reached the alarm standard;

[0034] If the predicted protective equipment performance index is less than the preset protective equipment performance index threshold, it is determined that the protective equipment does not meet the alarm standard.

[0035] Furthermore, if the alarm standard is met, the method of processing the physiological measurement image data to obtain a physiological characteristic data set includes:

[0036] The physiological measurement image is denoised by the non-local means algorithm. For each pixel p in the physiological measurement image, the similarity between all other pixels q and p is calculated within the search window Ω, and the weighted average value of these pixels is used to estimate

[0037] The non-local means algorithm is: in, is the value of the denoised physiological measurement image at pixel p; B(p) is the normalization constant that ensures that the sum of all weights is 1; I(q) is the value of the original physiological measurement image at pixel q; h is the filter parameter that controls the decay speed of the exponential function e; is the search window, which defines the pixel range considered when calculating the similarity; ||I(p)-I(q)|| 2 is the similarity between pixel q and the small blocks around p;

[0038] Use the U-Net model to segment physiological measurement images, extract individual soldier body regions, define a dataset containing physiological measurement images and corresponding pixel value labels, record the image region with a pixel value label of 1 as the individual soldier body region, and record the region with a pixel value label of 0 as the background region;

[0039] Select any pixel point from the single-soldier human body area as a candidate point u, preset the candidate point brightness threshold Q, select a circular area with a radius of r′ around the candidate point u, and select l equally spaced points on the circumference of the candidate point u as test points;

[0040] Take the brightness lu of the candidate point u and the brightness threshold Q as reference, detect the test points on the circumference, and preset the brightness range of the test points to be (lu-Q,lu+Q). When the brightness difference between l equally spaced points and the candidate point u is greater than lu+Q or less than lu-Q, the candidate point u is determined to be a feature point.

[0041] In the circular area with a radius of r′ around the candidate point u, the SIFT feature matching algorithm is used to find the matching feature point pairs in the area; the binocular stereo vision algorithm is used to reconstruct the 3D point cloud based on the matching feature point pairs: the position coordinate points of each pair of matching feature points in the 3D space are calculated using the triangulation principle and the camera's internal and external parameters, and all the position coordinate points obtained by triangulation are combined to form a 3D point cloud C;

[0042] The convex hull algorithm is used to estimate the volume of a single soldier by calculating the volume enclosed by the three-dimensional point cloud. Let the height H be the convex hull of the point cloud. The convex hull algorithm is: Where W is the number of triangle patches of the convex hull; a w , b w and c w is the vertex of the w-th triangle patch;

[0043] Calculate the distance between the highest point and the lowest point of a soldier on the three-dimensional point cloud as the soldier's height H, pre-set the soldier's body density as ρ, and multiply the soldier's volume V by the soldier's body density ρ to get the soldier's weight;

[0044] Each point C in the 3D point cloud C i The coordinates of (x i ,y i ,z i ), mark the reference points of the three key parts of the soldier's chest, waist and hip circumference as C es , C wa and C hp ; Choose any key point C from the reference points of the three key parts of the soldier's chest, waist and hips key , by the key point C key Extract the horizontal section contour line on the plane where it is located; assuming that the normal vector of the section plane is b = (0,0,1), then the point set on the horizontal section contour line is: A = {c∈C||(cC key )·b|<ε}; A is the point set on the horizontal section contour line; c is all the points in the three-dimensional point cloud C;

[0045] The point set A on the horizontal cross-section contour line of each key part is preset to contain N points {a1, a2, ..., a N}; where N is the number of points; calculate the perimeter L of the point set A on the horizontal cross-section contour line of each key part. The perimeter calculation formula is: Among them, a1 is the first point in point set A; a N is the Nth point in point set A; a i+1 is the i+1th point in point set A;

[0046] The reference points C of the three key parts es , C wa and C hp Substitute into the circumference calculation formula to find the chest circumference, waist circumference and hip circumference of a single soldier;

[0047] A density clustering algorithm is used on the extracted physiological measurement image data to identify the outliers in the physiological measurement image data, and the identified outliers are removed to obtain a processed physiological characteristic data set. The processed physiological characteristic data set is normalized by standard deviation and converted into a standard normal distribution, finally obtaining a normalized physiological characteristic data set.

[0048] Furthermore, the method of matching and analyzing the physiological characteristic data set with the size of protective equipment to determine whether the size of the protective equipment meets the needs of individual police dispatch includes:

[0049] The features in the physiological feature data set are combined into a feature vector, and the feature vector is: X = {x1, x2, x3, x4, x5, x6}; where x1 is the volume of the soldier; x2 is the height of the soldier; x3 is the weight of the soldier; x4 is the chest circumference of the soldier; x5 is the waist circumference of the soldier; x6 is the hip circumference of the soldier;

[0050] The physiological characteristic data set and the protective equipment size are used as training data. There are e types of protective equipment sizes preset. The mean vector and covariance matrix corresponding to each protective equipment size in the training data are calculated.

[0051] The mean vector is: Among them, μ e is the mean vector of the size of the e-th type of protective equipment; N e is the number of samples of the size of the e-th type of protective equipment; is the feature vector of the gth sample of the eth type of protective equipment size; e is the type of protective equipment size;

[0052] The covariance matrix is: Among them, v d is the covariance matrix of the size of the d-th type of protective equipment; for The transpose of

[0053] For a new sample feature vector X new , which is related to the mean vector μ of the e-th size e The Mahalanobis distance is defined as: Among them, d e (X new ,μ e ) is the new sample feature vector X new and the mean vector μ of the e-th size e The Mahalanobis distance of X new is the new sample feature vector X new ; is the inverse matrix of the covariance matrix of the size of the d-th type of protective equipment;

[0054] The preset Mahalanobis distance threshold is d′, and the Mahalanobis distance of each soldier is compared with the preset Mahalanobis distance threshold. If the Mahalanobis distance of each soldier is less than or equal to the preset Mahalanobis distance threshold, it is judged that the size of his equipment matches the physiological characteristic data set and meets the needs of individual soldiers to go out on police duty; if the Mahalanobis distance of each soldier is greater than the preset Mahalanobis distance threshold, it is judged that the size of his equipment does not match the physiological characteristic data set and does not meet the needs of individual soldiers to go out on police duty.

[0055] Furthermore, if the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen issues different police dispatch instructions according to the characteristics of the task characteristic data, and the method for the individual soldier to select protective equipment with appropriate task characteristics according to the police dispatch instructions includes:

[0056] The protective equipment control terminal receives mission parameter data, automatically analyzes the mission target type, mission target location and mission target priority, and automatically generates and issues specific alarm instructions based on the mission parameter data analysis results; if the mission is of a firefighting nature, a red instruction is issued, and if the mission is of a rescue nature, a yellow instruction is issued; individual soldiers receive alarm instructions through the control screen and select corresponding protective equipment according to the instruction type.

[0057] A single-soldier intelligent linkage standby device management system includes: a data acquisition module for collecting equipment parameter data, task status data and physiological measurement image data;

[0058] A data processing module is used to pre-process the equipment parameter data and the task parameter data to obtain an equipment feature data set and a task feature data set, and fuse the equipment feature data set and the task feature data set to obtain a comprehensive feature data set;

[0059] The model training module is used to train the equipment status prediction model, input the comprehensive feature data set into the equipment status prediction model, and predict the performance indicators of the protective equipment;

[0060] The equipment evaluation module is used to compare the predicted performance indicators of protective equipment with the preset thresholds of protective equipment performance indicators to determine whether the protective equipment meets the alarm standards;

[0061] The size judgment module processes the physiological measurement image data to obtain a physiological feature data set if the police standard is met, matches and analyzes the physiological feature data set with the size of the protective equipment, and determines whether the size of the protective equipment meets the needs of individual police dispatch; if the police standard is not met, the protective equipment is promptly repaired and replaced;

[0062] Equipment selection module: if the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen will issue different police dispatch instructions according to the characteristics of the task characteristic data, and the individual soldier will select protective equipment with appropriate task characteristics according to the police dispatch instructions; if the size of the protective equipment does not meet the needs of individual soldiers for police dispatch, the device will automatically rotate to protective equipment with matching size; each module is connected by wired and / or wireless means.

[0063] The technical effects and advantages of the single-soldier intelligent linkage standby device management method and system of the present invention are as follows:

[0064] The physiological measurement images were denoised using the non-local mean algorithm, which effectively reduced the noise in the images, improved the image quality, and provided a clearer image foundation for subsequent processing; the U-Net model was used to accurately segment the individual soldier body area in the physiological measurement images, providing an accurate area range for subsequent feature extraction; the SI FT feature matching algorithm and binocular stereo vision algorithm were used to reconstruct the three-dimensional point cloud, realizing the conversion from two-dimensional images to three-dimensional space, which made it possible to more accurately measure and calculate physiological characteristics; the convex hull algorithm was used to estimate the volume of the individual soldier, and the weight was calculated in combination with the body density, realizing the quantitative evaluation of the basic characteristics of the individual soldier's body;

[0065] By combining the physiological characteristics of individual soldiers into feature vectors and comparing them with the mean vector and covariance matrix of the protective equipment size, accurate matching of the personalized needs of each individual soldier is achieved, ensuring that the protective equipment fits the soldier's body better in size, thereby improving the comfort and flexibility of wearing and reducing the sense of restraint and discomfort caused by inappropriate equipment size. By calculating the Mahalanobis distance and comparing it with the preset threshold, it is possible to quickly determine whether the equipment size of the individual soldier matches the physiological characteristics, thereby improving the efficiency of preparation before the alarm. Through automated processing, it can reduce dependence on manual operations, thereby saving labor costs and time costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic diagram of a flow chart of a method for managing a single-soldier intelligent linkage standby device according to the present invention;

[0067] Figure 2 The present invention is a schematic diagram of the structure of a single-soldier intelligent linkage standby device management system. DETAILED DESCRIPTION

[0068] 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.

[0069] Example 1

[0070] See also Figure 1 As shown, this embodiment provides a method for managing a single-soldier intelligent linkage standby device, including:

[0071] S1, collecting equipment parameter data, task parameter data and physiological measurement image data;

[0072] S2. Preprocess the equipment parameter data and the task parameter data to obtain an equipment feature data set and a task feature data set, and fuse the equipment feature data set and the task feature data set to obtain a comprehensive feature data set;

[0073] S3, training the equipment status prediction model, inputting the comprehensive feature data set into the equipment status prediction model, and predicting the performance index of the protective equipment;

[0074] S4. Compare the predicted performance index of the protective equipment with the preset performance index threshold of the protective equipment to determine whether the protective equipment meets the alarm standard;

[0075] S5. If the police dispatch standard is met, the physiological measurement image data is processed to obtain a physiological characteristic data set, and the physiological characteristic data set is matched and analyzed with the size of the protective equipment to determine whether the size of the protective equipment meets the needs of individual police dispatch; if the police dispatch standard is not met, the protective equipment is promptly repaired and replaced;

[0076] S6. If the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen will issue different police dispatch instructions according to the characteristics of the task characteristic data, and the individual soldier will select protective equipment suitable for the task characteristics according to the police dispatch instructions; if the size of the protective equipment does not meet the needs of individual soldiers for police dispatch, the device will automatically rotate to protective equipment of matching size.

[0077] Equipment parameter data includes the storage location, storage time, usage frequency and maintenance status of the equipment; mission parameter data includes mission target type, mission target location and mission target priority; physiological measurement image data includes the three-dimensional body image data and key part size image data of a single soldier; the three-dimensional body image data of a single soldier includes the volume, height and weight of the single soldier; the key part size image data includes the chest circumference, waist circumference and hip circumference of the single soldier.

[0078] The storage location of the equipment is directly obtained through the location of the standby device; the storage time and frequency of use of the equipment are obtained through the protective equipment control terminal; the maintenance status of the equipment is manually inspected and recorded; the task parameter data is obtained through the task allocation terminal; the physiological measurement image data is obtained through the three-dimensional point cloud;

[0079] The method of preprocessing the equipment parameter data and the task parameter data to obtain the equipment feature data set and the task feature data set, and fusing the equipment feature data set and the task feature data set to obtain the comprehensive feature data set includes:

[0080] The outliers in the equipment parameter data and task parameter data are identified by the Laida criterion method, and the mean of each data in the equipment parameter data and task parameter data is calculated respectively. The mean of each data is: Wherein, O(i) is each data in the equipment parameter data and the task parameter data; c′ is the total number of each data;

[0081] Calculate the standard deviation of each data Among them, μ is the mean of each data; the region range of the outlier is determined to be (μ-3σ, μ+3σ), and the values ​​beyond the outlier region are determined as outliers;

[0082] The outliers identified in the equipment parameter data and the task parameter data are removed, and the missing values ​​in the equipment parameter data and the task parameter data are filled by the forward filling method to obtain the processed equipment feature data set and the task feature data set;

[0083] The processed equipment feature data set and task feature data set are normalized by standard deviation and converted into standard normal distribution, and finally the normalized equipment feature data set and task feature data set are obtained;

[0084] The equipment feature data set and the task feature data set are fused through a weighted model to form a comprehensive feature data set. The equipment feature data set is denoted as E1, and the task feature data set is denoted as E2. The weighted model is: ST = E1·δ1+E2·δ2; ​​wherein δ1 is the weight coefficient of the equipment feature data set; δ2 is the weight coefficient of the task feature data set.

[0085] The training method of the equipment status prediction model includes:

[0086] Divide the data set into training set, validation set and test set, train the model and evaluate the model performance; build an equipment status prediction model, which includes input layer, hidden layer and output layer; the hidden layer uses ReLU activation function;

[0087] The input layer of the model is a historical comprehensive feature data set, and the number of neurons in the input layer matches the number of features in the historical comprehensive feature data set; the output layer of the model is a protective equipment performance indicator, and the number of neurons in the output layer corresponds to the number of predicted targets, and the predicted value is output through one neuron; the equipment status prediction model is a fully connected neural network model;

[0088] The mean absolute error is used as the loss function to measure the error between the model's predicted value and the actual value; the mean absolute error loss function is: Among them, MAE is the mean absolute error; n is the number of data sets; y i′ is the true value of the performance index of the i′th protective equipment; is the predicted value of the performance index of the i′th protective equipment;

[0089] Use the training set data to train the model, minimize the loss function through the Adam optimizer, and tune the model's hyperparameters;

[0090] The model is evaluated through the test set. When the performance of the model in the prediction task reaches the preset performance threshold, the test is stopped to obtain the equipment status prediction model; the current comprehensive feature data set is input into the trained equipment status prediction model to obtain the protective equipment performance index.

[0091] Methods for tuning the model's hyperparameters include:

[0092] A hyperparameter space is preset, which contains all possible hyperparameter combinations. The hyperparameter space is: θ = {θ1, θ2, ..., θ n′}, where n′ is the number of hyperparameter combinations; the hyperparameter space is divided into m subsets, each of which contains a hyperparameter combination θ i″ ;

[0093] Define R as the set of available computing resources, R = {r1, r2, ..., r k}; where k is the number of available computing resources; each computing resource r j is assigned a hyperparameter combination θ i″ , for each hyperparameter combination θ i″ , in computing resources r j Perform training and evaluation on

[0094] The mean absolute error of the loss function in the model training process is used as the evaluation function, and S is preset to record the performance score set of the evaluation function. The performance score set is: S = {S1, S2, ..., S v}, where v is the number of performance scores;

[0095] Select the one with the best performance score S i′ The hyperparameter combination θ * , minimize the MAE value of the mean absolute error of the evaluation function; Among them, θ * is the best hyperparameter combination; argmin is the variable value operator that minimizes the evaluation function, using the best hyperparameter combination θ *The equipment status prediction model is trained on the training set, and the performance of the equipment status prediction model is evaluated using the test set.

[0096] For example, the hyperparameter space of a pre-set model can be defined as follows:

[0097] Learning rate: [0.001, 0.005, 0.01, 0.05]; Batch size: [32, 64, 128]; Number of hidden layers: [1, 2, 3]; Number of neurons in each layer: [64, 128, 256]; This hyperparameter space contains a total of 4×3×3×3=108 different hyperparameter combinations; Assume that there are enough computing resources, such as 108 independent GPUs or CPU clusters to process all hyperparameter combinations simultaneously, and each computing resource is assigned a unique hyperparameter combination;

[0098] Preset a performance score set to record the MAE value corresponding to each hyperparameter combination, for example: (0.001, 32, 1, 64): 0.05, (0.001, 32, 1, 128): 0.045, (0.05, 128, 3, 256): 0.03. Assuming this is the best performance, retrain the model on the full training set using the best hyperparameter combination found (0.05, 128, 3, 256) and evaluate its performance using the test set to ensure that the model performs well on unseen data.

[0099] The method of comparing the predicted performance index of protective equipment with the preset performance index threshold of protective equipment to determine whether the protective equipment meets the alarm standard includes:

[0100] If the predicted performance index of the protective equipment is greater than or equal to the preset performance index threshold of the protective equipment, it is determined that the protective equipment has reached the alarm standard;

[0101] If the predicted protective equipment performance index is less than the preset protective equipment performance index threshold, it is determined that the protective equipment does not meet the alarm standard.

[0102] If the police response standard is met, the method of processing the physiological measurement image data to obtain the physiological characteristic data set includes:

[0103] The physiological measurement image is denoised by the non-local means algorithm. For each pixel p in the physiological measurement image, the similarity between all other pixels q and p is calculated within the search window Ω, and the weighted average value of these pixels is used to estimate

[0104] The non-local means algorithm is: in, is the value of the denoised physiological measurement image at pixel p; B(p) is the normalization constant that ensures that the sum of all weights is 1; I(q) is the value of the original physiological measurement image at pixel q; h is the filter parameter that controls the decay speed of the exponential function e; is the search window, which defines the pixel range considered when calculating the similarity; ||I(p)-I(q)|| 2 is the similarity between pixel q and the small blocks around p;

[0105] Use the U-Net model to segment physiological measurement images, extract individual soldier body regions, define a dataset containing physiological measurement images and corresponding pixel value labels, record the image region with a pixel value label of 1 as the individual soldier body region, and record the region with a pixel value label of 0 as the background region;

[0106] Select any pixel point from the single-soldier human body area as a candidate point u, preset the candidate point brightness threshold Q, select a circular area with a radius of r′ around the candidate point u, and select l equally spaced points on the circumference of the candidate point u as test points;

[0107] Take the brightness lu of the candidate point u and the brightness threshold Q as reference, detect the test points on the circumference, and preset the brightness range of the test points to be (lu-Q,lu+Q). When the brightness difference between l equally spaced points and the candidate point u is greater than lu+Q or less than lu-Q, the candidate point u is determined to be a feature point.

[0108] In the circular area with a radius of r′ around the candidate point u, the SIFT feature matching algorithm is used to find the matching feature point pairs in the area; the binocular stereo vision algorithm is used to reconstruct the 3D point cloud based on the matching feature point pairs: the position coordinate points of each pair of matching feature points in the 3D space are calculated using the triangulation principle and the camera's internal and external parameters, and all the position coordinate points obtained by triangulation are combined to form a 3D point cloud C;

[0109] The convex hull algorithm is used to estimate the volume of a single soldier by calculating the volume enclosed by the three-dimensional point cloud. Let the height H be the convex hull of the point cloud. The convex hull algorithm is: Where W is the number of triangle patches of the convex hull; a w , b w and c w is the vertex of the w-th triangle patch;

[0110] Calculate the distance between the highest point and the lowest point of a soldier on the three-dimensional point cloud as the soldier's height H, pre-set the soldier's body density as ρ, and multiply the soldier's volume V by the soldier's body density ρ to get the soldier's weight;

[0111] Each point C in the 3D point cloud C i The coordinates of (x i ,yi ,z i ), the reference points of the three key parts of the soldier’s chest, waist and hip are C es , C wa and C hp ; Choose any key point C from the reference points of the three key parts of the soldier's chest, waist and hips key , by the key point C key Extract the horizontal section contour line on the plane where it is located; assuming that the normal vector of the section plane is b = (0,0,1), then the point set on the horizontal section contour line is: A = {c∈C||(cC key )·b|<ε}; A is the point set on the horizontal section contour line; c is all the points in the three-dimensional point cloud C;

[0112] The point set A on the horizontal cross-section contour line of each key part is preset to contain N points {a1, a2, ..., a N}; where N is the number of points; calculate the perimeter L of the point set A on the horizontal cross-section contour line of each key part. The perimeter calculation formula is: Among them, a1 is the first point in point set A; a N is the Nth point in point set A; a i+1 is the i+1th point in point set A;

[0113] The reference points C of the three key parts es , C wa and C hp Substitute into the circumference calculation formula to find the chest circumference, waist circumference and hip circumference of a single soldier;

[0114] For example, the search window is set to 31x31 pixels, the filter parameter controlling the decay speed of the exponential function e is set to 10, and a data set containing physiological measurement images and corresponding pixel value labels is prepared. The image area with a pixel value label of 1 is recorded as a single soldier human body area, and the area with a pixel value label of 0 is recorded as a background area;

[0115] From the segmented single-soldier human body area, randomly select a pixel point as a candidate point, such as a point with coordinates (512,512), preset a brightness threshold of 100, select a circular area with a radius of 20 pixels around the candidate point u, and select 8 equally spaced points on the circumference of the candidate point u as test points, so that the circumference can be relatively evenly covered. For each test point on the circumference, calculate its brightness value and compare it with the brightness value of the candidate point. Assuming that the brightness value of the candidate point is 120, when the brightness difference between the 8 equally spaced points and the candidate point u is greater than 220 or less than 20, the candidate point u is determined to be a feature point;

[0116] Assume that the brightness values ​​of 8 test points on the circumference are calculated, and the results are as follows:

[0117] Test point 1: 105, test point 2: 122, test point 3: 90, test point 4: 140, test point 5: 118, test point 6: 120, test point 7: 85, test point 8: 135; According to the above data, the brightness of each test point is compared with the brightness of the candidate point (120), test point 1: |105-120|=15<20; test point 2: |122-120|=2<20; test point 3: |90-120|=30>20; ...; test point 8: |135-120|=15<20; If the brightness difference between the total number of test points on the circumference and the candidate point is greater than 220 or less than 20, the candidate point is determined to be a feature point;

[0118] A density clustering algorithm is used on the extracted physiological measurement image data to identify the outliers in the physiological measurement image data, and the identified outliers are removed to obtain a processed physiological characteristic data set. The processed physiological characteristic data set is normalized by standard deviation and converted into a standard normal distribution, finally obtaining a normalized physiological characteristic data set.

[0119] The method of matching and analyzing the physiological characteristic data set with the size of protective equipment to determine whether the size of protective equipment meets the needs of individual police deployment includes:

[0120] The features in the physiological feature data set are combined into a feature vector, and the feature vector is: X = {x1, x2, x3, x4, x5, x6}; where x1 is the volume of the soldier; x2 is the height of the soldier; x3 is the weight of the soldier; x4 is the chest circumference of the soldier; x5 is the waist circumference of the soldier; x6 is the hip circumference of the soldier;

[0121] The physiological characteristic data set and the protective equipment size are used as training data. There are e types of protective equipment sizes preset. The mean vector and covariance matrix corresponding to each protective equipment size in the training data are calculated.

[0122] The mean vector is: Among them, μ e is the mean vector of the size of the e-th type of protective equipment; N e is the number of samples of the size of the e-th type of protective equipment; is the feature vector of the gth sample of the eth type of protective equipment size; e is the type of protective equipment size;

[0123] The covariance matrix is: Among them, ∑ d is the covariance matrix of the size of the d-th type of protective equipment; for The transpose of

[0124] For a new sample feature vector Xnew , which is related to the mean vector μ of the e-th size e The Mahalanobis distance is defined as: Among them, d e (X new ,μ e ) is the new sample feature vector X new and the mean vector μ of the e-th size e The Mahalanobis distance of X new is the new sample feature vector X new ; is the inverse matrix of the covariance matrix of the size of the d-th type of protective equipment;

[0125] The preset Mahalanobis distance threshold is d′, and the Mahalanobis distance of each soldier is compared with the preset Mahalanobis distance threshold. If the Mahalanobis distance of each soldier is less than or equal to the preset Mahalanobis distance threshold, it is judged that the size of his equipment matches the physiological characteristic data set and meets the needs of individual soldiers to go out on police duty; if the Mahalanobis distance of each soldier is greater than the preset Mahalanobis distance threshold, it is judged that the size of his equipment does not match the physiological characteristic data set and does not meet the needs of individual soldiers to go out on police duty.

[0126] For example, suppose there is the following physiological feature data set and the corresponding protective equipment sizes, 3 protective equipment sizes (small, medium, and large), the number of samples for each size is 5, and the new sample feature vector is X new =(6017070908095); the preset Mahalanobis distance threshold d′ is 3;

[0127] The sample mean vector of the size of small protective equipment is: (5516565857590); the sample covariance matrix is ​​a 6×6 matrix;

[0128] The Mahalanobis distance is calculated as:

[0129] The sample mean vector for medium protective gear size is: (6017070908095); the sample covariance matrix is ​​a 6×6 matrix;

[0130] The Mahalanobis distance is calculated as:

[0131] The sample mean vector of the large protective gear size is: (65175759585100); the sample covariance matrix is ​​also a 6×6 matrix;

[0132] The Mahalanobis distance is calculated as:

[0133] Compare the Mahalanobis distance of each size with the preset Mahalanobis distance threshold: small protective equipment size: Mahalanobis distance is about 12.25, greater than the threshold 3, no match; medium protective equipment size: Mahalanobis distance is 0, less than or equal to the threshold 3, match; large protective equipment size: Mahalanobis distance is about 12.25, greater than the threshold 3, no match; therefore, the most matching protective equipment size for the new sample is medium, which meets the needs of individual police dispatch.

[0134] If the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen will issue different police dispatch instructions according to the characteristics of the task characteristic data, and the method for individual soldiers to select protective equipment with appropriate task characteristics according to the police dispatch instructions includes:

[0135] The protective equipment control terminal receives mission parameter data, automatically analyzes the mission target type, mission target location and mission target priority, and automatically generates and issues specific alarm instructions based on the mission parameter data analysis results; if the mission is of a firefighting nature, a red instruction is issued, and if the mission is of a rescue nature, a yellow instruction is issued; individual soldiers receive alarm instructions through the control screen and select corresponding protective equipment according to the instruction type.

[0136] The preset protective equipment performance index threshold is set by the staff. Different protective equipment performance indexes are collected through the protective equipment control terminal, and the average value of multiple protective equipment performance indexes is taken as the preset protective equipment performance index threshold; the preset performance threshold and the preset Mahalanobis distance threshold are set in the same way;

[0137] In this embodiment, the physiological measurement image is denoised by the non-local mean algorithm, which effectively reduces the noise in the image, improves the image quality, and provides a clearer image basis for subsequent processing; the U-Net model is used to accurately segment the individual soldier body area in the physiological measurement image, providing an accurate area range for subsequent feature extraction; the SIFT feature matching algorithm and the binocular stereo vision algorithm are used to reconstruct the three-dimensional point cloud, realize the conversion from the two-dimensional image to the three-dimensional space, and provide the possibility for more accurate measurement and calculation of physiological characteristics; the convex hull algorithm is used to estimate the volume of the individual soldier, and the weight is calculated in combination with the body density, so as to realize the quantitative evaluation of the basic characteristics of the individual soldier's body;

[0138] By combining the physiological characteristics of individual soldiers into feature vectors and comparing them with the mean vector and covariance matrix of the protective equipment size, accurate matching of the personalized needs of each individual soldier is achieved, ensuring that the protective equipment fits the soldier's body better in size, thereby improving the comfort and flexibility of wearing and reducing the sense of restraint and discomfort caused by inappropriate equipment size. By calculating the Mahalanobis distance and comparing it with the preset threshold, it is possible to quickly determine whether the equipment size of the individual soldier matches the physiological characteristics, thereby improving the efficiency of preparation before the alarm. Through automated processing, it can reduce dependence on manual operations, thereby saving labor costs and time costs.

[0139] Example 2

[0140] See also Figure 2 As shown, this embodiment provides a single-soldier intelligent linkage standby device management system, including:

[0141] A data acquisition module, used to collect equipment parameter data, task status data and physiological measurement image data;

[0142] A data processing module is used to pre-process the equipment parameter data and the task parameter data to obtain an equipment feature data set and a task feature data set, and fuse the equipment feature data set and the task feature data set to obtain a comprehensive feature data set;

[0143] The model training module is used to train the equipment status prediction model, input the comprehensive feature data set into the equipment status prediction model, and predict the performance indicators of the protective equipment;

[0144] The equipment evaluation module is used to compare the predicted performance indicators of protective equipment with the preset performance indicator thresholds of protective equipment to determine whether the protective equipment meets the alarm standards;

[0145] The size judgment module processes the physiological measurement image data to obtain a physiological feature data set if the police standard is met, matches and analyzes the physiological feature data set with the size of the protective equipment, and determines whether the size of the protective equipment meets the needs of individual police dispatch; if the police standard is not met, the protective equipment is promptly repaired and replaced;

[0146] Equipment selection module: if the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen will issue different police dispatch instructions based on the characteristics of the task characteristic data, and the individual soldier will select protective equipment with appropriate task characteristics according to the police dispatch instructions; if the size of the protective equipment does not meet the needs of individual soldiers for police dispatch, the device will automatically rotate to protective equipment with matching size.

[0147] Since the electronic device introduced in this embodiment is an electronic device used to implement a single-soldier intelligent linkage standby device management method and system in the embodiment of the present application, based on the single-soldier intelligent linkage standby device management method and system introduced in the embodiment of the present application, the technical personnel of this field can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present application is not described in detail here. As long as the technical personnel of this field implement the electronic device used in the single-soldier intelligent linkage standby device management method and system in the embodiment of the present application, it belongs to the scope of protection of this application.

[0148] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0149] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for managing a single-soldier intelligent linkage standby device, characterized in that: include: S1, collecting equipment parameter data, task parameter data and physiological measurement image data; S2. Preprocess the equipment parameter data and the task parameter data to obtain an equipment feature data set and a task feature data set, and fuse the equipment feature data set and the task feature data set to obtain a comprehensive feature data set; S3, training the equipment status prediction model, inputting the comprehensive feature data set into the equipment status prediction model, and predicting the performance index of the protective equipment; S4. Compare the predicted performance index of the protective equipment with the preset performance index threshold of the protective equipment to determine whether the protective equipment meets the alarm standard; S5. If the police dispatch standard is met, the physiological measurement image data is processed to obtain a physiological characteristic data set, and the physiological characteristic data set is matched and analyzed with the size of the protective equipment to determine whether the size of the protective equipment meets the needs of individual police dispatch; if the police dispatch standard is not met, the protective equipment is promptly repaired and replaced; S6. If the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen will issue different police dispatch instructions according to the characteristics of the task characteristic data, and the individual soldier will select protective equipment suitable for the task characteristics according to the police dispatch instructions; if the size of the protective equipment does not meet the needs of individual soldiers for police dispatch, the device will automatically rotate to protective equipment with matching size; The equipment parameter data includes the storage location, storage time, usage frequency and maintenance status of the equipment; the mission parameter data includes the mission target type, mission target location and mission target priority; the physiological measurement image data includes the three-dimensional body image data and key part size image data of the individual soldier; the three-dimensional body image data of the individual soldier includes the volume, height and weight of the individual soldier; the key part size image data includes the chest circumference, waist circumference and hip circumference of the individual soldier; The method of preprocessing the equipment parameter data and the task parameter data to obtain the equipment feature data set and the task feature data set, and fusing the equipment feature data set and the task feature data set to obtain the comprehensive feature data set includes: The outliers in the equipment parameter data and the task parameter data are identified by the Laida criterion method, and the mean of each data in the equipment parameter data and the task parameter data is calculated respectively. The mean of each data is: Wherein, O(i) is each data in the equipment parameter data and the task parameter data; c′ is the total number of each data; Calculate the standard deviation of each data Among them, μ is the mean of each data; the region range of the outlier is determined to be (μ-3σ, μ+3σ), and the values ​​beyond the outlier region are determined as outliers; The outliers identified in the equipment parameter data and the task parameter data are removed, and the missing values ​​in the equipment parameter data and the task parameter data are filled by the forward filling method to obtain the processed equipment feature data set and the task feature data set; The processed equipment feature data set and task feature data set are normalized by standard deviation and converted into standard normal distribution, and finally the normalized equipment feature data set and task feature data set are obtained; The equipment feature data set and the task feature data set are fused through a weighted model to form a comprehensive feature data set, the equipment feature data set is denoted as E1, and the task feature data set is denoted as E2; the weighted model is: ST = E1·δ1+E2·δ2; ​​wherein δ1 is the weight coefficient of the equipment feature data set; δ2 is the weight coefficient of the task feature data set; The training method of the equipment status prediction model includes: Divide the data set into training set, validation set and test set, train the model and evaluate the model performance; build an equipment status prediction model, which includes input layer, hidden layer and output layer; the hidden layer uses ReLU activation function; The input layer of the model is a historical comprehensive feature data set, and the number of neurons in the input layer matches the number of features in the historical comprehensive feature data set; the output layer of the model is a protective equipment performance indicator, and the number of neurons in the output layer corresponds to the number of predicted targets, and the predicted value is output through one neuron; the equipment status prediction model is a fully connected neural network model; The mean absolute error is used as the loss function to measure the error between the model's predicted value and the actual value; the mean absolute error loss function is: Among them, MAE is the mean absolute error; n is the number of data sets; y i′ is the true value of the performance index of the i′th protective equipment; is the predicted value of the performance index of the i′th protective equipment; the model is trained using the training set data, the loss function is minimized through the Adam optimizer, and the hyperparameters of the model are tuned; The model is evaluated through the test set. When the performance of the model in the prediction task reaches the preset performance threshold, the test is stopped to obtain the equipment status prediction model; the current comprehensive feature data set is input into the trained equipment status prediction model to obtain the protective equipment performance index; The method for tuning the hyperparameters of the model includes: A hyperparameter space is preset, which contains all possible hyperparameter combinations. The hyperparameter space is: θ = {θ1, θ2, ..., θ n′ }, where n′ is the number of hyperparameter combinations; the hyperparameter space is divided into m subsets, each subset contains a hyperparameter combination θ i″ ; Define R as the set of available computing resources, R = {r1, r2, ..., r k }; where k is the number of available computing resources; each computing resource r j is assigned a hyperparameter combination θ i″ , for each hyperparameter combination θ i″ , in computing resources r j Perform training and evaluation on The mean absolute error of the loss function in the model training process is used as the evaluation function, and S is preset to record the performance score set of the evaluation function. The performance score set is: S = {S1, S2, ..., S v }, where v is the number of performance scores; Select the one with the best performance score S i′ The hyperparameter combination θ * , minimize the MAE value of the mean absolute error of the evaluation function; Among them, θ * is the best hyperparameter combination; arg min is the variable value operator that makes the evaluation function take the minimum value, using the best hyperparameter combination θ * Train the equipment status prediction model on the training set, and use the test set to evaluate the performance of the equipment status prediction model; The method of comparing the predicted protective equipment performance index with the preset protective equipment performance index threshold to determine whether the protective equipment meets the alarm standard includes: If the predicted performance index of the protective equipment is greater than or equal to the preset performance index threshold of the protective equipment, it is determined that the protective equipment has reached the alarm standard; If the predicted performance index of the protective equipment is less than the preset performance index threshold of the protective equipment, it is determined that the protective equipment does not meet the alarm standard; If the alarm standard is met, the method of processing the physiological measurement image data to obtain a physiological characteristic data set includes: The physiological measurement image is denoised by the non-local means algorithm. For each pixel p in the physiological measurement image, the similarity between all other pixels g and p is calculated within the search window Ω, and the weighted average value of these pixels is used to estimate The non-local means algorithm is: in, is the value of the denoised physiological measurement image at pixel ; B(p) is the normalization constant that ensures that the sum of all weights is 1; I(q) is the value of the original physiological measurement image at pixel g; h is the filter parameter that controls the decay speed of the exponential function e; is the search window; ||I(q)-I(q)|| 2 is the similarity between pixel g and the surrounding small blocks; Use the U-Net model to segment physiological measurement images, extract individual soldier body regions, define a dataset containing physiological measurement images and corresponding pixel value labels, record the image region with a pixel value label of 1 as the individual soldier body region, and record the region with a pixel value label of 0 as the background region; Select any pixel point from the single-soldier human body area as a candidate point u, preset the candidate point brightness threshold Q, select a circular area with a radius of r′ around the candidate point u, and select l equally spaced points on the circumference of the candidate point u as test points; Take the brightness lu of the candidate point u and the brightness threshold Q as reference, detect the test points on the circumference, and preset the brightness range of the test points to be (lu-Q, lu+Q). When the brightness difference between l equally spaced points and the candidate point u is greater than lu+Q or less than lu-Q, the candidate point u is determined to be a feature point. In the circular area with a radius of r′ around the candidate point u, the SIFT feature matching algorithm is used to find the matching feature point pairs in the area; the binocular stereo vision algorithm is used to reconstruct the 3D point cloud based on the matching feature point pairs: the position coordinate points of each pair of matching feature points in the 3D space are calculated using the triangulation principle and the camera's internal and external parameters, and all the position coordinate points obtained by triangulation are combined to form a 3D point cloud C; The convex hull algorithm is used to estimate the volume of a single soldier by calculating the volume enclosed by the three-dimensional point cloud. Let the height H be the convex hull of the point cloud. The convex hull algorithm is: Where W is the number of triangle patches of the convex hull; a w 、b w and c w is the vertex of the w-th triangle patch; Calculate the distance between the highest point and the lowest point of a soldier on the three-dimensional point cloud as the soldier's height H, pre-set the soldier's body density as ρ, and multiply the soldier's volume V by the soldier's body density ρ to get the soldier's weight; Each point C in the 3D point cloud C i The coordinates of (x i ,y i , z i ), the reference points of the three key parts of the soldier’s chest, waist and hip are C es , C wa and C hp ; Choose any key point C from the reference points of the three key parts of the soldier's chest, waist and hips key , by the key point C key Extract the horizontal section contour line on the plane where it is located. Assuming that the normal vector of the section plane is b = (0, 0, 1), the point set on the horizontal section contour line is: A = {c∈C||(cC key )·b|<ε}; A is the point set on the horizontal section contour line; c is all the points in the three-dimensional point cloud C; The point set A on the horizontal cross-section contour line of each key part is preset to contain N points {a1, a2, ..., a N }; where N is the number of points; calculate the perimeter L of the point set A on the horizontal cross-section contour line of each key part. The perimeter calculation formula is: Among them, a1 is the first point in point set A; a N is the Nth point in point set A; a i+1 is the i+1th point in point set A; The reference points C of the three key parts es , C wa and C hp Substitute into the circumference calculation formula to find the chest circumference, waist circumference and hip circumference of a single soldier; A density clustering algorithm is used on the extracted physiological measurement image data to identify the outliers in the physiological measurement image data, and the identified outliers are removed to obtain a processed physiological characteristic data set. The processed physiological characteristic data set is normalized by standard deviation and converted into a standard normal distribution, finally obtaining a normalized physiological characteristic data set.

2. The method for managing a single-soldier intelligent linkage standby device according to claim 1, characterized in that: The method of matching and analyzing the physiological characteristic data set with the size of protective equipment to determine whether the size of the protective equipment meets the needs of individual soldiers for police dispatch includes: The features in the physiological feature data set are combined into a feature vector, and the feature vector is: X = {x1, x2, x3, x4, x5, x6}; Among them, x1 is the volume of a single soldier; x2 is the height of a single soldier; x3 is the weight of a single soldier; x4 is the chest circumference of a single soldier; x5 is the waist circumference of a single soldier; x6 is the hip circumference of a single soldier; The physiological characteristic data set and the protective equipment size are used as training data. There are e types of protective equipment sizes preset. The mean vector and covariance matrix corresponding to each protective equipment size in the training data are calculated. The mean vector is: Among them, μ e is the mean vector of the size of the e-th type of protective equipment; N e is the number of samples of the size of the e-th type of protective equipment; is the feature vector of the gth sample of the eth type of protective equipment size; e is the type of protective equipment size; The covariance matrix is: Among them, ∑ e is the covariance matrix of the size of the e-th type of protective equipment; for The transpose of For a new sample feature vector X new , which is related to the mean vector μ of the e-th size e The Mahalanobis distance is defined as: Among them, d e (X new , μ e ) is the new sample feature vector X new and the mean vector μ of the e-th size e The Mahalanobis distance of X new is the new sample feature vector X new ; is the inverse matrix of the covariance matrix of the size of the e-th type of protective equipment; The preset Mahalanobis distance threshold is d′, and the Mahalanobis distance of each soldier is compared with the preset Mahalanobis distance threshold. If the Mahalanobis distance of each soldier is less than or equal to the preset Mahalanobis distance threshold, it is judged that the size of his equipment matches the physiological characteristic data set and meets the needs of individual soldiers to go out on police duty; if the Mahalanobis distance of each soldier is greater than the preset Mahalanobis distance threshold, it is judged that the size of his equipment does not match the physiological characteristic data set and does not meet the needs of individual soldiers to go out on police duty.

3. The method for managing a single-soldier intelligent linkage standby device according to claim 2 is characterized in that: If the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen issues different police dispatch instructions according to the characteristics of the task characteristic data, and the method for individual soldiers to select protective equipment with appropriate task characteristics according to the police dispatch instructions includes: The protective equipment control terminal receives mission parameter data, automatically analyzes the mission target type, mission target location and mission target priority, and automatically generates and issues specific alarm instructions based on the mission parameter data analysis results; if the mission is of a firefighting nature, a red instruction is issued, and if the mission is of a rescue nature, a yellow instruction is issued; individual soldiers receive alarm instructions through the control screen and select corresponding protective equipment according to the instruction type.

4. A single-soldier intelligent linkage standby device management system, used to implement the single-soldier intelligent linkage standby device management method described in any one of claims 1 to 3, characterized in that: include: A data acquisition module, used to collect equipment parameter data, task status data and physiological measurement image data; A data processing module is used to pre-process the equipment parameter data and the task parameter data to obtain an equipment feature data set and a task feature data set, and fuse the equipment feature data set and the task feature data set to obtain a comprehensive feature data set; The model training module is used to train the equipment status prediction model, input the comprehensive feature data set into the equipment status prediction model, and predict the performance indicators of the protective equipment; The equipment evaluation module is used to compare the predicted performance indicators of protective equipment with the preset performance indicator thresholds of protective equipment to determine whether the protective equipment meets the alarm standards; The size judgment module processes the physiological measurement image data to obtain a physiological feature data set if the police standard is met, matches and analyzes the physiological feature data set with the size of the protective equipment, and determines whether the size of the protective equipment meets the needs of individual police dispatch; if the police standard is not met, the protective equipment is promptly repaired and replaced; Equipment selection module: if the size of the protective equipment meets the needs of individual soldiers for police dispatch, the control screen will issue different police dispatch instructions according to the characteristics of the task characteristic data, and the individual soldier will select protective equipment with appropriate task characteristics according to the police dispatch instructions; if the size of the protective equipment does not meet the needs of individual soldiers for police dispatch, the device will automatically rotate to protective equipment with matching size; each module is connected by wired and / or wireless means.

Images