Large recreation facility safety detection method and system based on multiple sensors

Abstract

The invention discloses a large recreation facility safety detection method and system based on multiple sensors, and relates to the technical field of data encryption, and the method comprises the steps: based on a multi-modal data set, extracting vibration signal time-frequency features through fast Fourier transform and wavelet analysis, and generating a fault analysis report in combination with a support vector machine algorithm; based on the comprehensive characteristic spectrum, carrying out recreation facility health state quantitative evaluation through a Bayesian inversion algorithm, and generating a health state diagnosis report; based on a differential detection scheme, multi-source information fusion decision making is carried out through a D-S evidence theory, and graded safety early warning is generated. According to the method, the time-frequency characteristics of the vibration signals are extracted through fast Fourier transform and wavelet analysis, and a fault analysis report is generated in combination with a support vector machine algorithm, so that accurate analysis of the vibration state of the recreation facility is realized; and the accuracy of fault diagnosis is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent detection technology, and in particular to a multi-sensor based large-scale amusement facility safety detection method and system. Background Art

[0002] As a vital component of the modern entertainment industry, the safety of large-scale amusement facilities has always attracted considerable attention. With technological advancements, particularly breakthroughs in sensor technology, data processing techniques, and artificial intelligence algorithms, amusement facility safety monitoring methods have undergone significant evolution. In recent years, the widespread adoption of the Internet of Things (IoT) has driven the application of multiple sensor types, enabling more sophisticated data collection on the operating status of amusement facilities, providing technical support for more accurate safety monitoring.

[0003] With the advancement of sensor technology, it is now possible to utilize a variety of sensor types to monitor key aspects of amusement rides in real time, such as vibration, temperature, and acoustic signals. By integrating and analyzing this multimodal data, we can not only gain a more comprehensive understanding of the equipment's operating status but also predict potential failures. However, in this process, how to efficiently integrate data from different sensors and apply advanced data analysis methods to accurately assess the health of amusement rides has become an important research direction. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a large-scale amusement facility safety detection method based on multiple sensors to solve the problem of low efficiency of multimodal data fusion.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a safety detection method for large-scale amusement facilities based on multiple sensors, which includes real-time collection of vibration data, temperature data and acoustic signals of stress-bearing parts of large-scale amusement facilities and preprocessing to generate a multimodal data set; based on the multimodal data set, extracting the time-frequency characteristics of the vibration signal through fast Fourier transform and wavelet analysis, and generating a fault analysis report in combination with a support vector machine algorithm; based on the fault analysis report, extracting the temperature signal characteristics and acoustic signal characteristics of the fault area through principal component analysis, and performing multimodal feature fusion through a typical correlation analysis method to generate a comprehensive feature map; based on the comprehensive feature map, performing a quantitative assessment of the health status of the amusement facility through a Bayesian inversion algorithm to generate a health status diagnosis report; based on the health status diagnosis report, evaluating the risk levels of different areas through a fuzzy Bayesian network, and optimizing the detection parameters through a particle swarm optimization algorithm to generate differentiated detection schemes for different risk areas; based on the differentiated detection scheme, performing multi-source information fusion decision-making through DS evidence theory to generate a graded safety warning.

[0008] As a preferred solution of the multi-sensor based large-scale amusement facility safety detection method described in the present invention, the preprocessing includes time alignment of multi-source sensor data through a high-precision clock synchronization protocol, using Kalman filtering to eliminate noise interference, and performing normalization processing.

[0009] As a preferred solution of the multi-sensor based large-scale amusement facility safety detection method of the present invention, wherein: based on the multimodal data set, the time-frequency characteristics of the vibration signal are extracted by fast Fourier transform and wavelet analysis, and the fault analysis report is generated in combination with the support vector machine algorithm. The specific steps are as follows:

[0010] Based on the multimodal data set, the vibration signal is spectrally analyzed by fast Fourier transform to generate the spectrum feature vector;

[0011] Perform time-frequency analysis on the spectrum feature vector through wavelet analysis to generate a time-frequency feature matrix;

[0012] Based on the time-frequency feature matrix, the vibration state of the amusement facilities is analyzed through the support vector machine classifier to generate a fault analysis report.

[0013] As a preferred solution of the multi-sensor based large-scale amusement facility safety detection method of the present invention, wherein: based on the fault analysis report, the temperature signal characteristics and acoustic signal characteristics of the fault area are extracted by principal component analysis, and multimodal feature fusion is performed by canonical correlation analysis method to generate a comprehensive feature map. The specific steps are as follows:

[0014] Based on the fault analysis report, feature extraction is performed on the multimodal dataset through principal component analysis to generate temperature feature vectors and acoustic feature vectors;

[0015] Based on the temperature eigenvector and acoustic eigenvector, multimodal feature fusion is performed through the canonical correlation analysis method, and the fusion feature matrix is ​​generated through the canonical correlation analysis algorithm;

[0016] Based on the fused feature matrix, the feature space is probabilistically modeled through kernel density estimation to generate a comprehensive feature map.

[0017] As a preferred solution of the multi-sensor-based large-scale amusement facility safety detection method of the present invention, wherein: based on the comprehensive feature map, the health status of the amusement facility is quantitatively evaluated by the Bayesian inversion algorithm to generate a health status diagnosis report. The specific steps are as follows:

[0018] Based on the comprehensive feature map, the Bayesian inversion algorithm is used to perform parameter inversion on the fault probability distribution, generating a quantized matrix of the amusement facility state containing the posterior probability.

[0019] Based on the quantification matrix of the amusement facility status, the health of the amusement facility is evaluated at multiple levels through fuzzy logic reasoning to generate a health diagnosis score H;

[0020] Generate a health diagnosis report based on the health diagnosis score.

[0021] As a preferred solution of the multi-sensor-based large-scale amusement facility safety detection method of the present invention, the risk level of different areas is evaluated based on the health status diagnosis report through a fuzzy Bayesian network, and the detection parameters are optimized by a particle swarm optimization algorithm to generate differentiated detection solutions for different risk areas. The specific steps are as follows:

[0022] Based on the health status diagnosis report, the risk probability of each area of ​​the amusement facility is quantitatively calculated through fuzzy Bayesian to generate a regional safety assessment map;

[0023] Based on the regional safety assessment map, an improved particle swarm optimization algorithm is used to perform multi-objective optimization on core parameters such as monitoring frequency and sampling accuracy to generate an optimal detection parameter configuration table;

[0024] Based on the optimal detection parameter configuration table, detection resources are dynamically allocated through the space-time dual-dimensional scheduling algorithm to generate differentiated detection implementation plans.

[0025] As a preferred solution of the multi-sensor based large-scale amusement facility safety detection method of the present invention, wherein: the differentiated detection scheme is based on multi-source information fusion decision-making through DS evidence theory to generate graded safety warnings. The specific steps are as follows:

[0026] Based on the differentiated detection scheme, the probability distribution model of sensor data is carried out through the DS evidence theory algorithm to generate a set of evidence bodies for different risk levels;

[0027] Based on the evidence set, the conflicting evidence sets are orthogonally summed using the Dempster combination rule to generate a consistent probability distribution after fusion.

[0028] Based on the consistent probability distribution, the security risks are graded using the confidence interval decision criterion and graded security warnings are generated.

[0029] In a second aspect, the present invention provides a large-scale amusement facility safety detection system based on multiple sensors, comprising a data acquisition module, a fault feature analysis module, a feature fusion module, a health status assessment module, a risk detection optimization module and a safety warning module;

[0030] The data acquisition module collects vibration data, temperature data, and acoustic signals from the stress-bearing parts of large-scale amusement facilities in real time and pre-processes them to generate multimodal data sets;

[0031] The fault feature analysis module, based on a multimodal data set, extracts the time-frequency features of vibration signals through fast Fourier transform and wavelet analysis, and generates a fault analysis report in combination with the support vector machine algorithm;

[0032] The feature fusion module, based on the fault analysis report, extracts the temperature signal characteristics and acoustic signal characteristics of the fault area through principal component analysis, and performs multimodal feature fusion through canonical correlation analysis to generate a comprehensive feature map;

[0033] The health status assessment module uses the Bayesian inversion algorithm to quantitatively assess the health status of amusement facilities based on the comprehensive feature map and generate a health status diagnosis report;

[0034] The risk detection optimization module uses a fuzzy Bayesian network to assess the risk levels of different areas based on health status diagnostic reports, and optimizes detection parameters using a particle swarm optimization algorithm to generate differentiated detection plans for different risk areas.

[0035] The security warning module, based on differentiated detection schemes, makes multi-source information fusion decisions through DS evidence theory to generate graded security warnings.

[0036] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the large-scale amusement facility safety detection method based on multiple sensors as described in the first aspect of the present invention is implemented.

[0037] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the large-scale amusement facility safety detection method based on multiple sensors as described in the first aspect of the present invention is implemented.

[0038] The present invention achieves the following beneficial effects: By extracting the time-frequency characteristics of vibration signals using fast Fourier transform and wavelet analysis, and combining this with a support vector machine algorithm to generate a fault analysis report, it enables precise analysis of the vibration status of amusement rides and improves the accuracy of fault diagnosis. This step not only enhances the ability to identify early-stage faults but also provides a reliable basis for subsequent health assessments, ultimately improving overall safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 This is a flow chart of the safety detection method for large-scale amusement facilities based on multiple sensors.

[0041] Figure 2 Flowchart of risk warning decision-making.

[0042] Figure 3 Flowchart for fault feature analysis.

[0043] Figure 4 Flowchart of multimodal data processing. DETAILED DESCRIPTION

[0044] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0045] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0047] Example 1, with reference to Figures 1 to 4 This embodiment provides a large-scale amusement facility safety detection method based on multiple sensors, including the following steps:

[0048] S1. Real-time collection of vibration data, temperature data, and acoustic signals from the stress-bearing parts of large-scale amusement facilities and preprocessing to generate a multimodal dataset.

[0049] Preprocessing includes time alignment of multi-source sensor data through a high-precision clock synchronization protocol, Kalman filtering to eliminate noise interference, and normalization.

[0050] It should be noted that, first, a high-precision clock synchronization protocol is used to time-align the vibration data, temperature data, and acoustic signals to ensure strict synchronization of multi-source sensor data in the time domain; then, the original signals collected by various sensors are subjected to noise elimination processing through the Kalman filtering algorithm to effectively suppress environmental interference and measurement errors; finally, the time-aligned and denoised data are normalized, and the numerical ranges of the vibration data, temperature data, and acoustic signals are uniformly mapped to the standard interval, eliminating the magnitude differences caused by different physical dimensions, and generating a consistent and comparable multimodal data set.

[0051] S2. Based on the multimodal data set, the time-frequency characteristics of the vibration signal are extracted through fast Fourier transform and wavelet analysis, and the fault analysis report is generated by combining the support vector machine algorithm;

[0052] Based on the multimodal data set, the vibration signal is spectrally analyzed by fast Fourier transform to generate the spectrum feature vector;

[0053] It should be noted that based on the vibration signal components in the multimodal data set, the fast Fourier transform algorithm is used to convert the time domain vibration waveform into a frequency domain representation. The amplitude spectral density distribution of each frequency component is calculated through spectral analysis, and the energy distribution information of the characteristic frequency band including the fundamental frequency and harmonic components is extracted. At the same time, the resonance peak position and its relative intensity relationship are identified, and finally a multi-dimensional spectral feature vector is constructed to fully characterize the spectral characteristics of the vibration signal.

[0054] Perform time-frequency analysis on the spectrum feature vector through wavelet analysis to generate a time-frequency feature matrix;

[0055] It should be noted that the vibration signal is analyzed in a time-frequency joint manner using wavelet analysis. Daubechies wavelet basis functions are selected to perform multi-scale decomposition of the signal. Continuous wavelet transforms are used to calculate wavelet coefficients at different scales, thereby obtaining the energy distribution characteristics of the vibration signal in the time-frequency domain. The time-frequency localization characteristics of the transient impact component are also analyzed, ultimately constructing a time-frequency feature matrix containing three-dimensional information: time, frequency, and energy. The wavelet analysis uses the Mallat algorithm for rapid decomposition. The rows of the time-frequency feature matrix correspond to wavelet decomposition levels of different scales, while the columns correspond to time sampling points. The matrix elements store the normalized wavelet coefficient amplitudes.

[0056] Based on the time-frequency feature matrix, the vibration state of the amusement facilities is analyzed through the support vector machine classifier to generate a fault analysis report.

[0057] It should be noted that the vibration state of the amusement facility is analyzed by the support vector machine classifier. First, the time-frequency feature matrix is ​​expanded into a feature vector and input into the support vector machine classifier. The pre-trained kernel function is used to map the feature space. The optimal classification hyperplane is constructed according to the principle of structural risk minimization. The vibration state modes such as normal state, early fault and serious fault are classified and judged. Finally, a fault analysis report containing diagnostic conclusions such as fault type, severity and location is output.

[0058] S3. Based on the fault analysis report, the temperature signal characteristics and acoustic signal characteristics of the fault area are extracted through principal component analysis, and multimodal feature fusion is performed through canonical correlation analysis to generate a comprehensive feature map;

[0059] Based on the fault analysis report, feature extraction is performed on the multimodal dataset through principal component analysis to generate temperature feature vectors and acoustic feature vectors;

[0060] It should be noted that based on the fault analysis report, principal component analysis is performed on the temperature data and acoustic signals at the corresponding positions in the multimodal data set. First, the covariance matrix of the temperature data samples is calculated, and the eigenvalues ​​and eigenvectors are solved by the singular value decomposition algorithm. The main component directions are selected in order of eigenvalue size; at the same time, the same processing flow is performed on the acoustic signal samples. Subsequently, the original temperature data matrix is ​​projected into the temperature principal component space, and the acoustic signal matrix is ​​projected into the acoustic principal component space, achieving feature dimensionality reduction while retaining the main variance information of the original data. The two sets of eigenvectors finally generated are: temperature eigenvectors composed of the main temperature principal component coefficients, and acoustic eigenvectors composed of the main acoustic principal component coefficients. Among them, the principal component analysis uses an iterative algorithm to accelerate the calculation, and all eigenvectors are standardized to ensure that the features of each dimension have a uniform dimension.

[0061] Based on the temperature eigenvector and acoustic eigenvector, multimodal feature fusion is performed through the canonical correlation analysis method, and the fusion feature matrix is ​​generated through the canonical correlation analysis algorithm;

[0062] It should be noted that based on the temperature eigenvector and the acoustic eigenvector, a canonical correlation analysis method is used to establish a statistical association model between the two sets of features. By solving the cross-covariance matrix of the temperature eigenvector and the acoustic eigenvector, the canonical variable pair corresponding to the maximum correlation coefficient is calculated. The original temperature eigenvector and the acoustic eigenvector are projected into the canonical correlation space respectively to obtain the feature representation with the maximum correlation. Finally, the two sets of canonical variables are linearly combined using the canonical correlation analysis algorithm to generate a fused feature matrix containing both temperature and acoustic feature information. Among them, the canonical correlation analysis uses the singular value decomposition algorithm to solve the feature projection direction. Each row in the fused feature matrix represents the joint feature representation of different samples in the canonical correlation space.

[0063] Based on the fused feature matrix, the feature space is probabilistically modeled through kernel density estimation to generate a comprehensive feature map.

[0064] It should be noted that the Gaussian kernel function is selected as the basic kernel function to calculate the neighborhood density contribution of each sample point in the fusion feature matrix in the feature space, and the discrete sample points are converted into a continuous probability density distribution through kernel function smoothing. During the modeling process, the kernel function bandwidth parameters are automatically adjusted according to the sample distribution characteristics of the feature space to ensure that the probability density can be accurately estimated for both sparse and dense areas. Then a multidimensional joint probability density function is established to characterize the statistical correlation characteristics between temperature features and acoustic features. The final generated comprehensive feature map adopts a visualization form that combines three-dimensional surfaces and two-dimensional contour lines to fully present the probability density distribution status of different regions in the feature space, where the surface height represents the probability density value and the contour lines reflect the gradient change of the probability density. The maximum likelihood estimation method is used to optimize the kernel function parameters throughout the modeling process to ensure the goodness of fit between the probability model and the sample data.

[0065] S4. Based on the comprehensive feature map, the health status of the amusement facilities is quantitatively evaluated using the Bayesian inversion algorithm to generate a health status diagnosis report;

[0066] Based on the comprehensive feature map, the Bayesian inversion algorithm is used to perform parameter inversion on the fault probability distribution, generating a quantized matrix of the amusement facility state containing the posterior probability.

[0067] It should be noted that the probabilistic relationship between fault characteristics and equipment status is established based on the comprehensive feature map, and the initial probability distribution is determined by combining historical equipment operation data. The probability estimate is then iteratively calculated and updated using the Bayesian theorem. During the inversion process, the Markov Chain Monte Carlo sampling method is used to perform random sampling in the probability space and calculate the posterior probability distribution of each fault state. Ultimately, a quantization matrix for the amusement facility state is generated. The rows of this matrix correspond to different equipment detection locations, and the columns correspond to various potential fault states. The matrix element values ​​represent the posterior probability of each fault state occurring at each location, fully quantitatively representing the health status of the amusement facility. The probability calculation uses the conjugate gradient optimization method to ensure that the probability estimate of the state quantization matrix is ​​accurate and reliable.

[0068] Based on the quantification matrix of the amusement facility status, the health of the amusement facility is evaluated at multiple levels through fuzzy logic reasoning to generate a health diagnosis score. The formula is as follows:

[0069]

[0070] Among them, H is the health status score of the amusement facility, Q is the energy matrix of the amusement facility status, w is the failure mode weight vector, μ k is the membership function of the kth fuzzy rule, l is the discrete value of the health level, (l) represents the confidence of the kth fuzzy rule for health level l, which is used to quantify the strength of the rule's judgment on a specific health state; K represents the total number of rules in the fuzzy rule base, k is the rule index of the current operation, and T is the transposition symbol;

[0071] It should be noted that first, the weighted combination w of the fault mode weight vector w and the amusement facility state quantization matrix Q is calculated. T Q, forming a comprehensive state evaluation vector. Then the kth rule in the fuzzy rule base Calculate the Gaussian membership function μ k , determine the degree of satisfaction of the prerequisite. k (w T Q) and the confidence level of rule conclusions Perform min operation to get the rule output strength. Aggregate the support strength of all rules for health level l. Finally, select The health level with the maximum value is taken as the diagnosis result H.

[0072] Generate a health diagnosis report based on the health diagnosis score.

[0073] It should be noted that the current health status category of the amusement facility is determined according to the health level corresponding to the health diagnosis score and the preset health level classification standards. Then, combined with the failure probability distribution of each detection location in the amusement facility status quantification matrix, the health status assessment results of each key part are recorded in detail. The high-risk areas that require special attention and their corresponding failure modes are clearly marked in the report, and maintenance recommendations based on the health diagnosis score H are also provided. The report format strictly follows industry standards and specifications, and includes necessary content such as basic equipment information, detection time, overall health status evaluation, detailed diagnosis results of each part, and subsequent maintenance recommendations.

[0074] S5. Based on the health status diagnosis report, the risk level of different areas is evaluated through fuzzy Bayesian network, and the detection parameters are optimized through particle swarm optimization algorithm to generate differentiated detection plans for different risk areas;

[0075] Based on the health status diagnosis report, the risk probability of each area of ​​the amusement facility is quantitatively calculated through fuzzy Bayesian to generate a regional safety assessment map;

[0076] It should be noted that a fuzzy Bayesian approach is used to quantitatively assess the risk probability of each area of ​​an amusement facility. First, the health diagnosis scores and fault probability distributions in the health status diagnosis report are analyzed, and an inference framework combining fuzzy rules and Bayesian networks is established. The health diagnosis scores are converted into fuzzy membership degrees, which serve as soft evidence inputs for the Bayesian network. The probabilistic dependencies between areas are constructed, taking into account the structural characteristics of the equipment. The Bayesian theorem is used to recursively calculate the posterior risk probability of each area under different health status conditions, and a fuzzy comprehensive evaluation method is employed to address uncertainty in the assessment process. The resulting regional safety assessment map is presented as a heat map, where different color depths intuitively reflect the risk probability level of each area, fully displaying the overall safety status of the amusement facility. The assessment process strictly adheres to the mathematical principles of probabilistic graphical models and fuzzy logic to ensure the accuracy and reliability of the risk assessment results.

[0077] Based on the regional safety assessment map, an improved particle swarm optimization algorithm is used to perform multi-objective optimization on core parameters such as monitoring frequency and sampling accuracy to generate an optimal detection parameter configuration table;

[0078] It should be noted that, based on the risk level of each area in the regional safety assessment map, an optimization function is established with the goals of maximizing detection efficiency and minimizing resource consumption, and parameter value range constraints are set. During the optimization process, the improved particle swarm optimization algorithm uses a dynamic inertia weight adjustment mechanism to enable particles to maintain strong global exploration capabilities in the early stages of the search and enhance local development accuracy in the later stages. At the same time, the Pareto dominance relationship is used to sort particle positions for non-inferior solutions, maintaining the diversity and convergence of the optimal solution set. By iteratively updating particle speeds and positions, a parameter table containing the optimal monitoring frequency and sampling accuracy configuration for each area is ultimately generated. High-risk areas correspond to higher monitoring frequencies and sampling accuracy, while low-risk areas use lower parameter configurations. The optimization process strictly follows the computational process of the multi-objective evolutionary algorithm to ensure that the generated detection parameter configuration table achieves the optimal balance between safety and economy.

[0079] Based on the optimal detection parameter configuration table, detection resources are dynamically allocated through the space-time dual-dimensional scheduling algorithm to generate differentiated detection implementation plans.

[0080] It should be noted that the monitoring frequency and sampling accuracy requirements for each area in the optimal detection parameter configuration table are analyzed, and a scheduling framework with spatial location dimensions and time series dimensions as coordinates is established. In the spatial dimension, detection resources are preferentially allocated to high-risk key areas based on the physical layout of the amusement facilities and the regional risk level; in the time dimension, a round-robin scheduling strategy is used to reasonably arrange the detection sequence based on the monitoring frequency requirements specified for each area. The detection tasks are managed through dynamic priority queues to ensure that high-frequency monitoring areas obtain sufficient detection resources while ensuring the basic coverage requirements of low-frequency monitoring areas. The final differentiated detection implementation plan details the spatial deployment plan of each detection resource in different time periods, forming an optimal scheduling plan that takes into account both detection effect and resource utilization efficiency. The scheduling process strictly adheres to the basic principles of the real-time scheduling algorithm to ensure the feasibility and execution efficiency of the implementation plan.

[0081] S6. Based on the differentiated detection scheme, multi-source information fusion decision-making is carried out through DS evidence theory to generate graded security warnings.

[0082] Based on the differentiated detection scheme, the probability distribution model of sensor data is carried out through the DS evidence theory algorithm to generate a set of evidence bodies for different risk levels;

[0083] It should be noted that, based on the risk classification criteria determined by the differentiated detection scheme, various types of sensor data are classified and processed according to risk level, and the distribution characteristics of sensor data at each risk level are estimated using a probability density function. A basic probability allocation function is then constructed to assign a probability mass reflecting the credibility of each risk level, forming a set of evidence bodies containing various propositions such as normal state, warning state, and dangerous state. The DS evidence theory algorithm uses the maximum entropy principle to determine the prior probability distribution. Each evidence body in the evidence body set contains a quantitative value for the degree of support for each proposition within the identification framework, fully characterizing the statistical characteristics of sensor data at different risk levels.

[0084] Based on the evidence set, the conflicting evidence sets are orthogonally summed using the Dempster combination rule to generate a consistent probability distribution after fusion.

[0085] It should be noted that the evidence from different sensor sources in the evidence set is combined pairwise, and the joint support for each proposition is calculated through orthogonal summation. Conflicting evidence is handled using a normalization factor. During this combination process, the vibration data evidence, temperature data evidence, and acoustic signal evidence are fused step by step, and the basic probability assignment for each risk level proposition is recalculated with each fusion. The resulting consistent probability distribution fully reflects the comprehensive assessment results of the multi-source sensor data fusion, with each risk level corresponding to a composite probability value rigorously derived from evidence theory.

[0086] Based on the consistent probability distribution, the security risks are graded using the confidence interval decision criterion and graded security warnings are generated.

[0087] It should be noted that the confidence function value and plausibility function value of each risk level proposition are extracted from the consistent probability distribution to calculate the confidence interval range. By comparing the degree of overlap and boundary thresholds of the confidence intervals of different risk levels, the optimal risk level division scheme is determined, which takes into account the degree of confidence and uncertainty range supported by the evidence. The final generated graded safety warning includes multiple warning levels such as normal state, attention state, alert state and dangerous state. Each level corresponds to a clear description of the safety state and disposal recommendations. The decision-making process strictly follows the mathematical definition of the confidence interval in evidence theory to ensure the logical consistency of the risk level division results and the probability distribution. At the same time, the warning information fully covers all types of potential risk conditions.

[0088] This embodiment also provides a large-scale amusement facility safety detection system based on multiple sensors, including: a data acquisition module, a fault feature analysis module, a feature fusion module, a health status assessment module, a risk detection optimization module and a safety warning module; the data acquisition module collects vibration data, temperature data and acoustic signals of the stress-bearing parts of the large-scale amusement facility in real time and pre-processes them to generate a multi-modal data set; the fault feature analysis module extracts the time-frequency features of the vibration signal through fast Fourier transform and wavelet analysis based on the multi-modal data set, and generates a fault analysis report in combination with the support vector machine algorithm; the feature fusion module extracts the fault area through principal component analysis based on the fault analysis report The temperature signal characteristics and acoustic signal characteristics of the domain are analyzed, and multimodal features are fused through the canonical correlation analysis method to generate a comprehensive feature map. The health status assessment module, based on the comprehensive feature map, uses the Bayesian inversion algorithm to quantitatively assess the health status of the amusement facility and generate a health status diagnosis report. The risk detection optimization module, based on the health status diagnosis report, uses the fuzzy Bayesian network to evaluate the risk level of different areas, and optimizes the detection parameters through the particle swarm optimization algorithm to generate differentiated detection schemes for different risk areas. The safety warning module, based on the differentiated detection scheme, uses the DS evidence theory to make multi-source information fusion decisions and generate graded safety warnings. This embodiment also provides a computer device suitable for the case of a large-scale amusement facility safety detection method based on multiple sensors, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the large-scale amusement facility safety detection method based on multiple sensors as proposed in the above embodiment.

[0089] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0090] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the large-scale amusement facility safety detection method based on multiple sensors as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0091] In summary, this invention achieves precise analysis of amusement ride vibration status and improves the accuracy of fault diagnosis by using fast Fourier transform and wavelet analysis to extract the time-frequency characteristics of vibration signals and combining this with a support vector machine algorithm to generate fault analysis reports. This step not only enhances the ability to identify early-stage faults but also provides a reliable basis for subsequent health assessments, ultimately improving overall safety.

[0092] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A multi-sensor based safety detection method for large-scale amusement facilities, characterized by: include, Real-time collection of vibration data, temperature data, and acoustic signals from the stress-bearing parts of large-scale amusement facilities and pre-processing to generate multimodal data sets; Based on multimodal data sets, fast Fourier transform and wavelet analysis are used to extract the time-frequency characteristics of vibration signals, and support vector machine algorithms are used to generate fault analysis reports. Based on the fault analysis report, the temperature signal characteristics and acoustic signal characteristics of the fault area are extracted through principal component analysis, and multimodal feature fusion is performed through canonical correlation analysis to generate a comprehensive feature map; Based on the comprehensive feature map, the Bayesian inversion algorithm is used to quantitatively evaluate the health status of amusement facilities and generate a health status diagnosis report; Based on the health status diagnosis report, the risk level of different areas is evaluated through fuzzy Bayesian network, and the detection parameters are optimized through particle swarm optimization algorithm to generate differentiated detection plans for different risk areas; Based on the differentiated detection scheme, multi-source information fusion decision-making is carried out through DS evidence theory to generate graded security warnings.

2. The multi-sensor based large-scale amusement facility safety detection method according to claim 1, characterized in that: The preprocessing includes time alignment of multi-source sensor data through a high-precision clock synchronization protocol, elimination of noise interference by using Kalman filtering, and normalization processing.

3. The multi-sensor based large-scale amusement facility safety detection method according to claim 2, characterized in that: Based on the multimodal data set, the time-frequency characteristics of the vibration signal are extracted by fast Fourier transform and wavelet analysis, and the fault analysis report is generated by combining the support vector machine algorithm. The specific steps are as follows: Based on the multimodal data set, the vibration signal is spectrally analyzed by fast Fourier transform to generate the spectrum feature vector; Perform time-frequency analysis on the spectrum feature vector through wavelet analysis to generate a time-frequency feature matrix; Based on the time-frequency feature matrix, the vibration state of the amusement facilities is analyzed through the support vector machine classifier to generate a fault analysis report.

4. The multi-sensor based large-scale amusement facility safety detection method according to claim 3, characterized in that: Based on the fault analysis report, the temperature signal characteristics and acoustic signal characteristics of the fault area are extracted through principal component analysis, and multi-modal feature fusion is performed through the canonical correlation analysis method to generate a comprehensive feature map. The specific steps are as follows: Based on the fault analysis report, feature extraction is performed on the multimodal dataset through principal component analysis to generate temperature feature vectors and acoustic feature vectors; Based on the temperature eigenvector and acoustic eigenvector, multimodal feature fusion is performed through the canonical correlation analysis method, and the fusion feature matrix is ​​generated through the canonical correlation analysis algorithm; Based on the fusion feature matrix, the feature space is probabilistically modeled through kernel density estimation to generate a comprehensive feature map.

5. The multi-sensor based large-scale amusement facility safety detection method according to claim 4, characterized in that: Based on the comprehensive feature map, the Bayesian inversion algorithm is used to quantitatively evaluate the health status of the amusement facility and generate a health status diagnosis report. The specific steps are as follows: Based on the comprehensive feature map, the Bayesian inversion algorithm is used to perform parameter inversion on the fault probability distribution and generate a quantized matrix of the amusement facility state containing the posterior probability. Based on the quantification matrix of the amusement facility status, the health of the amusement facility is evaluated at multiple levels through fuzzy logic reasoning to generate a health diagnosis score H; Generate a health diagnosis report based on the health diagnosis score.

6. The multi-sensor based large-scale amusement facility safety detection method according to claim 5, characterized in that: Based on the health status diagnosis report, the risk level of different areas is evaluated through fuzzy Bayesian network, and the detection parameters are optimized through particle swarm optimization algorithm to generate differentiated detection plans for different risk areas. The specific steps are as follows: Based on the health status diagnosis report, the risk probability of each area of ​​the amusement facility is quantitatively calculated through fuzzy Bayesian to generate a regional safety assessment map; Based on the regional safety assessment map, an improved particle swarm optimization algorithm is used to perform multi-objective optimization on core parameters such as monitoring frequency and sampling accuracy to generate an optimal detection parameter configuration table; Based on the optimal detection parameter configuration table, detection resources are dynamically allocated through the space-time dual-dimensional scheduling algorithm to generate differentiated detection implementation plans.

7. The multi-sensor based large-scale amusement facility safety detection method according to claim 6, characterized in that: The differentiated detection scheme uses DS evidence theory to make multi-source information fusion decisions and generate hierarchical security warnings. The specific steps are as follows: Based on the differentiated detection scheme, the probability distribution model of sensor data is carried out through the DS evidence theory algorithm to generate a set of evidence bodies for different risk levels; Based on the evidence set, the conflicting evidence sets are orthogonally summed using the Dempster combination rule to generate a consistent probability distribution after fusion. Based on the consistent probability distribution, the security risks are graded using the confidence interval decision criterion and graded security warnings are generated.

8. A multi-sensor based large-scale amusement facility safety detection system, based on the multi-sensor based large-scale amusement facility safety detection method according to any one of claims 1 to 7, characterized in that: Including data acquisition module, fault feature analysis module, feature fusion module, health status assessment module, risk detection optimization module and safety warning module; The data acquisition module collects vibration data, temperature data, and acoustic signals from the stress-bearing parts of large-scale amusement facilities in real time and pre-processes them to generate multimodal data sets; The fault feature analysis module, based on a multimodal data set, extracts the time-frequency features of vibration signals through fast Fourier transform and wavelet analysis, and generates a fault analysis report in combination with the support vector machine algorithm; The feature fusion module, based on the fault analysis report, extracts the temperature signal characteristics and acoustic signal characteristics of the fault area through principal component analysis, and performs multimodal feature fusion through canonical correlation analysis to generate a comprehensive feature map; The health status assessment module uses the Bayesian inversion algorithm to quantitatively assess the health status of amusement facilities based on the comprehensive feature map and generate a health status diagnosis report; The risk detection optimization module uses a fuzzy Bayesian network to assess the risk levels of different areas based on health status diagnostic reports, and optimizes detection parameters using a particle swarm optimization algorithm to generate differentiated detection plans for different risk areas. The security warning module, based on differentiated detection schemes, makes multi-source information fusion decisions through DS evidence theory to generate graded security warnings.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the large-scale amusement facility safety detection method based on multiple sensors as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the multi-sensor based large-scale amusement facility safety detection method according to any one of claims 1 to 7 are implemented.

Images