Cableway daily report large screen system based on multi-source data alignment and analysis and operation management method

Abstract

The invention provides a cableway daily report large screen system based on multi-source data alignment and analysis and an operation management method, and belongs to the technical field of cableway operation management, and the system comprises a multi-source data collection module which uniformly accesses and collects cableway sensor data, video monitoring data, scenic spot ticketing system data and scenic spot meteorological data in real time; the data fusion and alignment module performs timestamp synchronization, format standardization and linear interpolation aggregation on the accessed heterogeneous data based on a dynamic sliding window algorithm to generate multi-dimensional feature vectors with uniform granularity; the intelligent analysis module inputs the multi-dimensional feature vectors into a safety early warning model, a passenger flow prediction model and a sales trend prediction model for cableway safety early warning analysis, passenger flow prediction analysis and sales trend prediction analysis to generate an intelligent analysis result; and the visual large-screen terminal displays and outputs the intelligent analysis result. Multi-source data are integrated, intelligent early warning and visualization are achieved, cableway operation is optimized, and safety, efficiency and benefits are improved.

Description

Technical Field

[0001] This application belongs to the field of cableway operation and management technology, specifically involving a cableway daily report screen system and operation management method based on multi-source data alignment and analysis. Background Technology

[0002] With the development of tourism, cableways, as a vital mode of transportation connecting tourists and core attractions in mountainous scenic areas, are receiving increasing attention for their safety, comfort, and operational efficiency. Traditional cableway management methods mainly rely on manual inspections, report statistics, and static analysis, which are insufficient to meet the real-time, intelligent, and visual requirements of modern scenic areas. The development of the Internet of Things, cloud computing, and big data has provided new possibilities for cableway management. Some scenic areas have initially introduced video surveillance systems, ticketing systems, and weather sensors to achieve localized data collection and display. However, these systems often operate independently, resulting in inconsistent data and a lack of effective integration, leading to serious information silos and hindering support for overall optimization and rational decision-making.

[0003] Current cableway management systems suffer from the following main problems: First, they only support structured data sources, making it difficult to integrate unstructured or semi-structured data such as video streams, image recognition results, and meteorological APIs. This results in data sources being singular or having significantly different formats, making data fusion difficult. Second, most cableway management systems remain at the data display level, unable to determine the impact of weather on cable lifespan, or to predict sales trends and provide strategic feedback on the impact of peak passenger flow on operational strategies. Third, different cableway management systems use different timestamps, units, and granularities, and rely heavily on manual configuration of rules for system synchronization, which is costly and prone to errors.

[0004] Therefore, there is an urgent need for a cableway large-screen data display system that integrates multi-source heterogeneous data and enables unified display and correlation analysis to improve the safety, efficiency and economic benefits of cableway operation. Summary of the Invention

[0005] In a first aspect, embodiments of this application provide a cableway daily report display system based on multi-source data alignment and analysis, including a data processor and a visualization display terminal; The data processor is equipped with: The multi-source data acquisition module uniformly accesses cableway sensor data, video surveillance data, scenic area ticketing system data, and scenic area meteorological data, and performs real-time capture of structured and unstructured heterogeneous data from various data sources. The data fusion and alignment module, based on the dynamic sliding window algorithm, performs timestamp synchronization, format standardization, and linear interpolation aggregation on the accessed heterogeneous data to generate a multi-dimensional feature vector with uniform granularity. The intelligent analysis module inputs multi-dimensional feature vectors into pre-configured and trained safety early warning models, passenger flow prediction models, and sales trend prediction models to perform cableway safety early warning analysis, passenger flow prediction analysis, and sales trend prediction analysis, and generates intelligent analysis results. The large-screen visualization terminal displays and outputs the intelligent analysis results of the data processor in a visual format.

[0006] Furthermore, the multi-source data acquisition module includes: The cableway data acquisition unit collects cableway tension data in real time through a cableway tension sensor and cableway motor temperature data in real time through a cableway motor temperature sensor, and uses the cableway tension data and cableway motor temperature data as cableway data. The video surveillance data acquisition unit connects to the cableway's intelligent camera via the RTSP video stream protocol to acquire pedestrian detection and counting results, generate passenger flow data according to a set cycle, and push the passenger flow data to the visualization large screen terminal; The scenic area meteorological data acquisition unit acquires scenic area meteorological data at a set frequency through the map's open weather service interface; the scenic area meteorological data includes temperature, humidity, wind force, wind direction, and precipitation; The sales data collection unit obtains the scenic area's daily sales volume, number of orders, and ticket refund amount as sales data through the scenic area's ticketing system interface according to a set request frequency.

[0007] Furthermore, the data fusion and alignment module includes: The window settings unit configures the time window length W and step size S; The data segmentation unit uses a sliding window algorithm to slide a time window of length W across scenic area weather data, visitor flow data, sales data, and cable car data with a step size S, thereby cutting various types of data into segments of length W. The data alignment unit extracts and aggregates features from various types of data within the time window of each segment, generating a set of feature vectors for each time window.

[0008] Furthermore, the data alignment unit includes: The weather feature extraction subunit performs a linear interpolation algorithm on the temperature, humidity, wind force, wind direction and precipitation as meteorological data within the time window, and calculates the window boundary values; The interpolation boundary points of temperature and humidity within the time window are sorted with the original temperature and humidity data within the time window by time, and the trapezoidal method is used to calculate the interval average temperature and average humidity. Extract the maximum wind force value across all interpolation intervals within the time window; Calculate the sum of all precipitation within the time window to obtain the cumulative precipitation; The average temperature, average humidity, maximum wind speed, and cumulative precipitation within the time window are used as meteorological characteristics. The passenger flow feature extraction subunit acquires passenger flow data from video surveillance within a time window. The total passenger flow is calculated by accumulating the number of passengers in all directions within the time window. The maximum number of people entering the direction reported in a single time window within each time window is taken as the peak passenger flow. If the same tourist appears repeatedly within a time window, it is counted only once to deduplicat passenger flow. The deduplicated passenger flow and peak passenger flow are used as passenger flow characteristics; The sales feature extraction subunit acquires incremental data on orders and refund records from the scenic area's ticketing system within a time window; The total number of orders is obtained by summing up the number of new orders added within the time window; The total sales amount is obtained by summing up the payment amounts within the time window. The refund records within the time window are summed up to obtain the number of refunds and the refund amount; The total number of orders, total sales, number of refunds, and refund amount are used as sales characteristics; The cableway feature extraction subunit acquires second-level cableway data within a time window; Calculate the average value of all cableway tension data within the time window, and use it as the tension characteristic; The average, maximum, and minimum values ​​of all cableway motor temperature data within the calculated time window are used as temperature characteristics. Tension characteristics and temperature characteristics are used as cableway characteristics.

[0009] Furthermore, the intelligent analysis module includes: The safety early warning unit includes: The safety early warning model construction and training subunit sets the sliding window length W and step size S, completes the window configuration, and receives the maximum wind speed, average humidity, average tension, maximum motor temperature, number of passengers, and peak passenger flow after data alignment as the first input features. A dataset is constructed using historical first input feature data. Samples that have fault alarms within a set time period after the end of the time window are marked as positive samples, while samples that have not had fault alarms within the set time period are marked as negative samples. A security warning model was built based on the LightGBM binary classification model. It was trained using a dataset with positive and negative samples. The parameters of the security warning model were optimized by cross-validation, and the model training process was verified by recall. The safety early warning analysis subunit receives the first input feature within the time window according to the set time period, inputs it into the safety early warning model, outputs the risk probability, and triggers an alarm when the risk probability exceeds the set threshold. The passenger flow prediction unit includes: The passenger flow prediction model construction and training subunit sets the sliding window length W and step size S, completes the window configuration, and receives the total passenger flow, peak passenger flow, average temperature, cumulative precipitation, and number of sales orders after data alignment, and constructs a time series as the second input feature; The dataset is constructed using historical second input feature data, and then divided into a training set and a validation set. Normalize the mean and standard deviation of the training set into continuous features; A passenger flow prediction model is constructed based on the LSTM sequence regression model, with the second input feature data of the first time period in the past as the input and the total passenger flow and peak passenger flow of the second time period in the future as the output. The passenger flow prediction model is trained using a training set normalized to continuous features, and training is stopped when the validation set loss does not decrease for N consecutive rounds. The passenger flow prediction and analysis subunit receives the second input features within the time window according to the set time period, inputs them into the passenger flow prediction model, and outputs the total passenger flow and peak passenger flow for the second time period in the future. The sales trend forecasting unit includes: The sales trend prediction model construction and training subunit sets the sliding window length W and step size S, completes the window configuration, and receives the daily sales total, total number of orders, average temperature, cumulative precipitation, total passenger flow, and holiday identifiers after data alignment as the third input features; The dataset is constructed using the third input feature data, the continuous variables in the dataset are normalized, and the normalized dataset is divided into a training set and a validation set. A sales trend prediction model is built based on the XGBoost regression model, using historical third-input feature data as input and sales in a future set time period as output. The sales trend prediction model is trained using the training set. During the training process, the hyperparameters are optimized through cross-validation. Training is stopped when the RMSE on the validation set does not decrease for N consecutive rounds. The sales trend forecasting and analysis sub-unit inputs historical third-party input features into the sales forecasting model according to the set forecasting period to predict sales revenue for the future set time period.

[0010] Furthermore, in the safety early warning model construction and training sub-unit, the first input feature data within a set time period from the current time point is divided into a training set and a validation set according to a set ratio in chronological order; Balance positive and negative samples by adjusting the classification parameters of the LightGBM binary classification algorithm model; After the trained security prediction model is deployed on the processing server, the accuracy of the early warning and the false alarm rate are monitored regularly.

[0011] Furthermore, in the passenger flow prediction model construction and training sub-unit, The second input feature data from the current time point are concatenated into an input matrix; LSTM sequence regression models have fully connected layers with memory units and neurons; After the LSTM sequence regression model is deployed to the processing server, the prediction bias is dynamically adjusted based on real-time passenger flow data.

[0012] Furthermore, in the sales trend prediction model construction and training sub-unit, Holiday identifiers are used as one-hot encodings, and together with the normalized third input feature as a continuous feature, they are used as inputs to the sales trend prediction model. Optimize the decision tree depth, learning rate, and subsampling rate parameters during the training process of the sales trend prediction model; The sales trend forecasting model outputs forecast results periodically.

[0013] Furthermore, the large-screen visualization terminal includes: The data receiving unit communicates with the data processor in real time via the WebSocket protocol to receive security warning levels, passenger flow forecast results, and sales trend data. The multi-view display unit dynamically displays the real-time values ​​and warning levels of cableway tension and motor temperature in the first view. The second view overlays real-time passenger flow distribution and predicted peak areas based on a geographic information system. The third view displays a comparison between the daily sales and the predicted value in the form of a line graph.

[0014] Secondly, embodiments of this application also provide a cableway operation management method for the system described in the first aspect, comprising the following steps: S1. Real-time collection of cableway sensor data, video surveillance data, scenic area ticketing system data, and scenic area weather data to obtain heterogeneous data; S2. Based on the dynamic sliding window algorithm, timestamp synchronization, format standardization and linear interpolation aggregation are performed on the collected heterogeneous data to generate multidimensional feature vectors with uniform granularity; S3. Input the multi-dimensional feature vector into the pre-configured trained safety early warning model, passenger flow prediction model and sales trend prediction model to perform cableway safety early warning analysis, passenger flow prediction analysis and sales trend prediction analysis, and generate intelligent analysis results; S4. Output the intelligent analysis structure to a large-screen visualization terminal for visual display.

[0015] Thirdly, embodiments of this application also provide a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the cableway operation management method as described in the second aspect.

[0016] As can be seen from the above technical solutions, this application has the following advantages: The cableway daily report screen system and operation management method based on multi-source data alignment and analysis provided in this application comprehensively integrates cableway operation-related data through the synergy of multi-source data collection, fusion alignment, intelligent analysis and visualization display. This enables all-round, dynamic and intelligent monitoring and decision support, timely detection and early warning of safety hazards, optimization of energy consumption configuration and precise marketing, improvement of cableway operation safety, efficiency and economic benefits, and realization of intelligent operation and refined management of scenic spots. Attached Figure Description

[0017] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the cableway daily report display system based on multi-source data alignment and analysis according to the present invention.

[0019] Figure 2 This is a flowchart illustrating the cableway daily report large screen operation management method of the present invention. Detailed Implementation

[0020] The various embodiments of this disclosure will be described more fully in the following detailed description of the cableway daily report display system based on multi-source data alignment and analysis. This disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of this disclosure to the specific embodiments disclosed herein, but rather this disclosure should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of this disclosure.

[0021] For example, with the development of tourism, cableways play a crucial role in high-altitude scenic areas, connecting tourists with core attractions. Their safety, comfort, and operational efficiency are receiving increasing attention. Traditional cableway management methods, relying mainly on manual inspections, report statistics, and static analysis, are no longer sufficient to meet the modern scenic area's demands for real-time, intelligent, and visualized solutions. The continuous advancements in the Internet of Things, cloud computing, and big data have brought convenience to cableway management. Currently, some scenic areas have initially introduced video surveillance systems, ticketing systems, and weather sensors, achieving preliminary local data collection and display functions. However, most of these systems operate independently, lacking unified data standards and hindering effective integration. This results in significant information silos, making it difficult to support overall optimization and scientifically sound decision-making.

[0022] The existing cableway management systems currently suffer from the following limitations: First, most existing cableway management systems are only compatible with structured data sources, and they face difficulties in integrating unstructured or semi-structured data such as video streams, image recognition results, and meteorological APIs. This results in data sources that are either singular or have vastly different formats, making the integration process extremely difficult. Second, most cableway management systems only focus on basic data display and cannot analyze the potential impact of weather factors on the lifespan of steel cables. They are unable to optimize operational strategies based on peak passenger flow information, let alone conduct sales trend forecasting and formulate corresponding strategy feedback. Third, different cableway management systems differ in terms of timestamps, units, and granularity, and the synchronization between systems largely relies on manual rule configuration, which is not only costly but also prone to errors.

[0023] Therefore, it is urgent to develop a cableway large-screen data display system that can integrate multi-source heterogeneous data and achieve unified data display and correlation analysis. This is of vital importance to comprehensively improving the safety, efficiency and economic benefits of cableway operation.

[0024] To address the aforementioned issues, this embodiment provides a cableway daily report display system based on multi-source data alignment and analysis. This system integrates heterogeneous data from multiple sources, breaks down information silos, and enables unified display and intelligent analysis of cableway data. Through real-time monitoring, intelligent early warning, and predictive analysis, it can identify potential safety hazards in advance, optimize scheduling strategies, improve operational efficiency, facilitate intelligent management of scenic areas, and enhance the safety and economic benefits of cableway operations.

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 The diagram shown is a schematic of a cableway daily report display system based on multi-source data alignment and analysis in a specific embodiment. The system includes a data processor and a visualization display terminal. The data processor is equipped with: The multi-source data acquisition module uniformly accesses cableway sensor data, video surveillance data, scenic area ticketing system data, and scenic area meteorological data, and performs real-time capture of structured and unstructured heterogeneous data from various data sources. It should be noted that the multi-source data acquisition module can access various heterogeneous data sources related to cableway operation, enabling real-time capture of structured and unstructured data. This provides comprehensive and dynamically updated data resources for subsequent data processing and analysis, serving as the data foundation for intelligent monitoring and decision support, and ensuring that the real-time status of cableway operation and changes in the external environment can be reflected. The data fusion and alignment module, based on the dynamic sliding window algorithm, performs timestamp synchronization, format standardization, and linear interpolation aggregation on the accessed heterogeneous data to generate a multi-dimensional feature vector with uniform granularity. It should be noted that the data fusion and alignment module solves the problem of differences in timestamps, formats and granularity of multi-source heterogeneous data. By using a dynamic sliding window algorithm, it aligns and integrates different data into a multi-dimensional feature vector with a unified granularity, which effectively improves the availability and consistency of the data, provides reliable data input for the intelligent analysis module, and ensures the accuracy and credibility of the intelligent analysis results. It is a key link in realizing intelligent functions. The intelligent analysis module inputs multi-dimensional feature vectors into pre-configured and trained safety early warning models, passenger flow prediction models, and sales trend prediction models to perform cableway safety early warning analysis, passenger flow prediction analysis, and sales trend prediction analysis, and generates intelligent analysis results. It should be noted that by using pre-trained safety early warning models, passenger flow prediction models, and sales trend prediction models, and conducting in-depth analysis and mining of multi-dimensional feature vectors, a comprehensive intelligent assessment and prediction of cableway operation can be achieved. This enables the timely detection of potential safety hazards, advance prediction of passenger flow changes and sales trends, and provides accurate decision support for scenic area managers. It facilitates the optimization of scenic area operation strategies, the rational allocation of resources, and the improvement of service quality, ensuring the safe and efficient operation of the cableway and improving economic benefits. The large-screen visualization terminal displays and outputs the intelligent analysis results of the data processor in a visual format; It should be noted that by presenting the intelligent analysis results in an intuitive and clear visualization, managers of peripheral scenic areas can quickly obtain key information and understand important dynamics such as cable car operation status, passenger flow distribution, and sales, which facilitates real-time monitoring and rapid decision-making. At the same time, the intuitive visualization also helps to convey important information to scenic area staff and tourists, improve the transparency and management efficiency of scenic area operations, and enhance the tourist experience.

[0027] This embodiment, through the cooperation of a data processor and a large-screen visualization terminal, realizes the entire process from data collection, fusion and alignment to intelligent analysis and visualization display, breaking down data silos, mining data value, and providing scenic area managers with comprehensive, real-time and accurate cableway operation information to facilitate reasonable decision-making.

[0028] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process in this embodiment, another cableway daily report screen system based on multi-source data alignment and analysis is provided. This system includes a data processor and a visualization screen terminal. The data processor is equipped with: The multi-source data acquisition module unifies the access to cableway sensor data, video surveillance data, scenic area ticketing system data, and scenic area meteorological data, and performs real-time capture of structured and unstructured heterogeneous data from various data sources; the multi-source data acquisition module includes: The cableway data acquisition unit collects cableway tension data in real time through a cableway tension sensor and cableway motor temperature data in real time through a cableway motor temperature sensor, and uses the cableway tension data and cableway motor temperature data as cableway data. It should be noted that the real-time acquisition of relevant data from the cableway sensors is achieved through the MQTT protocol; The video surveillance data acquisition unit connects to the cableway's intelligent camera via the RTSP video stream protocol to acquire pedestrian detection and counting results, generate passenger flow data according to a set cycle, and push the passenger flow data to the visualization large screen terminal; The scenic area meteorological data acquisition unit acquires scenic area meteorological data at a set frequency through the map's open weather service interface; the scenic area meteorological data includes temperature, humidity, wind force, wind direction, and precipitation; It should be noted that the map's open weather service interface is a RESTful HTTP-based interface, and each request needs to include the scenic area's geographical coordinates and service authentication information; the latest data is obtained once every 30 minutes through a scheduled task. The sales data collection unit obtains the scenic area's daily sales, number of orders, and ticket refund amount as sales data through the scenic area's ticketing system interface according to a set request frequency; It should be noted that the scenic area ticketing system interface is called using an HTTP RESTful API. Each request uses the update time of the latest record in the scenic area ticketing system's database as the start time in the request parameters, and the current time as the end time. The request step is one hour, and requests are initiated according to the request frequency (e.g., every 10 minutes). Each request can retrieve the latest sales orders and refund orders within the requested time period. After retrieving the latest data, the system summarizes the day's sales volume, order count, and refund amount in real time based on the retrieved orders and the request frequency. The data fusion and alignment module, based on a dynamic sliding window algorithm, performs timestamp synchronization, format standardization, and linear interpolation aggregation on the incoming heterogeneous data to generate a multidimensional feature vector with uniform granularity. The data fusion and alignment module includes: The window settings unit configures the time window length W and step size S; The data segmentation unit uses a sliding window algorithm to slide a time window of length W across scenic area weather data, visitor flow data, sales data, and cable car data with a step size S, thereby cutting various types of data into segments of length W. The data alignment unit extracts and aggregates features from various data types within each time window, generating a set of feature vectors for each time window. The data alignment unit includes: The weather feature extraction subunit performs a linear interpolation algorithm on the temperature, humidity, wind force, wind direction and precipitation as meteorological data within the time window, and calculates the window boundary values; The logic of the linear interpolation algorithm: If there is no meteorological data at the beginning or end of the time window, then linear interpolation is performed based on the two most recent observations to calculate the boundary values; The interpolation boundary points of temperature and humidity within the time window are sorted with the original temperature and humidity data within the time window by time, and the trapezoidal method is used to calculate the interval average temperature and average humidity. Extract the maximum wind force value across all interpolation intervals within the time window; Calculate the sum of all precipitation within the time window to obtain the cumulative precipitation; The average temperature, average humidity, maximum wind speed, and cumulative precipitation within the time window are used as meteorological characteristics. The passenger flow feature extraction subunit acquires passenger flow data from video surveillance within a time window. The total passenger flow is calculated by accumulating the number of passengers in all directions within the time window. The maximum number of people entering the direction reported in a single time window within each time window is taken as the peak passenger flow. If the same tourist appears repeatedly within a time window, it is counted only once to deduplicat passenger flow. The deduplicated passenger flow and peak passenger flow are used as passenger flow characteristics; The sales feature extraction subunit acquires incremental data on orders and refund records from the scenic area's ticketing system within a time window; The total number of orders is obtained by summing up the number of new orders added within the time window; The total sales amount is obtained by summing up the payment amounts within the time window. The refund records within the time window are summed up to obtain the number of refunds and the refund amount; The total number of orders, total sales, number of refunds, and refund amount are used as sales characteristics; The cableway feature extraction subunit acquires second-level cableway data within a time window; Calculate the average value of all cableway tension data within the time window, and use it as the tension characteristic; The average, maximum, and minimum values ​​of all cableway motor temperature data within the calculated time window are used as temperature characteristics. Tension characteristics and temperature characteristics are used as cableway characteristics; The intelligent analysis module inputs multi-dimensional feature vectors into pre-configured, trained safety early warning models, passenger flow prediction models, and sales trend prediction models to perform cableway safety early warning analysis, passenger flow prediction analysis, and sales trend prediction analysis, generating intelligent analysis results. The intelligent analysis module includes: The safety early warning unit includes: The safety early warning model construction and training subunit sets the sliding window length W and step size S, completes the window configuration, and receives the maximum wind speed, average humidity, average tension, maximum motor temperature, number of passengers, and peak passenger flow after data alignment as the first input features. A dataset is constructed using historical first input feature data. Samples that trigger fault alarms within a set time period (e.g., 30 minutes) after the end of the time window are marked as positive samples, while samples that do not trigger fault alarms within the set time period are marked as negative samples. A security warning model was built based on the LightGBM binary classification model. It was trained using a dataset with positive and negative samples. The parameters of the security warning model were optimized by cross-validation, and the model training process was verified by recall. The safety early warning analysis subunit receives the first input feature within the time window according to the set time period, inputs it into the safety early warning model, outputs the risk probability, and triggers an alarm when the risk probability exceeds the set threshold. The passenger flow prediction unit includes: The passenger flow prediction model construction and training subunit sets the sliding window length W and step size S, completes the window configuration, and receives the total passenger flow, peak passenger flow, average temperature, cumulative precipitation, and number of sales orders after data alignment, and constructs a time series as the second input feature; In the safety early warning model construction and training sub-unit, the first input feature data within a set time period from the current time point is divided into a training set and a validation set according to a set ratio in chronological order. Balance positive and negative samples by adjusting the classification parameters of the LightGBM binary classification algorithm model; After the trained security prediction model is deployed on the processing server, the accuracy of the early warning and the false alarm rate are monitored regularly. The dataset is constructed using historical second input feature data, and then divided into a training set and a validation set. Normalize the mean and standard deviation of the training set into continuous features; A passenger flow prediction model is constructed based on the LSTM sequence regression model, with the second input feature data of the first time period in the past as the input and the total passenger flow and peak passenger flow of the second time period in the future as the output. The passenger flow prediction model is trained using a training set normalized to continuous features, and training is stopped when the validation set loss does not decrease for N consecutive rounds. In the passenger flow prediction model construction and training sub-unit, The second input feature data from the current time point are concatenated into an input matrix; LSTM sequence regression models have fully connected layers with memory units and neurons; After the LSTM sequence regression model is deployed to the processing server, the prediction bias is dynamically adjusted based on real-time passenger flow data. The passenger flow prediction and analysis subunit receives the second input features within the time window according to the set time period, inputs them into the passenger flow prediction model, and outputs the total passenger flow and peak passenger flow for the second time period in the future. The sales trend forecasting unit includes: The sales trend prediction model construction and training subunit sets the sliding window length W and step size S, completes the window configuration, and receives the daily sales total, total number of orders, average temperature, cumulative precipitation, total passenger flow, and holiday identifiers after data alignment as the third input features; The dataset is constructed using the third input feature data, the continuous variables in the dataset are normalized, and the normalized dataset is divided into a training set and a validation set. A sales trend prediction model is built based on the XGBoost regression model, using historical third-input feature data as input and sales in a future set time period as output. The sales trend prediction model is trained using the training set. During the training process, the hyperparameters are optimized through cross-validation. Training is stopped when the RMSE on the validation set does not decrease for N consecutive rounds. In the sales trend prediction model construction and training sub-unit, Holiday identifiers are used as one-hot encodings, and together with the normalized third input feature as a continuous feature, they are used as inputs to the sales trend prediction model. Optimize the decision tree depth, learning rate, and subsampling rate parameters during the training process of the sales trend prediction model; The sales trend forecasting model outputs forecast results periodically. The sales trend forecasting and analysis sub-unit inputs historical third-party input features into the sales forecasting model according to the set forecasting period to predict the sales revenue for the future set time period. A large-screen visualization terminal displays and outputs the intelligent analysis results of the data processor in a visual format; the large-screen visualization terminal includes: The data receiving unit communicates with the data processor in real time via the WebSocket protocol to receive security warning levels, passenger flow forecast results, and sales trend data. The multi-view display unit dynamically displays the real-time values ​​and warning levels of cableway tension and motor temperature in the first view. The second view overlays real-time passenger flow distribution and predicted peak areas based on a geographic information system. The third view displays a comparison between the daily sales and the predicted value in the form of a line graph.

[0029] like Figure 2 As shown, the following are embodiments of the cableway operation management method provided in this disclosure. This method belongs to the same inventive concept as the cableway daily report screen system based on multi-source data alignment and analysis in the above embodiments. For details not described in detail in the embodiments of the cableway daily report screen system based on multi-source data alignment and analysis, please refer to the embodiments of the cableway operation management method described above.

[0030] The methods include: S1. Real-time collection of cableway sensor data, video surveillance data, scenic area ticketing system data, and scenic area weather data to obtain heterogeneous data; S2. Based on the dynamic sliding window algorithm, timestamp synchronization, format standardization and linear interpolation aggregation are performed on the collected heterogeneous data to generate multidimensional feature vectors with uniform granularity; S3. Input the multi-dimensional feature vector into the pre-configured trained safety early warning model, passenger flow prediction model and sales trend prediction model to perform cableway safety early warning analysis, passenger flow prediction analysis and sales trend prediction analysis, and generate intelligent analysis results; S4. Output the intelligent analysis structure to a large-screen visualization terminal for visual display.

[0031] This embodiment enables intelligent and refined management of cableway operations, ensuring the safe and stable operation of the cableway and improving the operational efficiency of the scenic area.

[0032] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0033] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A cableway daily report large screen system based on multi-source data alignment and analysis, characterized in that, The data processor and a visual large-screen terminal are included. The data processor is provided with: A multi-source data acquisition module, which uniformly accesses ropeway sensor data, video monitoring data, scenic spot ticketing system data and scenic spot meteorological data, and performs real-time grabbing of structured and unstructured heterogeneous data from each data source; A data fusion and alignment module, which, based on a dynamic sliding window algorithm, performs timestamp synchronization, format standardization and linear interpolation aggregation on the accessed heterogeneous data, to generate a multi-dimensional feature vector of uniform granularity; An intelligent analysis module, which inputs the multi-dimensional feature vector into a pre-configured trained safety early warning model, passenger flow prediction model and sales trend prediction model, to perform ropeway safety early warning analysis, passenger flow prediction analysis and sales trend prediction analysis, and generate intelligent analysis results; The visual large-screen terminal visualizes and outputs the intelligent analysis results of the data processor.

2. The gondola daily report large screen system based on multi-source data alignment and analysis according to claim 1, characterized in that, The multi-source data acquisition module includes: A ropeway data acquisition unit, which acquires ropeway tension data in real time through a ropeway tension sensor, and acquires ropeway motor temperature data in real time through a ropeway motor temperature sensor, and takes the ropeway tension data and the ropeway motor temperature data as ropeway data; A video monitoring data acquisition unit, which connects a ropeway intelligent camera through an RTSP video stream protocol, acquires pedestrian detection and counting results, generates passenger flow data according to a set period, and pushes the passenger flow data to the visual large-screen terminal; A scenic spot meteorological data acquisition unit, which acquires scenic spot meteorological data according to a set frequency through a map open weather service interface; the scenic spot meteorological data includes temperature, humidity, wind power, wind direction and precipitation; A sales data acquisition unit, which acquires daily sales, order quantity and refund amount of a scenic spot as sales data through a scenic spot ticketing system interface according to a set request frequency. 3.The gondola daily report large screen system based on multi-source data alignment and analysis of claim 2, wherein, The data fusion and alignment module includes: A window setting unit, which configures a time window length W and a step size S; A data segmentation unit, which uses a sliding window algorithm to slide a time window of length W on scenic spot meteorological data, passenger flow data, sales data and ropeway data according to the step size S, so as to cut the various types of data into segments of length W; A data alignment unit, which extracts and aggregates features of each type of data within each time window, and generates a set of feature vectors within each time window.

4. The cableway daily report large screen system based on multi-source data alignment and analysis according to claim 3, characterized in that, The data alignment unit includes: A weather feature extraction subunit, which performs a linear interpolation algorithm on temperature, humidity, wind power, wind direction and precipitation in the time window as meteorological data, to calculate window boundary values; The interpolated boundary points of temperature and humidity in the time window and the original temperature and humidity data in the time window are sorted by time, and the trapezoidal method is used to calculate the average temperature and average humidity in the interval; The maximum wind power in all interpolation intervals in the time window is taken out; The sum of all precipitation in the time window is calculated to obtain the cumulative precipitation; The average temperature, average humidity, maximum wind power and cumulative precipitation in the time window are taken as meteorological features; A passenger flow feature extraction subunit, which acquires passenger flow data in the video monitoring in the time window; The total passenger flow is calculated by adding up the passenger flow in all directions in the time window; The maximum value of the number of people in the same direction in each time window is reported as a peak passenger flow; The same visitor in the same time window is only recorded once, and the passenger flow is de-duplicated; The de-duplicated passenger flow and peak passenger flow are used as passenger flow characteristics; The sales feature extraction sub-unit obtains the incremental data of the scenic spot ticket system order and ticket refund record in the time window; The number of new orders in the time window is accumulated to obtain the total number of orders; The payment amount in the time window is accumulated to obtain the total sales; The ticket refund records in the time window are accumulated to obtain the number of refunds and the refund amount; The total number of orders, total sales, number of refunds, and refund amount are used as sales characteristics; The gondola feature extraction sub-unit obtains the second-level gondola data in the time window; The average value of all gondola tension data in the time window is calculated as a tension feature; The average value, maximum value, and minimum value of all gondola motor temperature data in the time window are calculated as temperature features; The tension feature and temperature feature are used as gondola features.

5. The gondola daily report large screen system based on multi-source data alignment and analysis according to claim 3, characterized in that, The intelligent analysis module includes: The safety warning unit includes: The safety warning model construction and training sub-unit sets the sliding window length W and step S, completes the window configuration, and receives the maximum wind speed, average humidity, average tension, maximum motor temperature, passenger flow, and peak passenger flow after data alignment as the first input features; The historical first input feature data is used to build a data set, and the samples in the set time period after the end of the time window that have occurred fault warnings are marked as positive samples, while the samples in the set time period that have not occurred fault warnings are marked as negative samples; The safety warning model is built based on the LightGBM binary classification model, the data set with positive samples and negative samples is used for training, and the safety warning model parameters are optimized through cross-validation, and the model training process is verified by recall rate; The safety warning analysis sub-unit receives the first input features in the time window according to the set time period, inputs the safety warning model, outputs the risk probability, and triggers the alarm when the risk probability exceeds the set threshold; The passenger flow prediction unit includes: The passenger flow prediction model construction and training sub-unit sets the sliding window length W and step S, completes the window configuration, and receives the total passenger flow, peak passenger flow, average temperature, cumulative precipitation, and sales order number after data alignment as the second input features, and builds a time series; The historical second input feature data is used to build a data set, and the data set is divided into a training set and a validation set; The mean and standard deviation of the training set are normalized as continuous features; The passenger flow prediction model is built based on the LSTM sequence regression model, and the second input feature data of the first time period in the past is used as the input, and the total passenger flow and peak passenger flow of the second time period in the future are used as the output; The training set normalized as continuous features is used to train the passenger flow prediction model, and the training is stopped when the validation set loss does not decrease continuously for N rounds; The passenger flow prediction analysis sub-unit receives the second input features in the time window according to the set time period, inputs the passenger flow prediction model, and outputs the total passenger flow and peak passenger flow of the second time period in the future; The sales trend prediction unit includes: The sales trend prediction model construction and training subunit sets a sliding window length W and a step S, completes window configuration, and receives the daily sales total, the total number of orders, the average temperature, the cumulative precipitation, the total passenger flow, and the holiday identifier after data alignment as the third input features; The third input feature data is used to construct a data set, the continuous variables in the data set are normalized, and the normalized data set is divided into a training set and a validation set; The sales trend prediction model is constructed based on an XGBoost regression model, the historical third input feature data is used as input, and the sales in a future set time period is used as output; The sales trend prediction model is trained using the training set, the hyperparameters are optimized through cross-validation during the training process, and the training is stopped when the RMSE of the validation set does not decrease for N consecutive rounds; The sales trend prediction analysis subunit inputs the historical third input features into the sales prediction model according to a set prediction period to predict the sales in a future set time period. 6.The gondola daily report large screen system based on multi-source data alignment and analysis of claim 5, wherein, In the safety warning model construction and training subunit, the first input feature data in a set time period from the current time point is divided into a training set and a validation set according to a set proportion in chronological order; The classification parameters of the LightGBM binary classification algorithm model are adjusted to balance the positive and negative samples; The trained safety prediction model is deployed on the processing server to periodically monitor the warning accuracy and false alarm rate. 7.The gondola daily report large screen system based on multi-source data alignment and analysis of claim 5, wherein, In the passenger flow prediction model construction and training subunit, The second input feature data from the current time point is concatenated into an input matrix; The LSTM sequence regression model has a fully connected layer with memory cells and neurons; After the LSTM sequence regression model is deployed to the processing server, the prediction bias is dynamically adjusted according to real-time passenger flow data. 8.The gondola daily report large screen system based on multi-source data alignment and analysis of claim 5, wherein, In the sales trend prediction model construction and training subunit, The holiday identifier is used as one-hot encoding, and the third input feature after normalization is used as continuous features as the input of the sales trend prediction model; The decision tree depth parameter, the learning rate, and the subsampling rate parameter are optimized during the training process of the sales trend prediction model; The sales trend prediction model outputs prediction results regularly. 9.The gondola daily report large screen system based on multi-source data alignment and analysis of claim 3, wherein, The visual large-screen terminal includes: The data receiving unit communicates with the data processor in real time through the WebSocket protocol, receives the safety warning level, the passenger flow prediction result, and the sales trend data; The multi-view display unit dynamically displays the real-time values and warning levels of the cableway tension and motor temperature through the first view, displays the real-time passenger flow distribution and predicted peak area based on the geographic information system through the second view, and displays the comparison between the daily sales and the predicted value in the form of a line chart through the third view.

10. A method of ropeway operation management based on the system according to any one of claims 1 to 9, characterized in that, The method includes the following steps: S1. Real-time acquisition of cableway sensor data, video monitoring data, scenic spot ticketing system data, and scenic spot meteorological data to obtain heterogeneous data; S2. Based on the dynamic sliding window algorithm, the collected heterogeneous data is synchronized by timestamp, standardized by format, and aggregated by linear interpolation to generate a unified multi-dimensional feature vector; S3. Input the multi-dimensional feature vector into the pre-configured trained safety warning model, passenger flow prediction model and sales trend prediction model for cableway safety warning analysis, passenger flow prediction analysis and sales trend prediction analysis, and generate intelligent analysis results; S4. Output the intelligent analysis structure to a visual large-screen terminal for visual display.

Images