A three-dimensional visualization-based intelligent park management platform

Abstract

The present invention relates to the technical field of three-dimensional visualization for park management, and specifically discloses a three-dimensional visualization-based intelligent park management platform, which includes an adaptive sensor calibration module that dynamically calculates the accuracy coefficient of sensor data to ensure the accuracy of data acquisition; a multi-dimensional data fusion and three-dimensional modeling module that calculates the adaptation coefficient by integrating multi-source data to evaluate the system's response ability to environmental changes; an intelligent image data registration module that optimizes the matching accuracy coefficient to improve the spatial docking accuracy between image data and the three-dimensional model; a real-time error monitoring and feedback optimization module that uses machine learning algorithms to correct the error between the sensor and the three-dimensional model, dynamically calculates the error correction coefficient to ensure data consistency; and a comprehensive performance evaluation module that obtains the comprehensive performance coefficient by normalizing and calculating the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient, and error correction coefficient, and optimizes the system operation strategy in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional visualization for park management, and particularly to a three-dimensional visualization-based intelligent park management platform. Background Art

[0002] With the continuous increase in the management requirements of intelligent parks, the traditional park management methods can no longer meet the modern needs. Currently, the management of intelligent parks is gradually developing towards digitalization, intelligence, and visualization. By collecting multi-dimensional data and performing three-dimensional modeling, precise monitoring and dynamic management of the park environment, equipment, and personnel can be achieved, which has become an important technical means for realizing efficient operation, safety management, and resource optimization. However, due to the complexity and dynamics of the park environment, how to accurately dock the data collected by sensors with the three-dimensional visualization model and achieve real-time correction and comprehensive analysis of the data remains the core challenge in the development of intelligent park technology.

[0003] The existing technologies have the following deficiencies:

[0004] In the existing technologies, the management platforms of intelligent parks usually have deficiencies in aspects such as the matching of sensor data with three-dimensional models, the adaptive adjustment to dynamic changes in the environment, error correction, and system performance evaluation. First, the data collected by sensors are affected by measurement errors and environmental interference, and spatial deviations often occur when docking with three-dimensional models, resulting in insufficient accuracy of the visualization results. Second, the existing systems have a weak response ability to dynamic changes in the park environment and cannot efficiently adapt to complex scenarios such as equipment position changes or building reconstructions. In addition, most of the error correction methods lack intelligent feedback mechanisms and are difficult to achieve dynamic optimization of sensor errors and modeling deviations. Finally, the existing management systems lack a comprehensive evaluation of the collaborative efficiency of multiple modules such as data collection, matching, and correction, resulting in poor overall system performance and restricting the accuracy and efficiency of park management. Summary of the Invention

[0005] The purpose of the present invention is to provide a three-dimensional visualization-based intelligent park management platform to solve the problems in the above background.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A three-dimensional visualization-based intelligent park management platform, comprising:

[0008] An adaptive sensor calibration module, which is used to automatically adjust the positions and measurement accuracies of sensors in the park according to dynamic changes in the park environment, including equipment position changes and building reconstructions, ensure the matching degree between the image data collected by the sensors and the actual space of the park, and calculate and dynamically update the sensor data accuracy coefficient for adjusting the measurement errors of the sensors;

[0009] Among them, the sensors include: camera sensors, lidar sensors, infrared sensors, and GPS sensors;

[0010] A multi-dimensional data fusion and 3D modeling module. The multi-dimensional data fusion and 3D modeling module collects multi-source data from lidar, UAV aerial photography, and video surveillance through sensors, generates and real-time updates the 3D model of the park, calculates and dynamically adjusts the adaptation coefficient, and is used to evaluate the self-adaptive adjustment ability of the system when the park changes;

[0011] An intelligent image data registration module. The intelligent image data registration module automatically identifies the calibration area in the park, matches the image of the calibration area with the 3D model in space, calculates and optimizes the matching accuracy coefficient, and is used to evaluate the accuracy of the docking between the image data and the 3D model;

[0012] A real-time error monitoring and feedback optimization module. The real-time error monitoring and feedback optimization module calculates and optimizes the error correction coefficient based on the real-time spatial error between the sensor and the 3D model, and is used to evaluate the error correction effect;

[0013] An overall performance evaluation module. The overall performance evaluation module evaluates the overall operation performance of the system according to the comprehensive calculation of the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient, and error correction coefficient, calculates the overall performance coefficient, and is used to dynamically optimize the system operation strategy and improve the accuracy and efficiency of park management.

[0014] As a further solution of the present invention: the process of obtaining the sensor data accuracy coefficient is as follows:

[0015] Deploy a number of sensors inside the park, and the position of each sensor is a three-dimensional coordinate ; among them, represents the sensor; use the sensor to collect the distance data of the preset reference points in the environment ; among them, represents the preset reference point;

[0016] Calculate the distance between the preset reference point and the sensor, and the calculation expression is: ; in the formula, represents the calculated distance between the th sensor and the th preset reference point; for each sensor and each preset reference point, calculate the measurement error between the th sensor and the th preset reference point, and the calculation expression is:

[0017] ; in the formula, represents the The measurement error of a sensor and the th preset reference point, represents the actual measured distance between the th sensor and the th preset reference point;

[0018] Calculate the total error of the sensor for all reference points. The calculation expression is: ; In the formula, represents the total error of the th sensor, represents the total number of preset reference points; For the th sensor, calculate the accuracy coefficient of the sensor data. The calculation expression is: ;

[0019] In the formula, represents the th sensor data accuracy coefficient.

[0020] As a further solution of the present invention: The acquisition process of the adaptation coefficient is as follows:

[0021] Real-time collect environmental data in the park through sensors, including sensor measurement error data and 3D model update data, to form a system state vector;

[0022] In the initial stage of the system, initialize the particle set through uniform distribution, where each particle represents a system state, and the state vector includes sensor error and 3D model error;

[0023] At each time step, predict the state of the current particle set through the state transition model, and add process noise for particle propagation to obtain the state estimate at the next moment;

[0024] According to the real-time collected sensor data and 3D model data, calculate the matching degree between each particle and the actual observation data, and update the weight of each particle through the observation model. The update of the weight is based on the likelihood of sensor data and 3D model update error;

[0025] Resample according to the weights of the particles to generate a new particle set to ensure the stability of the particle filter algorithm;

[0026] Based on the particle set obtained by particle filtering, calculate the adaptation coefficient of the system. The calculation expression is:

[0027] ; In the formula, represents the adaptation coefficient, represents the sensor, represents the th total error of the sensor, represents the The three-dimensional model update error of a sensor.

[0028] As a further solution of the present invention: by automatically identifying the calibration area in the park, matching the image of the calibration area with the three-dimensional model in space, calculating and optimizing the matching accuracy coefficient, specifically including:

[0029] Collecting the image data and three-dimensional model data in the park in real time through sensors, and automatically identifying the calibration area in the park to form a preliminary matching point pair of the image data and the three-dimensional model data;

[0030] Using a convolutional neural network to extract features from the image data, obtaining the deep feature information in the image, and representing it as a feature vector for subsequent spatial docking optimization;

[0031] Processing the three-dimensional model data, using an adapted deep learning network to extract features from the three-dimensional model to obtain the feature representation of the three-dimensional model for matching with the image features;

[0032] Matching the image features and the three-dimensional model features, and using the minimization of the matching error to calculate the spatial transformation matrix between the image and the three-dimensional model, including: rotation transformation and translation transformation, and optimizing the docking accuracy of the image features and the three-dimensional model features;

[0033] Based on the spatial docking result of the image and the three-dimensional model, calculate the matching error between each matching feature point, and the calculation expression is: ; In the formula, represents the matching error of the th matching point, represents the matching point, represents the th coordinate of the feature point in the image of the matching point, represents the th coordinate of the feature point in the three-dimensional model of the matching point, represents the calculated transformation matrix;

[0034] According to the calculated matching error, calculate the matching accuracy coefficient, and the calculation expression is: ; In the formula, represents the total number of matching points, represents the matching accuracy coefficient.

[0035] As a further solution of the present invention: evaluating the accuracy of the docking of the image data and the three-dimensional model,

[0036] According to the matching accuracy coefficient, optimize the matching process, automatically adjust the parameters of the convolutional neural network, and optimize the spatial matching accuracy of the image and the three-dimensional model;

[0037] Determine whether the matching accuracy coefficient is greater than or equal to the preset threshold. If so, the image data is accurately docked with the 3D model. If not, the image data is inaccurately docked with the 3D model.

[0038] As a further solution of the present invention: Calculate and optimize the error correction coefficient based on the real-time spatial error between the sensor and the 3D model, specifically including:

[0039] Collect real-time measurement data in the park through the sensor, including distance, angle, and speed eigenvalue;

[0040] Combine the actual reference values of the 3D model to form an input feature set and a target error value set for subsequent error correction modeling;

[0041] Calculate the mean value of the target error value as the initial prediction value, and the calculation expression is: ; In the formula, represents the sample in the target error value set, represents the total number of samples in the target error value set, represents the initial prediction value, represents the th sample's actual error;

[0042] Calculate the residual of the th sample, and the calculation expression is: ; In the formula, represents the residual of the th sample;

[0043] Adopt a recursive iteration method to construct several weak learners to gradually fit the residual and update the current prediction value, and the calculation expression is: ; In the formula, represents the learning rate, represents the updated prediction value, represents the prediction value at the previous moment before update, represents the th moment, represents the th sample at the th moment's regression decision tree;

[0044] Calculate the new residual according to the updated prediction value, and the calculation expression is: ; In the formula, represents the new residual of the th sample;

[0045] Based on the prediction error value and the actual error value obtained from the new residual, calculate the error correction coefficient, and the calculation expression is: ; In the formula, represents the error correction coefficient.

[0046] As a further solution of the present invention: The evaluation of the error correction effect specifically includes:

[0047] Compare the error correction coefficient between the sensor and the 3D model with a preset threshold.

[0048] If the error correction coefficient is greater than or equal to the preset threshold, the error correction effect between the corresponding sensor and the 3D model reaches the preset standard.

[0049] If the error correction coefficient is less than the preset threshold, the error correction effect between the corresponding sensor and the 3D model does not reach the preset standard.

[0050] As a further solution of the present invention: The process of obtaining the comprehensive performance coefficient is as follows:

[0051] Obtain the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient, and error correction coefficient in the process of intelligent park management, perform normalization processing on the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient, and error correction coefficient, and calculate the comprehensive performance coefficient.

[0052] As a further solution of the present invention: The overall operating performance of the evaluation system specifically includes:

[0053] Compare the comprehensive performance with a preset threshold.

[0054] If the comprehensive performance coefficient is greater than or equal to the preset threshold, it indicates that the overall operating performance of the intelligent park management system is qualified.

[0055] If the comprehensive performance coefficient is less than the preset threshold, it indicates that the overall operating performance of the intelligent park management system is unqualified.

[0056] Advantages of the present invention:

[0057] (1) Through the organic cooperation of multiple modules, the present invention realizes the precise management and efficient operation of an intelligent park in a dynamic environment, improving the adaptability, flexibility and data accuracy of the system. The adaptive sensor calibration module can intelligently sense the dynamic changes in the device position, the adjustment of the building structure and the changes in the environmental conditions, optimize the installation position and measurement parameters of the sensor in real time, dynamically calculate the sensor data accuracy coefficient, and ensure that the collected data is consistent with the actual environment; the multi-dimensional data fusion and three-dimensional modeling module comprehensively utilizes multi-source data such as lidar, UAV aerial photography and video surveillance to generate and update the high-precision three-dimensional model of the park in real time, and comprehensively evaluates the adaptability of the system to dynamic environmental changes by calculating the adaptation coefficient, ensuring the synchronization and reliability of the model and the real scene. Through the cooperation of each module, the present invention breaks through the limitation of insufficient management accuracy caused by sensor errors, data delays and environmental changes in the prior art, provides intelligent and highly stable technical support for the precise monitoring of park equipment, environmental status evaluation and resource optimization allocation, and at the same time lays a solid technical foundation for the sustainable development of the intelligent park.

[0058] (2) Through the deep integration of the intelligent image data registration module and the real-time error monitoring and feedback optimization module, the present invention improves the accuracy of the docking between the three-dimensional model of the park and the real-time image data and the error correction ability, and comprehensively optimizes the intelligent level of the system. The intelligent image data registration module adopts advanced convolutional neural network technology to extract high-dimensional features from the image data and the three-dimensional model, reduces the deviation in the spatial docking between the image and the model by dynamically optimizing the matching accuracy coefficient, and realizes the highly consistent three-dimensional scene visualization, providing a solid data foundation for device status perception and real-time scene rendering. The real-time error monitoring module, combined with the deviation information of the sensor measurement data and the three-dimensional model, dynamically calculates and optimizes the error correction coefficient, thereby efficiently correcting the sensor error and the modeling error to ensure the consistency of the data. Each module works in cooperation to improve the real-time performance and reliability of the park management system in a dynamic environment, provides intelligent technical support for the safety warning, equipment operation status monitoring, personnel behavior management and resource scheduling of the park, and can achieve stable operation in the changeable and complex park environment, laying the leading position of the intelligent park management system in technology. Brief Description of the Drawings

[0059] The present invention will be further described below with reference to the drawings.

[0060] Figure 1 It is a flow chart of an intelligent park management platform based on three-dimensional visualization according to the present invention. Detailed Embodiment

[0061] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0062] Please refer to Figure 1 as shown, the present invention is a three-dimensional visualization intelligent park management platform, including:

[0063] An adaptive sensor calibration module, which is used to automatically adjust the positions and measurement accuracies of sensors in the park according to the dynamic changes in the park environment, including equipment position changes and building reconstructions, ensure the matching degree between the image data collected by the sensors and the actual space of the park, calculate and dynamically update the sensor data accuracy coefficient for adjusting the measurement error of the sensors;

[0064] Among them, the sensors include: camera sensors, lidar sensors, infrared sensors, and GPS sensors;

[0065] A multi-dimensional data fusion and three-dimensional modeling module, which collects multi-source data of lidar, UAV aerial photography, and video surveillance through sensors, generates and real-time updates the three-dimensional model of the park, calculates and dynamically adjusts the adaptation coefficient for evaluating the adaptive adjustment ability of the system when the park changes;

[0066] An intelligent image data registration module, which automatically identifies the calibration area in the park, matches the image of the calibration area with the three-dimensional model in space, calculates and optimizes the matching accuracy coefficient for evaluating the accuracy of the docking between the image data and the three-dimensional model;

[0067] A real-time error monitoring and feedback optimization module, which calculates and optimizes the error correction coefficient based on the real-time spatial error between the sensor and the three-dimensional model for evaluating the error correction effect;

[0068] A comprehensive performance evaluation module, which evaluates the overall operation performance of the system according to the comprehensive calculation of the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient, and error correction coefficient, calculates the comprehensive performance coefficient for dynamically optimizing the system operation strategy and improving the accuracy and efficiency of park management.

[0069] In the adaptive sensor calibration module, the sensor data accuracy coefficient is calculated and dynamically updated for adjusting the measurement error of the sensors. The acquisition process of the sensor data accuracy coefficient is as follows:

[0070] Deploy a number of sensors within the park, and the position of each sensor is a three-dimensional coordinate ;

[0071] Among them, represents a sensor;

[0072] Use the sensor to collect the distance data of the preset reference points in the environment ;

[0073] Among them, represents a preset reference point;

[0074] Calculate the distance between the preset reference point and the sensor, and the calculation expression is: ;

[0075] In the formula, represents the calculated distance between the th sensor and the th preset reference point;

[0076] For each sensor and each preset reference point, calculate the measurement error of the th sensor and the th preset reference point, and the calculation expression is:

[0077] ;

[0078] In the formula, represents the measurement error of the th sensor and the th preset reference point, represents the actual measured distance between the th sensor and the th preset reference point;

[0079] Calculate the total error of the sensor for all reference points, and the calculation expression is: ;

[0080] In the formula, represents the total error of the th sensor, represents the total number of preset reference points;

[0081] For the th sensor, calculate the accuracy coefficient of the sensor data, and the calculation expression is: ;

[0082] In the formula, represents the accuracy coefficient of the th sensor data.

[0083] It should be noted that: The sensor data accuracy coefficient is used to quantify the measurement accuracy of sensors in the park and reflects the matching degree between the data collected by the sensors and the actual environment; in the process of smart park management, sensors are the basis for data collection, and accuracy directly determines the reliability of subsequent analysis and decision-making; by regularly calibrating and calculating the sensor data accuracy coefficient in real time, outliers or errors in the sensor data can be identified, and the measurement accuracy of the sensors can be dynamically adjusted, so as to provide high-quality data support for park equipment monitoring, environmental monitoring and personnel tracking.

[0084] In the multi-dimensional data fusion and 3D modeling module, multi-source data from lidar, UAV aerial photography and video surveillance are collected through sensors to generate and real-time update the 3D model of the park, calculate and dynamically adjust the adaptation coefficient, which is used to evaluate the adaptive adjustment ability of the system when the park changes;

[0085] Among them, the acquisition process of the adaptation coefficient is as follows:

[0086] The environmental data in the park are collected in real time through sensors, including sensor measurement error data and 3D model update data, to form a system state vector;

[0087] In the initial stage of the system, the particle set is initialized by uniform distribution, where each particle represents a system state, and the state vector includes sensor error and 3D model error;

[0088] At each time step, the state of the current particle set is predicted through the state transition model, and process noise is added for particle propagation to obtain the state estimate at the next moment;

[0089] According to the sensor data and 3D model data collected in real time, the matching degree between each particle and the actual observation data is calculated, and the weight of each particle is updated through the observation model. The update of the weight is based on the likelihood of sensor data and 3D model update error;

[0090] Resampling is performed according to the weights of the particles to generate a new particle set to ensure the stability of the particle filter algorithm;

[0091] Based on the particle set obtained by particle filtering, the adaptation coefficient of the system is calculated, and the calculation expression is:

[0092] ;

[0093] In the formula, represents the adaptation coefficient, represents the sensor, represents the total error of the th sensor, represents the 3D model update error of the

[0094] It should be noted that the adaptation coefficient is used to evaluate the response ability of the smart park system to dynamic environmental changes, such as building reconstruction, equipment position changes, or sensor failures; in smart park management, the system needs to dynamically adjust models and parameters according to real-time environmental changes, reflecting the flexibility and adaptability of the system under different change conditions; by monitoring the adaptation coefficient, the layout adjustment process and resource allocation efficiency within the park can be optimized to ensure that the management system can continuously operate efficiently.

[0095] In the intelligent image data registration module, by automatically identifying the calibration area within the park, the spatial matching of the image of the calibration area with the 3D model is performed, and the matching accuracy coefficient is calculated and optimized to evaluate the accuracy of the docking of image data with the 3D model;

[0096] Among them, the process of obtaining the matching accuracy coefficient specifically includes:

[0097] Real-time collect image data and 3D model data within the park through sensors, and automatically identify the calibration area within the park to form a preliminary matching point pair of image data and 3D model data;

[0098] Use a convolutional neural network to extract features from the image data, obtain the deep feature information in the image, and represent it as a feature vector for subsequent spatial docking optimization;

[0099] Process the 3D model data, use an adapted deep learning network to extract features from the 3D model, and obtain the feature representation of the 3D model for matching with the image features;

[0100] Match the image features and the 3D model features, use the minimization of the matching error to calculate the spatial transformation matrix between the image and the 3D model, including: rotation transformation and translation transformation, and optimize the docking accuracy of the image features and the 3D model features;

[0101] Based on the spatial docking result of the image and the 3D model, calculate the matching error between each matching feature point, and the calculation expression is: ;

[0102] In the formula, represents the matching error of the th matching point, represents the matching point, represents the th coordinate of the feature point in the image of the matching point, represents the th coordinate of the feature point in the 3D model of the matching point, represents the calculated transformation matrix;

[0103] According to the calculated matching error, calculate the matching accuracy coefficient, and the calculation expression is: ;

[0104] In the formula, represents the total number of matching points, represents the matching accuracy coefficient;

[0105] Evaluate the accuracy of the docking between the evaluation image data and the 3D model,

[0106] According to the matching accuracy coefficient, optimize the matching process, automatically adjust the parameters of the convolutional neural network, and optimize the spatial matching accuracy between the image and the 3D model;

[0107] Judge whether the matching accuracy coefficient is greater than or equal to the preset threshold. If so, the docking between the image data and the 3D model is accurate. If not, the docking between the image data and the 3D model is inaccurate.

[0108] It should be noted that: The matching accuracy coefficient is used to measure the accuracy of the docking between the image data and the 3D model, and is a key indicator for the visual management of smart parks; In park monitoring and analysis, the precise registration of the 3D model and real-time image data determines the accuracy of equipment status monitoring, anomaly detection, and planning analysis; By optimizing the matching error, the spatial deviation in the docking between the image data and the 3D model can be reduced, ensuring that managers can obtain accurate park scene displays and decision-making support.

[0109] In the real-time error monitoring and feedback optimization module, based on the real-time spatial error between the sensor and the 3D model, calculate and optimize the error correction coefficient for evaluating the error correction effect;

[0110] Among them, the process of obtaining the error correction coefficient is as follows:

[0111] Collect real-time measurement data in the park through sensors, including distance, angle, and speed eigenvalue;

[0112] Combine the actual reference values of the 3D model to form an input feature set and a target error value set for subsequent error correction modeling;

[0113] Calculate the mean value of the target error value as the initial prediction value, and the calculation expression is: ;

[0114] In the formula, represents the sample in the target error value set, represents the total number of samples in the target error value set, represents the initial prediction value, represents the th sample's actual error;

[0115] Calculate the The residual of a sample, the calculation expression is: ;

[0116] In the formula, represents the residual of the th sample;

[0117] Adopt a recursive iteration method to construct several weak learners to gradually fit the residual and update the current predicted value. The calculation expression is: ;

[0118] In the formula, represents the learning rate, represents the updated predicted value, represents the predicted value at the previous moment before update, represents the th moment, represents the th sample at the th moment of the regression decision tree;

[0119] Calculate the new residual according to the updated predicted value. The calculation expression is: ;

[0120] In the formula, represents the new residual of the th sample;

[0121] Based on the predicted error value and the actual error value obtained from the new residual, calculate the error correction coefficient. The calculation expression is: ;

[0122] In the formula, represents the error correction coefficient;

[0123] The evaluation of the error correction effect specifically includes:

[0124] Compare the error correction coefficient between the sensor and the 3D model with the preset threshold;

[0125] If the error correction coefficient is greater than or equal to the preset threshold, the error correction effect between the corresponding sensor and the 3D model reaches the preset standard;

[0126] If the error correction coefficient is less than the preset threshold, the error correction effect between the corresponding sensor and the 3D model does not reach the preset standard.

[0127] It should be noted that: the error correction coefficient is used to evaluate the error correction effect between the sensor and the 3D model data, reflecting the reliability and efficiency of the correction algorithm; in the intelligent park management, real-time error correction can eliminate data deviation caused by sensor errors, inaccurate modeling or environmental interference, thereby improving the overall accuracy of the system; by monitoring the error correction coefficient, the error correction model can be dynamically optimized to ensure a high degree of consistency between the park monitoring data and the actual situation, and guarantee safety warning, equipment maintenance and resource optimization.

[0128] In the comprehensive performance evaluation module, according to the comprehensive calculation of the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient and error correction coefficient, the overall operation performance of the system is evaluated, and the comprehensive performance coefficient is calculated, which is used to dynamically optimize the system operation strategy and improve the accuracy and efficiency of park management;

[0129] The process of obtaining the comprehensive performance coefficient is as follows:

[0130] Obtain the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient and error correction coefficient in the process of intelligent park management, perform normalization processing on the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient and error correction coefficient, calculate the comprehensive performance coefficient, and the calculation expression is: ;

[0131] In the formula, represents the comprehensive performance coefficient, represents the sensor data accuracy coefficient, represents the adaptation coefficient, represents the matching accuracy coefficient, represents the error correction coefficient, , , and represent preset proportional coefficients, and , , and are all greater than 0;

[0132] Compare the comprehensive performance with the preset threshold;

[0133] If the comprehensive performance coefficient is greater than or equal to the preset threshold, it indicates that the overall operation performance of the intelligent park management system is qualified;

[0134] If the comprehensive performance coefficient is less than the preset threshold, it indicates that the overall operation performance of the intelligent park management system is unqualified.

[0135] It should be noted that the comprehensive performance coefficient is used to quantify the overall operating performance of the intelligent park management system. It is obtained through normalization calculation by combining the sensor data accuracy coefficient, adaptation coefficient, matching accuracy coefficient, and error correction coefficient. During the park management process, the comprehensive performance coefficient is a key indicator for evaluating the system's efficiency and accuracy, reflecting the comprehensive performance of the system in a dynamic environment. By real-time monitoring the comprehensive performance coefficient, the collaborative effect of the system in data acquisition, model adaptation, spatial matching, and error correction can be evaluated, guiding system optimization and resource allocation to ensure that the park management system can operate efficiently, accurately, and intelligently in a changing environment.

[0136] The working principle of the present invention: Through the collaborative work of the adaptive sensor calibration module, multi-dimensional data fusion and three-dimensional modeling module, intelligent image data registration module, real-time error monitoring and feedback optimization module, and comprehensive performance evaluation module, precise monitoring and management of dynamic changes in the park are achieved. The adaptive sensor calibration module dynamically adjusts the position and measurement accuracy of the sensor according to environmental changes such as equipment position changes and building reconstructions, and calculates the sensor data accuracy coefficient to quantify the measurement accuracy. The multi-dimensional data fusion and three-dimensional modeling module collects multi-source data from lidar, drone aerial photography, and video surveillance, generates and real-time updates the three-dimensional model of the park, and calculates the adaptation coefficient to evaluate the system's response ability to environmental changes. The intelligent image data registration module extracts the features of the image and the three-dimensional model through a convolutional neural network, optimizes the spatial matching between the image and the three-dimensional model, and calculates the matching accuracy coefficient to measure the docking accuracy. The real-time error monitoring and feedback optimization module corrects the error between the sensor and the three-dimensional model, and calculates the error correction coefficient to evaluate the correction effect. The comprehensive performance evaluation module combines the normalization calculations of the above four coefficients to obtain the comprehensive performance coefficient, evaluates the overall operating performance of the system, and dynamically optimizes the operating strategy to ensure the accuracy and efficiency of park management, providing strong technical support for equipment monitoring, personnel management, environmental safety, and resource optimization.

[0137] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0138] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired or wireless (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more sets of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0139] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context before and after.

[0140] It should be understood that in various embodiments of the present application, the sequence numbers of the above processes do not indicate the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0141] The above has described in detail one embodiment of the present invention, but the content described is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. A three-dimensional visualization-based smart park management platform, characterized in that: include: Adaptive sensor calibration module, used to automatically adjust the location and measurement accuracy of sensors in the park according to the dynamic changes of the park environment, calculate and dynamically update the sensor data accuracy coefficient to adjust the sensor's measurement error; The process of obtaining the sensor data accuracy coefficient is as follows: Several sensors are deployed in the park, and the position of each sensor is a three-dimensional coordinate ;in, Represents a sensor; uses a sensor to collect distance data from preset reference points in the environment ;in, Indicates the preset reference point; calculate the distance between the preset reference point and the sensor, the calculation expression is: ; In the formula, Indicates The sensor and For each sensor and each preset reference point, calculate the distance The sensor and The measurement error of the preset reference point is calculated as follows: ; In the formula, Indicates The sensor and The measurement error of the preset reference point, Indicates The sensor and The actual measured distance of the preset reference points; calculate the total error of the sensor to all reference points, the calculation expression is: ; In the formula, Indicates The total error of the sensor is Indicates the total number of preset reference points; Sensors, calculate the accuracy coefficient of sensor data, the calculation expression is: ; In the formula, Indicates The sensor data accuracy coefficient; The multi-dimensional data fusion and 3D modeling module collects multi-source data from lidar, drone aerial photography and video surveillance through sensors, generates and updates the 3D model of the park in real time, calculates and dynamically adjusts the adaptation coefficient to evaluate the system's adaptive adjustment capabilities when the park changes; The intelligent image data registration module automatically identifies the calibration area in the park, spatially matches the image of the calibration area with the 3D model, calculates and optimizes the matching accuracy coefficient, and is used to evaluate the accuracy of the docking between the image data and the 3D model; Real-time error monitoring and feedback optimization module, which calculates and optimizes the error correction coefficient based on the real-time spatial error between the sensor and the 3D model to evaluate the error correction effect; The comprehensive performance evaluation module evaluates the overall operating performance of the system based on the comprehensive calculation of the sensor data accuracy coefficient, adaptability coefficient, matching precision coefficient and error correction coefficient, and calculates the comprehensive performance coefficient, which is used to dynamically optimize the system operation strategy and improve the accuracy and efficiency of park management.

2. According to claim 1, a three-dimensional visualization-based smart park management platform is characterized in that: The process of obtaining the adaptation coefficient is as follows: The environmental data in the park is collected in real time through sensors, including sensor measurement error data and 3D model update data, to form a system state vector; At the initial stage of the system, the particle set is initialized by uniform distribution, where each particle represents a system state, and the state vector includes sensor error and three-dimensional model error; At each time step, the state of the current particle set is predicted through the state transition model, and process noise is added to propagate the particles to obtain the state estimate at the next moment; According to the real-time collected sensor data and 3D model data, the matching degree between each particle and the actual observation data is calculated, and the weight of each particle is updated through the observation model. The weight update is based on the likelihood of the sensor data and the 3D model update error. Resample according to the particle weights to generate a new particle set to ensure the stability of the particle filter algorithm; Based on the particle set obtained by particle filtering, the adaptation coefficient of the system is calculated, and the calculation expression is: ; In the formula, represents the adaptation coefficient, Indicates the sensor, Indicates The total error of the sensor is Indicates The 3D model update error of each sensor.

3. According to claim 1, a three-dimensional visualization-based smart park management platform is characterized in that: The method of automatically identifying the calibration area in the park, spatially matching the image of the calibration area with the three-dimensional model, and calculating and optimizing the matching accuracy coefficient specifically includes: The sensor collects the image data and 3D model data in the park in real time, and automatically identifies the calibration area in the park to form preliminary matching point pairs of image data and 3D model data; Convolutional neural networks are used to extract features from image data, obtain deep feature information in the image, and represent it as a feature vector for subsequent spatial docking optimization; Process the 3D model data, use an adapted deep learning network to extract features from the 3D model, and obtain a feature representation of the 3D model for matching with image features; Match the image features with the 3D model features, and calculate the spatial transformation matrix between the image and the 3D model by minimizing the matching error, including rotation transformation and translation transformation, and optimize the docking accuracy between the image features and the 3D model features; Based on the spatial docking results of the image and the 3D model, the matching error between each matching feature point is calculated. The calculation expression is: ; In the formula, Indicates The matching error of the matching points is represents the matching point, Indicates The coordinates of the feature points in the matching point image, Indicates The coordinates of the feature points in the 3D model of the matching points, Represents the calculated transformation matrix; according to the calculated matching error, the matching accuracy coefficient is calculated, and the calculation expression is: ; In the formula, represents the total number of matching points, Indicates the matching accuracy coefficient.

4. The three-dimensional visualization-based smart park management platform according to claim 1 is characterized in that: The accuracy of the docking between the evaluation image data and the three-dimensional model, According to the matching accuracy coefficient, the matching process is optimized, the parameters of the convolutional neural network are automatically adjusted, and the spatial matching accuracy between the image and the 3D model is optimized; It is determined whether the matching accuracy coefficient is greater than or equal to a preset threshold. If so, the image data is accurately matched with the three-dimensional model. If not, the image data is not accurately matched with the three-dimensional model.

5. The three-dimensional visualization-based smart park management platform according to claim 1 is characterized in that: The calculating and optimizing of the error correction coefficient based on the real-time spatial error between the sensor and the three-dimensional model specifically includes: Collect real-time measurement data in the park through sensors, including distance, angle and speed characteristic values; Combined with the actual reference value of the 3D model, an input feature set and a target error value set are formed for subsequent error correction modeling; Calculate the mean of the target error values ​​as the initial prediction value. The calculation expression is: ; In the formula, represents the samples in the target error value set, represents the total number of samples in the target error value set, represents the initial prediction value, Indicates The actual error of the samples; Calculate the The residual of samples is calculated as: ; In the formula, Indicates The residual of the sample; Using a recursive iterative method, several weak learners are constructed to gradually fit the residuals and update the current prediction value. The calculation expression is: ; In the formula, represents the learning rate, represents the updated predicted value, Indicates the predicted value at the last moment before the update. Indicates a moment, Indicates Sample No. Regression decision tree at time instant; Calculate the new residual based on the updated prediction value. The calculation expression is: ; In the formula, Indicates The new residuals of samples; Based on the predicted error value and actual error value obtained by the new residual, the error correction coefficient is calculated. The calculation expression is: ; In the formula, Represents the error correction factor.

6. The three-dimensional visualization-based smart park management platform according to claim 1 is characterized in that: The evaluation of the error correction effect specifically includes: comparing an error correction coefficient between the sensor and the three-dimensional model with a preset threshold; If the error correction coefficient is greater than or equal to the preset threshold, the error correction effect between the corresponding sensor and the three-dimensional model reaches the preset standard; If the error correction coefficient is less than the preset threshold, the error correction effect between the corresponding sensor and the three-dimensional model does not meet the preset standard.

7. The three-dimensional visualization-based smart park management platform according to claim 1 is characterized in that: The process of obtaining the comprehensive performance coefficient is as follows: The accuracy coefficient, adaptability coefficient, matching accuracy coefficient and error correction coefficient of sensor data in the process of smart park management are obtained, the accuracy coefficient, adaptability coefficient, matching accuracy coefficient and error correction coefficient of the sensor data are normalized, and the comprehensive performance coefficient is calculated.

8. The three-dimensional visualization-based smart park management platform according to claim 1 is characterized in that: The overall operational performance of the evaluation system specifically includes: Compare the comprehensive performance with the preset threshold; if the comprehensive performance coefficient is greater than or equal to the preset threshold, it means that the overall operating performance of the smart park management system is qualified; if the comprehensive performance coefficient is less than the preset threshold, it means that the overall operating performance of the smart park management system is unqualified.

Images