Intelligent parking management method and system

Abstract

The invention relates to the field of parking lot management, in particular to an intelligent parking management method and system. The method comprises the following steps: collecting a multi-angle monitoring image of an entering vehicle; performing image multi-view registration reconstruction on the multi-angle monitoring image of the entering vehicle, performing vehicle three-dimensional point cloud modeling, and constructing a three-dimensional morphological structure model of the entering vehicle; acquiring real-time state information of the three-dimensional parking lot; idle parking space recognition and idle lifting platform space distribution analysis are carried out on the real-time state information of the three-dimensional parking lot, and a three-dimensional parking lot twin model is constructed; performing optimal disable matching on the three-dimensional parking lot twin model according to the entering vehicle three-dimensional morphological structure model to obtain an optimal disable matching parking space; a real-time parking management decision is made, and a real-time parking management strategy is constructed; and obtaining a parking lot historical operation log. According to the invention, parking management with high efficiency and high utilization rate is realized.

Description

Technical Field

[0001] The present invention relates to the field of parking lot management, and in particular to a smart parking management method and system. Background Art

[0002] With the acceleration of urbanization and the continuous increase in the number of motor vehicles, traditional flat parking lots can no longer meet the growing parking demand. As an efficient and space-efficient parking solution, multi-story parking lots have been widely used in urban commercial areas, residential areas, transportation hubs and other places. The main advantage of multi-story parking lots is that they can make full use of vertical space and achieve rapid storage and retrieval of vehicles through mechanization, automation or intelligent means, greatly improving the parking capacity and utilization efficiency of parking lots.

[0003] However, with the continuous expansion of the scale of multi-story parking lots and the increase in the frequency of use, how to improve the level of intelligent parking management and solve the problems existing in traditional parking management methods has become a technical problem that needs to be solved urgently. Traditional multi-story parking lot management methods usually rely on manual operation or simple mechanized management systems, which have problems such as difficulty in obtaining parking information in real time, cumbersome parking procedures, and unreasonable allocation of parking resources. This not only affects the efficiency of parking lot use, but also leads to waste of parking resources and even brings hidden dangers in vehicle management and safety.

[0004] In addition, in the long-term operation of multi-story parking lots, due to factors such as large traffic volume and a wide variety of vehicles, they are also prone to information asymmetry, low parking space vacancy rate, and unsmooth parking process, which affects the experience of both car owners and parking lot managers. Therefore, a smart parking management method based on modern information technology and intelligent means is urgently needed. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a smart parking management method and system to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a smart parking management method, comprising the following steps: Step S1: collecting multi-angle monitoring images of vehicles entering the venue; performing multi-view registration and reconstruction on the multi-angle monitoring images of vehicles entering the venue, and performing three-dimensional point cloud modeling of the vehicles to construct a three-dimensional morphological structure model of the vehicles entering the venue; Step S2: obtaining real-time status information of the three-dimensional parking lot; identifying vacant parking spaces and analyzing the spatial distribution of vacant lifting platforms based on the real-time status information of the three-dimensional parking lot, and constructing a twin model of the three-dimensional parking lot; Step S3: optimally matching the parking lot twin model according to the three-dimensional morphological structure model of the incoming vehicle to obtain the optimal parking space; and making real-time parking management decisions to build a real-time parking management strategy; Step S4: Obtain the parking lot's historical operation log; analyze the owner's historical vehicle use behavior in the parking lot's historical operation log, and then perform dynamic vehicle use time point prediction, thereby generating a vehicle owner's vehicle use prediction time point; Step S5: Based on the predicted time point of the car owner's car use, the twin model of the three-dimensional parking lot is analyzed for the optimal vehicle scheduling time and the advance vehicle scheduling decision, and the advance vehicle scheduling strategy is constructed; Step S6: Conduct intelligent collaborative parking vehicle transfer learning on the real-time parking management strategy and the advance vehicle scheduling strategy, and perform iterative reinforcement optimization to build a smart parking management model for the three-dimensional parking lot.

[0007] The present invention establishes a three-dimensional point cloud model of the vehicle through multi-angle monitoring images, accurately reconstructs the three-dimensional morphological structure of the incoming vehicle, and constructs a three-dimensional morphological structure model to help accurately identify the type, size and shape of the vehicle, provide basic data for subsequent parking space matching, analyze real-time status information and identify idle parking spaces and analyze the spatial distribution of idle lifting platforms, and can construct a twin model of the three-dimensional parking lot. The twin model reflects the real-time status of the parking lot, helps to accurately match available parking spaces and optimize parking management decisions, performs optimal available parking matching based on the three-dimensional model of the incoming vehicle, determines the most suitable parking space, optimizes the parking layout, and constructs a real-time parking management strategy, which can improve the utilization and efficiency of the parking lot, reduce parking waiting time, and improve Improve user experience, analyze historical operation logs and predict the time when car owners use their cars, which helps to optimize vehicle scheduling and parking space allocation. Dynamic car use time prediction improves the operating efficiency of parking lots, reduces vacancy rate and chaos, and makes optimal vehicle scheduling time analysis and advance vehicle scheduling decisions based on the predicted time when car owners use their cars, which helps to arrange vehicle parking and retrieval in advance. Advance vehicle scheduling strategies reduce parking lot congestion, improve parking efficiency, and enhance user satisfaction. Through intelligent collaborative parking vehicle transfer learning and iterative reinforcement optimization, a smart parking management model is built to realize the intelligent allocation and management of parking resources. Model optimization continuously improves the intelligence level of parking lot management and realizes efficient, intelligent and sustainable development of parking lot operations.

[0008] Preferably, step S1 comprises the following steps: Step S11: When a vehicle is detected entering the venue, a multi-angle monitoring image of the vehicle entering the venue is collected by a camera; Step S12: performing image multi-view registration and reconstruction on the multi-angle monitoring image of the incoming vehicle to construct a panoramic reconstructed vehicle monitoring image; Step S13: performing visual recognition of vehicle structure key points on the panoramic reconstructed vehicle monitoring image, and marking a plurality of vehicle structure key points; Step S14: performing vehicle geometry analysis on a plurality of key points of the vehicle structure to generate geometric features of the incoming vehicle; Step S15: performing vehicle size type recognition on the geometric features of the incoming vehicle to obtain the size type features of the incoming vehicle; Step S16: Perform vehicle three-dimensional point cloud modeling based on the size and type characteristics of the incoming vehicle to construct a three-dimensional morphological structure model of the incoming vehicle.

[0009] The present invention collects multi-angle monitoring images of entering vehicles through cameras, realizes real-time monitoring and recording of vehicle entry, provides a data basis for subsequent vehicle identification and management, performs multi-view registration and reconstruction of images, generates panoramic reconstructed vehicle monitoring images, provides complete vehicle information under a comprehensive perspective, and is conducive to accurate vehicle analysis and identification. The geometric morphology analysis of key points of the vehicle structure is performed to generate geometric morphology features of the entering vehicles, which is conducive to an in-depth understanding of the morphological features and structure of the vehicles. By performing size type recognition on the geometric morphology features of the entering vehicles, the type features of the entering vehicles are obtained, which provides a basis for subsequent vehicle classification and management. Three-dimensional point cloud modeling is performed based on the size type features of the entering vehicles, and a three-dimensional morphological structure model of the entering vehicles is constructed to achieve accurate modeling and reconstruction of the vehicle morphology.

[0010] Preferably, the specific steps of step S12 are: Performing image brightness enhancement on the multi-angle monitoring image of the vehicle entering the venue to obtain a brightness enhanced monitoring image; Identify abnormal image noise points on the brightness-enhanced monitoring image and mark multiple abnormal image noise points; Perform non-local mean filtering on multiple image abnormal noise points to obtain a filtered optimized monitoring image; Perform vehicle multi-angle edge feature analysis on the filtered optimized monitoring image and extract vehicle multi-angle edge feature data; Perform detail sharpening and enhancement processing based on multi-angle edge feature data of the vehicle, thereby obtaining multi-angle sharp detail features; The filtered optimized monitoring image is reconstructed by multi-view registration based on the multi-angle sharpening detail features to construct a panoramic reconstructed vehicle monitoring image.

[0011] The present invention improves the visual quality and clarity of the image by enhancing the image brightness, thereby improving the effect of subsequent processing steps. Brightness enhancement helps to highlight the details and features in the image, making the monitoring image more readable and recognizable. Identifying abnormal noise points in the image helps to remove interference and reduce errors in subsequent analysis. Marking abnormal noise points helps to further process and repair these problem areas, thereby improving the accuracy and reliability of the image. Non-local mean filtering is performed on abnormal noise points to effectively reduce noise and retain image details. Filter optimization helps to improve image quality, thereby making subsequent analysis more accurate and reliable. Multi-angle edge features of the vehicle are analyzed to extract the shape and contour information of the vehicle, thereby helping to identify the vehicle. Extracting edge features helps to perform subsequent vehicle classification, tracking and identification. Detail sharpening and enhancement of edge features highlights the details and contours of the image, thereby enhancing the clarity of the image. Detail enhancement helps to more clearly display the features of the vehicle, thereby improving the accuracy of identification and analysis. A panoramic vehicle monitoring image is obtained through multi-view registration and reconstruction, thereby providing a more comprehensive perspective and information. Constructing a panoramic image helps to monitor and manage the overall situation of a parking lot, thereby improving the management efficiency and safety of the parking lot.

[0012] Preferably, the specific steps of step S2 are: Step S21: Acquire real-time status information of the three-dimensional parking lot; Step S22: Identify the vacant parking spaces based on the real-time status information of the multi-story parking lot, and extract the vacant parking space status information; Step S23: positioning the idle lifting platform of the three-dimensional parking lot based on the idle parking space status information, and extracting the positioning data of the idle lifting platform; Step S24: performing spatial distribution analysis on the idle lifting platform positioning data to generate idle lifting platform spatial distribution data; Step S25: Render the spatial distribution data of idle lifting platforms in real time according to the parking restriction standards to construct a twin model of the three-dimensional parking lot.

[0013] The present invention provides real-time parking lot data by acquiring parking lot status information in real time, thereby helping managers understand the current situation of the parking lot. The real-time status information provides basic data support for subsequent intelligent management decisions. Identifying idle parking spaces and extracting status information helps to accurately understand the usage of the parking lot and guide vehicle guidance and parking management. Idle parking space information helps to optimize the utilization of parking lot resources and improve the utilization efficiency of the parking lot. Positioning the lifting platform based on the idle parking space information effectively guides vehicles to enter the idle position and improves parking efficiency. The positioning data of the idle lifting platform helps to achieve accurate management and scheduling of parking spaces. Analysis of the spatial distribution of idle lifting platforms helps to optimize the design and layout of parking lots and improve the capacity and efficiency of parking lots. Generating spatial distribution data helps to understand the distribution of lifting platforms and provides a reference for the planning and management of parking spaces. Rendering the parking limit standards based on the spatial distribution data of idle lifting platforms realizes the intelligent allocation and management of parking spaces. A twin model is constructed to simulate and monitor the overall situation of the parking lot, thereby providing support and reference for parking lot management decisions.

[0014] Preferably, step S25 specifically comprises the following steps: Conducting a three-dimensional topological structure analysis on the three-dimensional parking lot to generate the three-dimensional topological structure features of the parking lot; Based on the three-dimensional topological structure characteristics of the parking lot, the spatial distribution data of the idle lifting platforms are fitted to construct a three-dimensional distribution network of idle platforms. Perform deactivation type analysis on the idle parking space status information to generate the deactivation vehicle type for each idle parking space; Analyze the parking limit standard of each idle parking space based on the type of parked vehicles, and extract the parking limit standard of each idle parking space; Based on the size limit parking standards of each vacant parking space, the vacant parking space attributes of the 3D distribution network of vacant platforms are rendered in real time to construct a twin model of the 3D parking lot.

[0015] The present invention can reveal the layout and spatial relationship of parking spaces by analyzing the three-dimensional topological structure of the parking lot, and provide basic information for parking management. The generation of three-dimensional topological structure features is helpful to understand the overall structure of the parking lot, and provides a basis for subsequent analysis and optimization. The spatial distribution data of the idle lifting platform is fitted based on the three-dimensional topological structure features to better understand the space utilization and optimize the allocation of parking spaces. The construction of a three-dimensional distribution network of idle platforms is helpful to realize the intelligent management and scheduling of parking spaces. The analysis of the parking type of the idle parking space status information helps to distinguish the parking needs of different vehicle types, and manages them in a targeted manner. The generation of the type of idle vehicles for each idle parking space is helpful to arrange the use of parking spaces more effectively and improve the utilization rate of the parking lot. The vehicle size limit parking standard analysis of the type of idle vehicles in the idle parking space ensures the safety and convenience of vehicle parking. The extraction of the size limit parking standard of each idle parking space helps to avoid size conflicts when vehicles are parked, and improves the management efficiency of the parking lot. The idle platform three-dimensional distribution network is rendered in real time based on the size limit parking standard to realize the intelligent allocation and management of parking spaces. The construction of a twin model can realize real-time monitoring and simulation of the parking lot status, and provide data support for management decisions.

[0016] Preferably, step S3 specifically comprises the following steps: Step S31: performing a parking simulation of an idle parking space on the twin model of the three-dimensional parking lot according to the three-dimensional morphological structure model of the incoming vehicle to generate idle parking space parking simulation data; Step S32: performing optimal deactivatable matching on the idle parking space parking simulation data to obtain the optimal deactivatable matching parking space; Step S33: performing lifting and lowering scheduling of the three-dimensional parking space for the optimal deactivatable matching parking space, and generating lifting and lowering scheduling data of the three-dimensional parking space; Step S34: performing parking path planning based on the parking space scheduling data to generate a parking planning path; Step S35: Make real-time parking management decisions based on the parking planning path and the three-dimensional parking space lifting and dispatching data, and build a real-time parking management strategy.

[0017] The present invention simulates the parking situation of vehicles in the parking lot by performing a parking simulation on the three-dimensional morphological structure model of the incoming vehicles, optimizes the utilization of parking spaces, and generates parking simulation data of idle parking spaces, which is helpful for understanding the real-time status of parking spaces, and provides a basis for subsequent vehicle scheduling and management. The present invention performs optimal deactivatable matching on the parking simulation data of idle parking spaces to achieve the best matching between vehicles and parking spaces, and improves the utilization rate of parking spaces. Obtaining the optimal deactivatable matching parking spaces helps optimize the allocation of vehicles in the parking lot, improves the efficiency of vehicle entry and exit, and performs three-dimensional parking lifting and lowering scheduling on the optimal deactivatable matching parking spaces to achieve intelligent scheduling of parking spaces, and improves The capacity and efficiency of parking lots, generating data on the lifting and dispatching of multi-story parking spaces helps optimize the use of parking spaces, reduces vehicle congestion and waiting time, and conducts parking route planning based on parking dispatching data to guide vehicles to find the best parking route, reducing parking time and congestion. Generating parking planning routes helps optimize parking processes and improve the operating efficiency of parking lots. Real-time parking management decisions are made based on parking planning routes and lifting and dispatching data of multi-story parking spaces to achieve real-time monitoring and management of parking lots. Building real-time parking management strategies helps adjust parking space allocation and vehicle dispatch according to real-time conditions, providing more intelligent parking services.

[0018] Preferably, the specific steps of step S4 are: Step S41: Obtain parking lot historical operation log; Step S42: Analyze the historical car owner's car-taking behavior in the parking lot's historical operation log to generate dynamic car owner behavior features; Step S43: quantify the vehicle usage time based on the dynamic vehicle owner behavior characteristics to obtain the vehicle owner's multiple vehicle usage timestamps; Step S44: mining personalized car-use behavior features based on the car owner's multiple car-use timestamps to generate the car owner's car-use behavior rules; Step S45: Dynamically predict the time point of vehicle use based on the vehicle owner's vehicle use behavior pattern, thereby generating a predicted time point of vehicle use by the vehicle owner.

[0019] The present invention includes the past operation data of the parking lot through the historical operation log, providing valuable information for analyzing the operation status and trend of the parking lot. The purpose of obtaining the historical operation log is to conduct subsequent owner behavior analysis and prediction based on historical data, help optimize the management strategy of the parking lot, and analyze the owner's historical car-taking behavior in the historical operation log to understand the owner's behavior habits and preferences, provide a basis for personalized services, and generate dynamic owner behavior characteristics to help classify and understand the owner's behavior, and provide support for subsequent predictions and decisions. By quantifying the use time of the dynamic owner behavior characteristics, the owner's car use frequency and time distribution are analyzed to reveal the use of the car Obtaining the timestamps of car owners' multiple car usage helps determine the regular time points of car owners' car usage, and provides basic data for predicting the time points of car owners' car usage. Personalized mining of the timestamps of car owners' multiple car usage can discover the regularities and characteristics of different car owners' car usage behaviors, and provide support for personalized services. Generating the regularities of car owners' car usage behaviors helps understand the behavioral habits of car owners, and provides guidance for improving the quality of parking lot services and user experience. Dynamic car usage time point prediction based on the regularities of car owners' car usage behaviors predicts car owners' future car usage needs and helps optimize parking lot resource allocation. Generating predicted time points for car owners' car usage helps prepare and schedule parking spaces in advance, and improves the utilization rate and management efficiency of parking lots.

[0020] Preferably, the specific steps of step S5 are: Step S51: Calculate the current vehicle position on the three-dimensional parking lot twin model to obtain the parking lot position of the current vehicle; Step S52: performing a planar platform dispatching path analysis according to the current parking lot position of the vehicle to generate a current vehicle platform dispatching path; Step S53: Calculate the path scheduling time for the current vehicle platform scheduling path, so as to obtain the current vehicle path scheduling time; Step S54: performing an optimal scheduling time analysis on the current vehicle path scheduling time based on the predicted time point of the vehicle owner's vehicle use, and generating an optimal vehicle scheduling time; Step S55: Make a pre-vehicle scheduling decision for the optimal vehicle scheduling time and construct a pre-vehicle scheduling strategy.

[0021] The present invention calculates the current vehicle position through a twin model, obtains the specific position information of the vehicle in the parking lot in real time, and helps to perform vehicle scheduling and management. Obtaining the parking lot position of the current vehicle is helpful to realize the accurate positioning and path planning of the vehicle in the parking lot, and improves parking efficiency. According to the parking lot position of the current vehicle, a plane platform scheduling path analysis is performed to determine the optimal platform scheduling path, reduce vehicle driving time and congestion, and generate the current vehicle platform scheduling path to help optimize the parking position of the vehicle and improve the operation efficiency inside the parking lot. The path scheduling time calculation is performed on the current vehicle platform scheduling path to estimate the time required for the vehicle to reach the destination, and help to make scheduling preparations in advance. Analyzing the optimal vehicle scheduling time is helpful to determine the best vehicle scheduling opportunity, avoid congestion and conflict, and improve parking efficiency. Combined with the predicted time point of the car owner's car use, the optimal scheduling time analysis is performed on the current vehicle path scheduling time to make more accurate scheduling arrangements according to the car owner's needs. Generating the optimal vehicle scheduling time is helpful to meet the needs of the car owner to the greatest extent within the predicted time point, improve the service quality of the parking lot, and make a pre-vehicle scheduling decision on the optimal vehicle scheduling time to arrange the parking position and time of the vehicle in advance, reduce waiting time and congestion, and construct a pre-vehicle scheduling strategy to help optimize the resource utilization of the parking lot and improve the management efficiency and user experience of the parking lot.

[0022] Preferably, the specific steps of step S6 are: Step S61: performing intelligent collaborative parking vehicle transfer learning on the real-time parking management strategy and the pre-use vehicle scheduling strategy to generate an intelligent collaborative scheduling strategy; Step S62: Perform dynamic management efficiency analysis on the intelligent collaborative scheduling strategy to generate collaborative scheduling strategy management efficiency; Step S63: Iteratively strengthen and optimize the intelligent collaborative scheduling strategy according to the management efficiency of the collaborative scheduling strategy to construct a smart parking management model for the multi-story parking lot.

[0023] The present invention realizes intelligent coordination of vehicle parking and scheduling through intelligent collaborative parking vehicle transfer learning, combined with real-time parking management strategy and pre-vehicle scheduling strategy. Generating intelligent collaborative scheduling strategy is helpful to improve resource utilization and management efficiency of parking lot, reduce parking waiting time and congestion, conduct dynamic management efficiency analysis on intelligent collaborative scheduling strategy, evaluate the real-time effect of scheduling strategy, discover and solve management efficiency problems, and generate collaborative scheduling strategy management efficiency to help optimize scheduling strategy, improve parking lot operation efficiency and service quality to cope with changing needs. Iterate, strengthen and optimize intelligent collaborative scheduling strategy according to its management efficiency, continuously improve scheduling strategy, adapt to different situations and needs, and construct a smart parking management model for three-dimensional parking lot, which is helpful to integrate various strategies and optimization schemes, realize intelligent and automated parking lot management, and improve management efficiency and user experience.

[0024] In this specification, a smart parking management system is provided, which is used to execute the smart parking management method as described above, including: An image registration and reconstruction module is used to collect multi-angle monitoring images of vehicles entering the venue; perform image multi-view registration and reconstruction on the multi-angle monitoring images of vehicles entering the venue, perform vehicle three-dimensional point cloud modeling, and construct a three-dimensional morphological structure model of vehicles entering the venue; The twin model module is used to obtain the real-time status information of the three-dimensional parking lot; identify the vacant parking spaces based on the real-time status information of the three-dimensional parking lot, analyze the spatial distribution of the idle lifting platforms, and build a twin model of the three-dimensional parking lot; The deactivatable matching module is used to perform optimal deactivatable matching on the twin model of the three-dimensional parking lot according to the three-dimensional morphological structure model of the incoming vehicle, obtain the optimal deactivatable matching parking space, and make real-time parking management decisions to build a real-time parking management strategy; The vehicle use time prediction module is used to obtain the parking lot's historical operation logs; analyze the owner's historical vehicle use behavior in the parking lot's historical operation logs, and then predict the dynamic vehicle use time points, thereby generating the owner's vehicle use prediction time points; The pre-use vehicle scheduling module is used to analyze the optimal vehicle scheduling time and pre-use vehicle scheduling decisions for the three-dimensional parking lot twin model based on the predicted time point of the car owner's car use, and to build a pre-use vehicle scheduling strategy; The smart parking management module is used to conduct intelligent collaborative parking vehicle transfer learning for real-time parking management strategies and pre-use vehicle scheduling strategies, and to perform iterative reinforcement optimization to build a smart parking management model for three-dimensional parking lots.

[0025] The present invention helps to accurately capture the three-dimensional morphological information of the vehicle by performing multi-angle image registration and reconstruction on the multi-angle monitoring images of the incoming vehicles, thereby realizing accurate modeling of the vehicle. The module uses image processing technology to reconstruct the monitoring images from multiple angles into a three-dimensional point cloud model, providing high-precision vehicle morphological structure data for subsequent intelligent parking management, and obtaining real-time status information of the three-dimensional parking lot, including vacant parking space identification and spatial distribution analysis of vacant lifting platforms, to construct a twin model of the three-dimensional parking lot. The module provides accurate control of the parking lot status by combining real-time data and model prediction, and provides real-time support for subsequent parking space scheduling and management. It performs optimal deactivatable matching based on the vehicle three-dimensional morphological structure model and the parking lot twin model, helping to achieve optimal parking space allocation and real-time parking management decisions. The module provides the best parking position for the vehicle through an intelligent matching algorithm, thereby improving Parking lot utilization and management efficiency, through historical operation logs and car owner's car usage behavior analysis, to achieve accurate prediction of the time point of car owner's car use, and then generate the predicted time point of car owner's car use. This module uses data analysis and prediction algorithms to help parking lot managers better understand car owners' car usage habits, and provide important reference for parking lot resource scheduling and management. According to the predicted time point of car owners' car use, the optimal vehicle scheduling and pre-vehicle scheduling decisions are made, and an efficient pre-vehicle scheduling strategy is constructed. This module optimizes vehicle scheduling plans and improves parking lot resource utilization efficiency by analyzing the predicted car use time point and parking lot status. It conducts intelligent collaborative learning and iterative optimization of parking management strategies and pre-vehicle scheduling strategies to construct a smart parking management model. This module integrates the data and decisions of each module, and continuously improves parking lot management strategies through machine learning and optimization algorithms, thereby improving parking lot operation efficiency and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A schematic diagram of the steps of a smart parking management method of the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0027] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0028] The present application example provides a smart parking management method and system. The execution subjects of the smart parking management method and system include but are not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0029] See also Figures 1 to 4 The present invention provides a smart parking management method, which includes the following steps: Step S1: collecting multi-angle monitoring images of vehicles entering the venue; performing multi-view registration and reconstruction on the multi-angle monitoring images of vehicles entering the venue, and performing three-dimensional point cloud modeling of the vehicles to construct a three-dimensional morphological structure model of the vehicles entering the venue; Step S2: obtaining real-time status information of the three-dimensional parking lot; identifying vacant parking spaces and analyzing the spatial distribution of vacant lifting platforms based on the real-time status information of the three-dimensional parking lot, and constructing a twin model of the three-dimensional parking lot; Step S3: optimally matching the parking lot twin model according to the three-dimensional morphological structure model of the incoming vehicle to obtain the optimal parking space; and making real-time parking management decisions to build a real-time parking management strategy; Step S4: Obtain the parking lot's historical operation log; analyze the owner's historical vehicle use behavior in the parking lot's historical operation log, and then perform dynamic vehicle use time point prediction, thereby generating a vehicle owner's vehicle use prediction time point; Step S5: Based on the predicted time point of the car owner's car use, the twin model of the three-dimensional parking lot is analyzed for the optimal vehicle scheduling time and the advance vehicle scheduling decision, and the advance vehicle scheduling strategy is constructed; Step S6: Conduct intelligent collaborative parking vehicle transfer learning on the real-time parking management strategy and the advance vehicle scheduling strategy, and perform iterative reinforcement optimization to build a smart parking management model for the three-dimensional parking lot.

[0030] The present invention establishes a three-dimensional point cloud model of the vehicle through multi-angle monitoring images, accurately reconstructs the three-dimensional morphological structure of the incoming vehicle, and constructs a three-dimensional morphological structure model to help accurately identify the type, size and shape of the vehicle, provide basic data for subsequent parking space matching, analyze real-time status information and identify idle parking spaces and analyze the spatial distribution of idle lifting platforms, and can construct a twin model of the three-dimensional parking lot. The twin model reflects the real-time status of the parking lot, helps to accurately match available parking spaces and optimize parking management decisions, performs optimal available parking matching based on the three-dimensional model of the incoming vehicle, determines the most suitable parking space, optimizes the parking layout, and constructs a real-time parking management strategy, which can improve the utilization and efficiency of the parking lot, reduce parking waiting time, and improve Improve user experience, analyze historical operation logs and predict the time when car owners use their cars, which helps to optimize vehicle scheduling and parking space allocation. Dynamic car use time prediction improves the operating efficiency of parking lots, reduces vacancy rate and chaos, and makes optimal vehicle scheduling time analysis and advance vehicle scheduling decisions based on the predicted time when car owners use their cars, which helps to arrange vehicle parking and retrieval in advance. Advance vehicle scheduling strategies reduce parking lot congestion, improve parking efficiency, and enhance user satisfaction. Through intelligent collaborative parking vehicle transfer learning and iterative reinforcement optimization, a smart parking management model is built to realize the intelligent allocation and management of parking resources. Model optimization continuously improves the intelligence level of parking lot management and realizes efficient, intelligent and sustainable development of parking lot operations.

[0031] In the embodiment of the present invention, refer to Figure 1 , is a flowchart of a smart parking management method of the present invention. In this example, the steps of the smart parking management method include: Step S1: collecting multi-angle monitoring images of vehicles entering the venue; performing multi-view registration and reconstruction on the multi-angle monitoring images of vehicles entering the venue, and performing three-dimensional point cloud modeling of the vehicles to construct a three-dimensional morphological structure model of the vehicles entering the venue; In this embodiment, appropriate monitoring equipment is selected, such as a high-definition camera or a depth camera (such as structured light or lidar), to ensure that images of vehicles entering the parking lot can be captured from multiple angles. Multiple cameras are installed at the entrance of the parking lot, and reasonable angles and heights are set to cover different perspectives of vehicles entering the parking lot. When a vehicle enters the parking lot, the system automatically triggers the camera to collect images, ensuring that each vehicle is photographed at multiple time points (such as the front, side, and rear) to obtain sufficient viewing angle information, record the timestamp and camera position information of each image for subsequent registration, and pre-process the collected images, including denoising and enhancement. Contrast and image cropping to improve the accuracy of subsequent processing, use image processing technology (such as OpenCV) to convert image formats to ensure image consistency, use feature extraction algorithms (such as SIFT, SURF or ORB) to extract key feature points from multi-angle images, these feature points will be used for matching in subsequent registration, use feature matching algorithms (such as FLANN or brute force matching) to match features from different perspectives, find the correspondence between the same feature points in different images, use camera calibration technology (such as Zhang Zhengyou calibration method) and geometric transformation (such as homography matrix or basic matrix) to match the features. The features are registered, and the transformation relationship between each image is calculated. The outliers are removed by the RANSAC (random sampling consistency) algorithm to enhance the robustness of the matching results. According to the registration results, multiple images are fused to generate a unified panoramic image or a preliminary model of 3D reconstruction. If a depth camera is used, the depth image is directly acquired to generate point cloud data. If an ordinary camera is used, the depth information is calculated by stereo vision technology to generate a depth map. The pixel information in the multi-view image is combined with the depth information to generate a 3D point cloud. The coordinates of each point are determined by its position in the 3D space (X, Y, Z). The generated point cloud is processed, including denoising, downsampling and feature extraction. It is processed using tools such as PCL (Point Cloud Library) to ensure the quality and accuracy of the point cloud. Based on the processed point cloud data, a 3D reconstruction algorithm (such as Poisson Surface Reconstruction or Ball Pivoting) is used to generate a 3D morphological structure model of the incoming vehicle. The generated 3D model is visualized to ensure the details and accuracy of the model. The constructed 3D morphological structure model is saved in a standard 3D model format (such as OBJ or PLY) for subsequent use and analysis.

[0032] Step S2: obtaining real-time status information of the three-dimensional parking lot; identifying vacant parking spaces and analyzing the spatial distribution of vacant lifting platforms based on the real-time status information of the three-dimensional parking lot, and constructing a twin model of the three-dimensional parking lot; In this embodiment, sensors and cameras are installed in the three-dimensional parking lot to ensure that all parking spaces and lifting platforms are covered. Common equipment includes: ultrasonic sensors: used to detect the occupancy status of parking spaces, cameras: used for real-time monitoring and image recognition, RFID readers: used to identify vehicles entering the parking lot, and the status information of the parking lot is collected in real time through sensors and cameras, including: the occupancy status of each parking space (free / occupied), the working status of the lifting platform (free / busy), the current traffic flow and vehicle entry and exit records, and the collected real-time status information is integrated into the central database to ensure the real-time and accuracy of the data. Data acquisition software or IoT platform is used to integrate information to ensure smooth data communication between various devices. The acquired real-time status information is pre-processed, including data cleaning and format conversion, and redundant and erroneous status information is removed to ensure data quality. Image processing and machine learning algorithms (such as convolutional neural network CNN) are used to analyze the images acquired by the camera to identify which parking spaces are free. For sensor data, the feedback of the ultrasonic sensor is directly read to determine the occupancy status of the parking space, and the identified free parking space information is organized into a structure. Structured data (such as JSON format), including parking space number, location and status, collect status information of lift platforms, record the current working status (idle / busy) and location of each platform, use spatial analysis methods (such as Voronoi diagram or spatial cluster analysis) to analyze the distribution of idle lift platforms, determine the optimal distribution of idle platforms, evaluate the accessibility and utilization efficiency of different platforms, organize the analysis results, generate a spatial distribution map of idle lift platforms, point out the optimal usage plan, define the structure of the twin model, including parking spaces, lift platforms, channels and their relationships, integrate real-time status information, idle parking space identification results and lift platform distribution analysis results into the twin model, use modeling tools (such as Unity or Blender) to create a three-dimensional model to ensure that the model can reflect the status changes of the parking lot in real time, visualize the twin model so that it can dynamically display the real-time status of the parking lot, use virtual reality (VR) or augmented reality (AR) technology to enhance the user interaction experience, integrate the constructed twin model into the parking management system to ensure that it can interact with the real-time data stream and update the parking lot status in real time.

[0033] Step S3: optimally matching the parking lot twin model according to the three-dimensional morphological structure model of the incoming vehicle to obtain the optimal parking space; and making real-time parking management decisions to build a real-time parking management strategy; In this embodiment, a three-dimensional morphological structure model of the incoming vehicle is obtained, including the size, shape and space occupancy information of the vehicle, which is obtained through the three-dimensional model generated in the previous step. The information of the twin model of the three-dimensional parking lot is obtained, including the size, shape, position and current status (free or occupied) of each parking space. A suitable optimal matching algorithm is selected, such as the Hungarian algorithm, the greedy algorithm or the heuristic algorithm. These algorithms can effectively handle the matching problem between the vehicle and the parking space, consider the actual shape of the vehicle and the adaptability of the parking space to ensure the validity of the matching result, match the three-dimensional morphological structure model of the incoming vehicle with the parking space information of the three-dimensional parking lot, and the algorithm will compare the size of each vehicle with the size of the free parking space to determine the optimal deactivatable matching parking space, record the optimal matching parking space for each vehicle, including the parking space number and the matching available status, generate a list of the optimal deactivatable matching parking spaces, record the matching results, and store them in the parking management system for subsequent use, collect real-time status information, including the current occupancy of all parking spaces A real-time parking management decision model is constructed based on the status of the vehicle, the status of the lifting platform, and the information of the vehicles that are about to arrive. The following factors are taken into consideration: the result of the priority matching algorithm (i.e., the optimal available matching parking space), the occupancy of the current parking space, the working status and availability of the lifting platform. According to the matching results and real-time information, a parking management strategy is formulated. The strategy includes: specifying the target parking space for each vehicle and arranging the scheduling of the lifting platform, optimizing the order of parking space usage, ensuring that vehicles can quickly find vacant parking spaces and reduce waiting time, considering special circumstances (such as large vehicles or emergencies), adjusting the scheduling strategy to improve flexibility, applying the formulated parking management strategy to actual operations, instructing vehicles to go to the corresponding parking spaces in real time, updating the real-time data in the parking management system, and ensuring that the system can dynamically respond to changes in the parking lot. During the execution process, the parking management effect is monitored in real time, including the efficiency of vehicle parking and the utilization rate of parking spaces, collecting feedback information, and evaluating the effectiveness of the parking management strategy for subsequent strategy adjustments and optimizations.

[0034] Step S4: Obtain the parking lot's historical operation log; analyze the owner's historical vehicle use behavior in the parking lot's historical operation log, and then perform dynamic vehicle use time point prediction, thereby generating a vehicle owner's vehicle use prediction time point; In this embodiment, the storage location of the historical operation log in the parking lot management system is determined to ensure access to relevant data. The log should include vehicle entry and exit records, parking time, owner information (such as license plate number, identity information), etc. The historical operation log is extracted from the database using a data extraction tool (such as SQL query, Python script, etc.) to ensure that the extracted data contains necessary fields such as timestamp, owner identification and parking space information. The extracted historical operation log is cleaned to remove invalid records (such as duplicate data, erroneous data, etc.), and the data format is unified to ensure data consistency and integrity. In preparation for subsequent analysis, the historical operation log is analyzed to extract the owner's vehicle usage behavior characteristics, including: vehicle usage frequency: counting the number of times each owner uses the vehicle in a specific time period, vehicle usage period: analyzing the owner's vehicle usage peak period, such as morning peak, evening peak, etc., parking time: calculating the parking time of each vehicle, identifying the behavior patterns of short-term parking and long-term parking, and selecting a suitable analysis method, such as time series analysis, cluster analysis or association rule mining, and analyzing according to the vehicle usage habits of different owners. Based on the analysis results, a vehicle owner's vehicle usage behavior model is generated, including The car usage patterns and characteristics of different types of car owners are recorded and analyzed to facilitate the subsequent prediction of car usage time points. Suitable prediction models are selected, such as time series prediction models (ARIMA), machine learning models (such as random forests, support vector machines) or deep learning models (such as LSTM) for car usage time point prediction. Feature selection and construction Based on the results of the analysis of the car owner's historical car usage behavior, relevant features are selected as model inputs, such as: the car owner's car usage frequency, historical data of car usage timestamps, holidays, weather and other external factors. The prediction model is trained using historical data, and the model parameters are adjusted to improve the accuracy of the prediction. The performance of the model is evaluated using methods such as cross-validation to ensure its generalization ability on new data. Based on the trained model, the future car usage time points are predicted to generate the car owner's car usage prediction time points. The prediction results, including the estimated car usage time of each car owner, are recorded to facilitate the parking management system to schedule and allocate resources. The accuracy of the prediction results is monitored in actual applications, and the comparison data between the actual car usage time and the predicted time is collected to evaluate the effect of the prediction model. The prediction model is adjusted according to the feedback information to continuously optimize the accuracy of the prediction.

[0035] Step S5: Based on the predicted time point of the car owner's car use, the twin model of the three-dimensional parking lot is analyzed for the optimal vehicle scheduling time and the advance vehicle scheduling decision, and the advance vehicle scheduling strategy is constructed; In this embodiment, the predicted time point data of the car owners are collected, including the estimated arrival time of each car owner and the corresponding parking space information, and the real-time status information of the multi-story parking lot is obtained, especially the current occupancy of the parking space and the working status of the lifting platform. A suitable scheduling time analysis model is selected, such as the shortest path algorithm (Dijkstra or A* algorithm) or a scheduling optimization algorithm (such as genetic algorithm, mixed integer linear programming), and the model input parameters are determined, including the current status of the vehicle, the estimated arrival time, the parking space information and the status of the lifting platform. According to the predicted time point of the car owner's car use, the optimal scheduling time required for the vehicle to arrive at the reserved parking space from the current position is calculated, and factors including traffic flow, processing time of the lifting platform, layout of the parking lot and availability of parking spaces are considered. The optimal scheduling time for each vehicle is generated and recorded in the parking management system for subsequent scheduling decision-making. Based on the analysis results of the optimal scheduling time, a pre-vehicle scheduling decision is constructed. The decision model is used to determine the goals of the decision strategy, such as minimizing vehicle waiting time, improving parking space utilization, or optimizing the efficiency of the lift platform. The pre-vehicle scheduling strategy is formulated, which mainly includes the following aspects: Scheduling priority: According to the owner's expected arrival time, the scheduling priority is assigned to the vehicle to ensure that the vehicle is picked up and dropped off on time. Parking space allocation: The optimal parking space is assigned to each vehicle to ensure that there is an available parking space when the vehicle arrives. Lift platform scheduling: Arrange necessary lift platform scheduling to ensure that the platform can be ready to receive the vehicle before the optimal scheduling time. The formulated pre-vehicle scheduling strategy is applied to the parking management system, which indicates the scheduling of lift platforms and vehicles in real time, monitors the execution of scheduling, ensures that the vehicle can arrive at the target parking space at the scheduled time, collects feedback information after scheduling execution, including actual arrival time, parking space occupancy and user satisfaction, evaluates the scheduling strategy, identifies potential room for improvement, and optimizes the scheduling strategy based on feedback information.

[0036] Step S6: Conduct intelligent collaborative parking vehicle transfer learning on the real-time parking management strategy and the advance vehicle scheduling strategy, and perform iterative reinforcement optimization to build a smart parking management model for the three-dimensional parking lot.

[0037] In this embodiment, historical data of real-time parking management strategies and pre-use vehicle scheduling strategies are collected, including scheduling decisions, execution results, and user feedback. The data is collated to ensure that the data includes the diversity of various parking lot environments and owner behaviors to improve the adaptability of the model. A suitable transfer learning model is selected, such as a convolutional neural network (CNN) or a recurrent neural network (RNN) in a deep learning framework (such as TensorFlow or PyTorch). The input features of the model are determined, including the owner's behavioral characteristics, historical scheduling strategies, and real-time status information. The transfer learning model is trained using historical data to optimize model parameters so that it can recognize and learn effective collaborative scheduling. Strategy, use cross-validation method to evaluate model performance and ensure its generalization ability on new data. After training, the model generates intelligent collaborative scheduling strategy, determines how to effectively coordinate between real-time management and pre-scheduling, stores the generated strategy in the system for real-time application, selects appropriate reinforcement learning algorithm, such as Q-learning or deep reinforcement learning (such as DQN), to iteratively optimize the intelligent collaborative scheduling strategy, determines the reward function, and designs a reasonable reward mechanism to guide the model to optimize the scheduling strategy. Simulate the parking environment, including owner behavior, scheduling decisions and real-time feedback, build a training environment for reinforcement learning, set the state space (such as vehicle state, The model continuously updates its strategy through interaction with the environment, collects data generated during the training process, including the results and feedback information of each dispatch, for subsequent analysis, evaluates the performance of the model in practical applications, and analyzes the actual effect of the dispatch strategy, including parking space utilization, user satisfaction and average waiting time. Based on the evaluation results, the model is adjusted and optimized to ensure that it can adapt to the dynamically changing parking environment, and the optimized intelligent collaborative dispatch strategy and reinforcement learning model are integrated into the smart parking management system of the three-dimensional parking lot to ensure that the model can Receive parking lot status information in real time and dynamically update strategies based on the latest data; design a user interface so that parking lot managers can easily monitor the system operating status and the effectiveness of scheduling strategies; display real-time data and the decision-making process of the model through visualization tools to improve managers' decision-making support capabilities; conduct system testing in an actual environment to verify the effectiveness and stability of the smart parking management model; make final adjustments based on test feedback to ensure that the system can be successfully launched and run; establish a continuous monitoring mechanism for the system; collect parking lot operation data and user feedback in real time; regularly evaluate and update the model to ensure that the smart parking management model can always adapt to changes in the parking lot environment.

[0038] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: When a vehicle is detected entering the venue, a multi-angle monitoring image of the vehicle entering the venue is collected by a camera; Step S12: performing image multi-view registration and reconstruction on the multi-angle monitoring image of the incoming vehicle to construct a panoramic reconstructed vehicle monitoring image; Step S13: performing visual recognition of vehicle structure key points on the panoramic reconstructed vehicle monitoring image, and marking a plurality of vehicle structure key points; Step S14: performing vehicle geometry analysis on a plurality of key points of the vehicle structure to generate geometric features of the incoming vehicle; Step S15: performing vehicle size type recognition on the geometric features of the incoming vehicle to obtain the size type features of the incoming vehicle; Step S16: Perform vehicle three-dimensional point cloud modeling based on the size and type characteristics of the incoming vehicle to construct a three-dimensional morphological structure model of the incoming vehicle.

[0039] In this embodiment, multiple surveillance cameras are installed in the entrance area to ensure that the images of vehicles can be captured from multiple angles at the same time. These cameras should cover all viewing angles of vehicle entry. When a vehicle entry is detected, the camera system is triggered to automatically capture multi-angle images of the entering vehicle, ensure the time synchronization of image acquisition, record the shooting time of each camera for subsequent alignment, store the acquired image data in a local server or cloud for subsequent processing and analysis, pre-process the acquired images, including denoising, adjusting brightness and contrast to improve image quality, and use feature extraction algorithms (such as SIFT, ORB or SURF) to identify key points and feature descriptors in the image to ensure that the image is not Find corresponding points in images with the same perspective, use camera calibration information and feature matching algorithms to align images from different perspectives, use the RANSAC algorithm to eliminate false matches and ensure the accuracy of alignment, use image stitching technology to synthesize the aligned images into a panoramic reconstructed vehicle monitoring image to form a continuous view that shows the full picture of the vehicle. In the panoramic reconstructed image, apply deep learning models (such as YOLO and MaskR-CNN) to detect and identify key points of the vehicle structure, find important parts of the vehicle body (such as wheels, lights, windows, etc.), mark the identified key points, and form key point coordinates. These coordinates will be used for subsequent geometric analysis. The data is stored in a structured format (such as JSON or CSV) for subsequent processing and analysis. Based on the marked multiple key points of the vehicle structure, a geometric model of the vehicle is constructed. Three-dimensional modeling software (such as Blender or CAD software) is used to assist in the construction. The geometric features of the vehicle, including the length, width, height, wheelbase, etc. of the vehicle, are calculated by calculating the distance between the key points. The generated geometric feature data is sorted out to form a report that records the basic geometric information of the vehicle. Based on the geometric features of the vehicle, a classification algorithm (such as a decision tree, support vector machine, or deep learning model) is applied to identify the size type of the vehicle (such as a small car, a mid-size car, an SUV, etc.). The extracted geometric features are Match with the predefined vehicle size type library to determine the size type characteristics of the incoming vehicles, organize and record the identified vehicle size type characteristics for subsequent use and analysis, and use 3D modeling tools (such as PCL or Open3D) to generate a 3D point cloud model of the vehicle based on the geometric features and size type of the incoming vehicles. Optimize and filter the generated point cloud model to remove noise points and smooth the surface to enhance the model quality. Convert the optimized point cloud model into a visual 3D morphological structure model to ensure that the model can accurately reflect the appearance and structure of the vehicle. Store the constructed 3D morphological structure model in a suitable format (such as OBJ or STL) for subsequent application or analysis.

[0040] In this embodiment, the specific steps of step S12 are: Performing image brightness enhancement on the multi-angle monitoring image of the vehicle entering the venue to obtain a brightness enhanced monitoring image; Identify abnormal image noise points on the brightness-enhanced monitoring image and mark multiple abnormal image noise points; Perform non-local mean filtering on multiple image abnormal noise points to obtain a filtered optimized monitoring image; Perform vehicle multi-angle edge feature analysis on the filtered optimized monitoring image and extract vehicle multi-angle edge feature data; Perform detail sharpening and enhancement processing based on multi-angle edge feature data of the vehicle, thereby obtaining multi-angle sharp detail features; The filtered optimized monitoring image is reconstructed by multi-view registration based on the multi-angle sharpening detail features to construct a panoramic reconstructed vehicle monitoring image.

[0041] In this embodiment, multi-angle monitoring images of vehicles entering the venue are collected to ensure that the image format is uniform and the quality is good. A suitable image enhancement method is selected, such as histogram equalization, gamma correction or adaptive histogram equalization (CLAHE). Each monitoring image is processed using the selected enhancement algorithm. For example, the cv2.equalizeHist() function of the OpenCV library is used to perform histogram equalization. The enhanced image is recorded and marked as a "brightness enhanced monitoring image". A suitable abnormal noise point recognition method is selected, such as threshold-based detection, edge detection or image segmentation method. Noise point recognition is performed on the brightness enhanced monitoring image. For example, a high-pass filter or Sobel operator is used for edge detection to identify abnormal noise points, mark the identified abnormal noise points, generate a marked image, and select a non-local mean (NLM) filtering method. This method can effectively remove noise while maintaining image details. The non-local mean filtering function ( FastNlMeansDenoisingColored()), record the processed image, mark it as "filter optimization monitoring image", select an edge detection method, such as Canny edge detection or Laplacian method, apply the edge detection algorithm to the filter optimization monitoring image, extract the edge features of the vehicle, record the extracted edge feature data, form an edge feature image, select a suitable image sharpening method, such as Laplace sharpening or high-pass filtering, apply the sharpening algorithm to the multi-angle edge feature data of the vehicle, enhance the details, record the enhanced image, mark it as "multi-angle sharpening detail features", select a multi-view registration algorithm, such as feature matching, phase-correlation-based method or structured light registration, use the extracted multi-angle sharpening detail features to register and reconstruct the filter optimization monitoring image, ensure that images from different perspectives can be accurately aligned, record the reconstructed panoramic image, and mark it as "panoramic reconstructed vehicle monitoring image".

[0042] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: Acquire real-time status information of the three-dimensional parking lot; Step S22: Identify the vacant parking spaces based on the real-time status information of the multi-story parking lot, and extract the vacant parking space status information; Step S23: positioning the idle lifting platform of the three-dimensional parking lot based on the idle parking space status information, and extracting the positioning data of the idle lifting platform; Step S24: performing spatial distribution analysis on the idle lifting platform positioning data to generate idle lifting platform spatial distribution data; Step S25: Render the spatial distribution data of idle lifting platforms in real time according to the parking restriction standards to construct a twin model of the three-dimensional parking lot.

[0043] In this embodiment, real-time monitoring equipment, such as sensors, cameras, and Internet of Things devices, are installed in the three-dimensional parking lot to collect parking lot status information, ensure that the equipment can transmit data in real time, use a wireless network or a wired network for data transmission, and regularly collect real-time status information of the parking lot through the equipment, including parking space occupancy, lifting platform status, and other related information. The collected data is uploaded to the central processing system for subsequent analysis, and the acquired real-time status information is pre-processed to filter out irrelevant data to ensure the accuracy and completeness of the data. Image processing or data analysis algorithms are used to identify vacant parking spaces in the parking lot. For example, a target detection algorithm based on deep learning (such as YOLO or FasterR-CNN) effectively identifies vacant parking spaces, extracts status information of vacant parking spaces, including parking space numbers and location coordinates, and stores the identified vacant parking space status information in a database for use in subsequent steps. Based on the vacant parking space status information, a positioning algorithm is designed to determine which lifting platform serves the vacant parking space. A geometric calculation or a graphical algorithm is used to calculate the distance between the lifting platform and the vacant parking space in combination with the plane layout of the parking lot. Extract the positioning data of available lifts, including the correspondence between the positions of lifts and vacant parking spaces, organize and store the positioning data of vacant lifts to form a data set for subsequent analysis, select appropriate spatial analysis methods, such as cluster analysis, heat map analysis, or spatial interpolation, to analyze the distribution of vacant lifts, perform spatial distribution analysis on the extracted positioning data of vacant lifts, identify areas where lifts are highly idle and have low utilization rates, generate spatial distribution data of vacant lifts, including distribution heat maps or other visual representations, organize the generated spatial distribution data, and record them in the form of reports or visual graphics for subsequent use, set real-time parking restriction standards based on the management rules and real-time status information of the parking lot, including usage restrictions for each parking space and lift, use 3D modeling tools (such as Unity or Blender) to render the spatial distribution data of vacant lifts and parking restriction standards into the digital twin model of the stereoscopic parking lot, ensure that the model can be dynamically updated to reflect the status changes of the parking lot in real time, integrate the rendered twin model, and apply it to the real-time monitoring and management system to optimize the operational efficiency of the parking lot.

[0044] In this embodiment, the specific steps of step S25 are: Conducting a three-dimensional topological structure analysis on the three-dimensional parking lot to generate the three-dimensional topological structure features of the parking lot; Based on the three-dimensional topological structure characteristics of the parking lot, the spatial distribution data of the idle lifting platforms are fitted to construct a three-dimensional distribution network of idle platforms. Perform deactivation type analysis on the idle parking space status information to generate the deactivation vehicle type for each idle parking space; Analyze the parking limit standard of each idle parking space based on the type of parked vehicles, and extract the parking limit standard of each idle parking space; Based on the size limit parking standards of each vacant parking space, the vacant parking space attributes of the 3D distribution network of vacant platforms are rendered in real time to construct a twin model of the 3D parking lot.

[0045] In this embodiment, the spatial layout data of the parking lot is collected, including the specific location and size information of parking spaces, lifting platforms, passages, etc. The structural information of the parking lot is obtained by using CAD drawings or three-dimensional modeling software (such as Revit or AutoCAD). A topological structure model of the parking lot is created in the analysis tool. The parking lot is regarded as a graph using graph theory methods. The nodes represent parking spaces and lifting platforms, and the edges represent passages and connections. The connection relationship between the elements is identified, and their relative positions and access paths are recorded. The key features in the topological structure, such as the connectivity of parking spaces, the accessibility of lifting platforms, the width of passages, etc., are analyzed to generate three-dimensional topological structure feature data of the parking lot, including connectivity, efficiency, and accessibility indicators. The spatial distribution data of idle lifting platforms, including their locations and idle states, are collected. A suitable spatial distribution fitting algorithm, such as Kriging interpolation, local weighted regression, or polynomial regression, is selected. The data of idle lifting platforms are analyzed, the spatial distribution data of the idle platforms are fitted, and a three-dimensional distribution network model of the idle platforms is generated. The fitting results are visualized to display the spatial distribution trend of the idle platforms. The status information of each idle parking space, including the parking space number, size, location, etc., is extracted from the real-time status information. The size, shape and location of parking spaces are analyzed. The types of vehicles parked in each vacant parking space (e.g., small cars, medium-sized cars, large cars) are analyzed. A rule engine or classification algorithm (e.g., decision tree or support vector machine) is used to automatically identify the types of vehicles that can be parked. The types of vehicles that can be parked in each vacant parking space are organized into a data set for subsequent use. Size restrictions are defined for each type of vehicle that can be parked, including length, width and height restrictions. The types of vehicles that can be parked in each vacant parking space are analyzed, and the corresponding size restrictions are extracted. The applicable vehicle size range for each vacant parking space is calculated, and the relevant data is recorded. The size limit parking standards for each vacant parking space are formed into a data table or database for subsequent use. Based on the size limit parking standards for each vacant parking space and the three-dimensional distribution network of idle platforms, a real-time rendering solution is designed. A three-dimensional visualization tool (such as Unity or UnrealEngine) is used to render the twin model of the parking lot in real time to display the attributes of each vacant parking space. This ensures that the rendering results can be dynamically updated to reflect the parking lot status and parking space attributes in real time. The rendered three-dimensional parking lot twin model is integrated into the parking management system to provide a user-friendly interface for displaying vacant parking space information and real-time status.

[0046] In this embodiment, refer to Figure 4, is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: performing a parking simulation of an idle parking space on the twin model of the three-dimensional parking lot according to the three-dimensional morphological structure model of the incoming vehicle to generate idle parking space parking simulation data; Step S32: performing optimal deactivatable matching on the idle parking space parking simulation data to obtain the optimal deactivatable matching parking space; Step S33: performing lifting and lowering scheduling of the three-dimensional parking space for the optimal deactivatable matching parking space, and generating lifting and lowering scheduling data of the three-dimensional parking space; Step S34: performing parking path planning based on the parking space scheduling data to generate a parking planning path; Step S35: Make real-time parking management decisions based on the parking planning path and the three-dimensional parking space lifting and dispatching data, and build a real-time parking management strategy.

[0047] In this embodiment, a three-dimensional morphological structure model of the incoming vehicle is obtained to ensure that it includes the size, shape and space occupancy information of the vehicle. The vehicle model is integrated with the twin model of the three-dimensional parking lot to ensure that the model can reflect the actual layout and structure of the parking lot. A simulation tool (such as MATLAB, Simulink or Unity) is used to simulate the parking of the vacant parking spaces. According to the size and shape of the incoming vehicle, the parking process of the vehicle in the parking lot is simulated to generate parking simulation data of the vacant parking spaces. The generated parking simulation data is recorded, including the occupancy status, parking time and space utilization rate of each parking space. etc., design the optimal deactivatable matching algorithm, use the vehicle size and parking space information for matching, use the greedy algorithm, heuristic algorithm or optimization algorithm, perform the optimal deactivatable matching based on the idle parking space parking simulation data, identify the parking space suitable for parking, generate the optimal deactivatable matching parking space data set, including the parking space number and the corresponding vehicle information, organize the matching results, form a report or data table for subsequent scheduling, design the lifting and lowering scheduling strategy according to the optimal deactivatable matching result, consider the vehicle entry and exit sequence and parking space utilization efficiency, perform the three-dimensional parking space lifting and lowering scheduling for the optimal deactivatable matching parking space, and determine The lifting and lowering sequence of each parking space is used to generate the lifting and lowering scheduling data of the three-dimensional parking space, including lifting instructions, time schedule and operation sequence. The scheduling data is organized into documents or databases to ensure the accuracy and traceability of the scheduling information. A suitable path planning algorithm, such as A* algorithm, Dijkstra algorithm or RRT (fast random tree), is selected to generate the parking path. Based on the parking scheduling data of the parking space, parking path planning is performed to calculate the optimal path from the entry position to the target parking space, considering the actual layout of the parking lot, obstacles and traffic flow to ensure the feasibility of the path, and the generated parking planning path, including path points and driving direction, is recorded for subsequent use. Based on the parking planning path and the lifting and lowering scheduling data of the three-dimensional parking space, a real-time parking management decision model is constructed, considering the vehicle entry and exit, parking space status and lifting and lowering scheduling. The real-time parking management strategy is implemented in the parking lot management system. The parking scheduling and path planning are dynamically adjusted according to the vehicle entry and exit. The effectiveness and flexibility of the decision are ensured through real-time feedback from the monitoring system. The status of the parking lot is continuously monitored, the effect of the parking management decision is recorded, and necessary adjustments and optimizations are made. The feedback information is used for model improvement to improve the efficiency of parking management in the future.

[0048] In this embodiment, step S4 includes the following steps: Step S41: Obtain parking lot historical operation log; Step S42: Analyze the historical car owner's car-taking behavior in the parking lot's historical operation log to generate dynamic car owner behavior features; Step S43: quantify the vehicle usage time based on the dynamic vehicle owner behavior characteristics to obtain the vehicle owner's multiple vehicle usage timestamps; Step S44: mining personalized car-use behavior features based on the car owner's multiple car-use timestamps to generate the car owner's car-use behavior rules; Step S45: Perform dynamic vehicle use time point prediction based on the vehicle owner's vehicle use behavior pattern, thereby generating a vehicle owner's vehicle use prediction time point.

[0049] In this embodiment, historical operation logs are extracted from the parking lot management system to ensure that the logs contain the owner's entry and exit records, parking time and parking space usage. The record data format should include timestamp, owner identity information, license plate number, parking space number, etc. The collected historical operation logs are stored in a database to ensure data integrity and accessibility for subsequent analysis. The historical operation logs are cleaned to remove redundant data and erroneous records to ensure data accuracy. A suitable analysis model is selected, such as cluster analysis or sequence pattern mining, to analyze the owner's historical vehicle pick-up and use behavior, extract dynamic owner behavior characteristics, such as vehicle use frequency, vehicle use time period and parking time, etc., to generate the owner's behavior characteristic data set, organize the analysis results into a report, record the behavior characteristics of each owner for subsequent use, extract the vehicle use timestamp from the owner's behavior characteristic data, including the time of each vehicle pick-up and return, count the vehicle use time of each owner, analyze the peak and off-peak hours of vehicle use, generate quantitative statistical results, and use statistical analysis tools (such as Python's Pan das library) for data processing and statistics, organize the generated timestamps of multiple car usage by car owners to form a data set for subsequent analysis, select appropriate mining methods, such as decision trees, association rule learning or time series analysis, and deeply explore the car owner's car usage behavior patterns. Based on the car owner's multiple car usage timestamps, analyze the characteristics of individual car usage behaviors, identify personalized car usage patterns and patterns, and generate car owner's car usage behavior pattern data, including the difference in car usage on weekends and weekdays, and the car usage trend on specific holidays. According to the car owner's car usage behavior patterns, select appropriate prediction models, such as time series prediction models (ARIMA) and machine learning models (such as random forests or support vector machines) to predict the time points of car usage. Use historical car usage timestamps and behavior feature data to train the prediction model, optimize model parameters, ensure the accuracy of the prediction, and predict dynamic car usage time points based on the trained model to generate future car usage prediction time points. Organize the generated car owner's car usage prediction time points to form a report for use by the management system to help optimize parking management.

[0050] In this embodiment, step S5 includes the following steps: Step S51: Calculate the current vehicle position on the three-dimensional parking lot twin model to obtain the parking lot position of the current vehicle; Step S52: performing a planar platform dispatching path analysis according to the current parking lot position of the vehicle to generate a current vehicle platform dispatching path; Step S53: Calculate the path scheduling time for the current vehicle platform scheduling path, so as to obtain the current vehicle path scheduling time; Step S54: performing an optimal scheduling time analysis on the current vehicle path scheduling time based on the predicted time point of the vehicle owner's vehicle use, and generating an optimal vehicle scheduling time; Step S55: Make a pre-vehicle scheduling decision for the optimal vehicle scheduling time and construct a pre-vehicle scheduling strategy.

[0051] In this embodiment, the identification information (such as the license plate number) of the current vehicle is obtained from the parking lot management system, and the real-time position of the vehicle is queried. The twin model (3D model) of the parking lot is used to determine the exact position of the current vehicle in the three-dimensional parking lot. The parking lot coordinates of the current vehicle are calculated by matching the position coordinates of the vehicle with the parking space information in the model, including the parking space number, floor information and relative position. The position data of the current vehicle is recorded in the system to ensure that the subsequent steps can smoothly access this information. In the plane layout of the parking lot, a scheduling path model is constructed to identify the position and connection relationship of the parking space, lifting platform and passage. A suitable path analysis algorithm, such as the A* algorithm or the Dijkstra algorithm, is selected to calculate the optimal path from the current vehicle position to the lifting platform. According to the current vehicle position, the path analysis algorithm is executed to generate the platform scheduling path of the current vehicle, record each key point in the path, including the entrance and exit of the passage and the lifting platform, calculate the length of the scheduling path, and determine the time required for the platform where the vehicle is located to move on the path. If the path involves the movement of the lifting platform, the scheduling time of the lifting platform needs to be calculated separately, including the time required for lifting. time and waiting time, add the path scheduling time to the lifting platform movement time to obtain the total path scheduling time of the current vehicle, record the calculated total path scheduling time in the system for subsequent analysis and decision-making, obtain the predicted vehicle use time point from the owner's vehicle use prediction module, determine when the vehicle should arrive at the lifting platform, compare the path scheduling time of the current vehicle with the predicted vehicle use time point, analyze whether the scheduling time needs to be adjusted, use optimization algorithms (such as linear programming or heuristic algorithms) to calculate the optimal scheduling time to ensure that the vehicle can arrive at the lifting platform before the predicted vehicle use time point, organize the generated optimal vehicle scheduling time, and form a record for subsequent scheduling decisions. According to the optimal vehicle scheduling time, design a pre-use vehicle scheduling strategy, clarify the scheduling instructions and execution time of the lifting platform, apply the pre-use vehicle scheduling strategy to the parking management system, and automatically generate scheduling instructions to ensure that the lifting platform is ready to receive the vehicle before the optimal scheduling time. During the scheduling execution process, monitor the movement and scheduling execution of the vehicle to ensure that the predetermined scheduling strategy is effective, collect feedback information, and use it to optimize future scheduling strategies to ensure the flexibility and responsiveness of the system.

[0052] In this embodiment, step S6 includes the following steps: Step S61: performing intelligent collaborative parking vehicle transfer learning on the real-time parking management strategy and the pre-use vehicle scheduling strategy to generate an intelligent collaborative scheduling strategy; Step S62: Perform dynamic management efficiency analysis on the intelligent collaborative scheduling strategy to generate collaborative scheduling strategy management efficiency; Step S63: Iteratively strengthen and optimize the intelligent collaborative scheduling strategy according to the management efficiency of the collaborative scheduling strategy to construct a smart parking management model for the multi-story parking lot.

[0053] In this embodiment, historical data of real-time parking management strategies and pre-scheduling strategies are collected, including scheduling decisions, execution results, and system feedback, to ensure that the data set covers different parking lot conditions and owner behavior patterns to improve the generalization ability of the model, select a suitable transfer learning framework, use a deep learning model (such as a convolutional neural network or a recurrent neural network) for strategy learning, define a model structure so as to simultaneously process data of real-time management strategies and pre-scheduling strategies, use the collected data to train the transfer learning model, adjust the model parameters through an optimization algorithm (such as Adam or RMSprop) so that it can identify and learn effective collaborative scheduling strategies, and use cross-validation and other methods to evaluate the model performance to ensure that it can be effectively executed on new data. After the training is completed, the model is used to generate an intelligent collaborative scheduling strategy to determine how to effectively collaborate between real-time management and pre-scheduling, and establish a management efficiency analysis framework, including key performance indicators (KPIs), such as average scheduling time, parking space utilization, and user satisfaction. After implementing the intelligent collaborative scheduling strategy, relevant performance data is collected to record the scheduling execution status and results, including vehicle entry and exit time, parking space utilization, and user satisfaction. occupancy, etc., use statistical analysis methods (such as regression analysis or decision trees) to analyze the collected data, calculate the management efficiency of the collaborative scheduling strategy and compare the management efficiency indicators before and after implementation, generate a management efficiency report, identify improvement points, organize the analysis results, and form a report to facilitate subsequent strategy optimization and decision support, select a suitable reinforcement learning algorithm (such as Q-learning or deep reinforcement learning) to optimize the intelligent collaborative scheduling strategy, ensure that the algorithm can handle the dynamically changing parking environment, and iterate the strategy based on the results of the dynamic management efficiency analysis. Use the feedback mechanism of reinforcement learning to optimize scheduling decisions and improve the effect of collaborative scheduling. In actual applications, continuously collect feedback data, update the model and adjust the strategy, integrate the optimized scheduling strategy into the smart parking management model of the multi-story parking lot, ensure that the model can adapt to changes in the parking lot in real time, design the model interface to facilitate data interaction with the real-time monitoring system of the parking lot, integrate the final smart parking management model, monitor and evaluate it, ensure its effectiveness and stability in actual applications, set monitoring indicators, regularly evaluate the performance of the model, and make necessary adjustments and optimizations.

[0054] In this embodiment, a smart parking management system is provided, which is used to execute the smart parking management method as described above, including: An image registration and reconstruction module is used to collect multi-angle monitoring images of vehicles entering the venue; perform image multi-view registration and reconstruction on the multi-angle monitoring images of vehicles entering the venue, perform vehicle three-dimensional point cloud modeling, and construct a three-dimensional morphological structure model of vehicles entering the venue; The twin model module is used to obtain the real-time status information of the three-dimensional parking lot; identify the vacant parking spaces based on the real-time status information of the three-dimensional parking lot, analyze the spatial distribution of the idle lifting platforms, and build a twin model of the three-dimensional parking lot; The deactivatable matching module is used to perform optimal deactivatable matching on the twin model of the three-dimensional parking lot according to the three-dimensional morphological structure model of the incoming vehicle, obtain the optimal deactivatable matching parking space, and make real-time parking management decisions to build a real-time parking management strategy; The vehicle use time prediction module is used to obtain the parking lot's historical operation logs; analyze the owner's historical vehicle use behavior in the parking lot's historical operation logs, and then predict the dynamic vehicle use time points, thereby generating the owner's vehicle use prediction time points; The pre-use vehicle scheduling module is used to analyze the optimal vehicle scheduling time and pre-use vehicle scheduling decisions for the three-dimensional parking lot twin model based on the predicted time point of the car owner's car use, and to build a pre-use vehicle scheduling strategy; The smart parking management module is used to conduct intelligent collaborative parking vehicle transfer learning for real-time parking management strategies and pre-use vehicle scheduling strategies, and to perform iterative reinforcement optimization to build a smart parking management model for three-dimensional parking lots.

[0055] The present invention helps to accurately capture the three-dimensional morphological information of the vehicle by performing multi-angle image registration and reconstruction on the multi-angle monitoring images of the incoming vehicles, thereby realizing accurate modeling of the vehicle. The module uses image processing technology to reconstruct the monitoring images from multiple angles into a three-dimensional point cloud model, providing high-precision vehicle morphological structure data for subsequent intelligent parking management, and obtaining real-time status information of the three-dimensional parking lot, including vacant parking space identification and spatial distribution analysis of vacant lifting platforms, to construct a twin model of the three-dimensional parking lot. The module provides accurate control of the parking lot status by combining real-time data and model prediction, and provides real-time support for subsequent parking space scheduling and management. It performs optimal deactivatable matching based on the vehicle three-dimensional morphological structure model and the parking lot twin model, helping to achieve optimal parking space allocation and real-time parking management decisions. The module provides the best parking position for the vehicle through an intelligent matching algorithm, thereby improving Parking lot utilization and management efficiency, through historical operation logs and car owner's car usage behavior analysis, to achieve accurate prediction of the time point of car owner's car use, and then generate the predicted time point of car owner's car use. This module uses data analysis and prediction algorithms to help parking lot managers better understand car owners' car usage habits, and provide important reference for parking lot resource scheduling and management. According to the predicted time point of car owners' car use, the optimal vehicle scheduling and pre-vehicle scheduling decisions are made, and an efficient pre-vehicle scheduling strategy is constructed. This module optimizes vehicle scheduling plans and improves parking lot resource utilization efficiency by analyzing the predicted car use time point and parking lot status. It conducts intelligent collaborative learning and iterative optimization of parking management strategies and pre-vehicle scheduling strategies to construct a smart parking management model. This module integrates the data and decisions of each module, and continuously improves parking lot management strategies through machine learning and optimization algorithms, thereby improving parking lot operation efficiency and user experience.

[0056] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0057] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be 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 present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A smart parking management method, characterized in that: The following steps are involved: Step S1: Collect multi-angle monitoring images of vehicles entering the venue; Performing multi-view registration and reconstruction on the multi-angle monitoring images of the vehicle entering the venue, and performing three-dimensional point cloud modeling of the vehicle to construct a three-dimensional morphological structure model of the vehicle entering the venue; Step S2: obtaining real-time status information of the three-dimensional parking lot; identifying vacant parking spaces and analyzing the spatial distribution of vacant lifting platforms based on the real-time status information of the three-dimensional parking lot, and constructing a twin model of the three-dimensional parking lot; Step S3: performing optimal deactivatable matching on the twin model of the stereoscopic parking lot according to the three-dimensional morphological structure model of the incoming vehicle to obtain the optimal deactivatable matching parking space; And make real-time parking management decisions and build real-time parking management strategies; Step S4: Obtain the parking lot's historical operation log; analyze the owner's historical vehicle use behavior in the parking lot's historical operation log, and then perform dynamic vehicle use time point prediction, thereby generating the owner's vehicle use prediction time point; Step S5: Based on the predicted time point of the car owner's car use, the twin model of the three-dimensional parking lot is analyzed for the optimal vehicle scheduling time and the advance vehicle scheduling decision, and the advance vehicle scheduling strategy is constructed; Step S6: Conduct intelligent collaborative parking vehicle transfer learning on the real-time parking management strategy and the advance vehicle scheduling strategy, and perform iterative reinforcement optimization to build a smart parking management model for the three-dimensional parking lot.

2. The smart parking management method according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: When a vehicle is detected entering the venue, a multi-angle monitoring image of the vehicle entering the venue is collected by a camera; Step S12: performing image multi-view registration and reconstruction on the multi-angle monitoring image of the incoming vehicle to construct a panoramic reconstructed vehicle monitoring image; Step S13: performing visual recognition of vehicle structure key points on the panoramic reconstructed vehicle monitoring image, and marking a plurality of vehicle structure key points; Step S14: performing vehicle geometry analysis on a plurality of key points of the vehicle structure to generate geometric features of the incoming vehicle; Step S15: performing vehicle size type recognition on the geometric features of the incoming vehicle to obtain the size type features of the incoming vehicle; Step S16: Perform vehicle three-dimensional point cloud modeling based on the size and type characteristics of the incoming vehicle to construct a three-dimensional morphological structure model of the incoming vehicle.

3. The smart parking management method according to claim 2, characterized in that: The specific steps of step S12 are: Performing image brightness enhancement on the multi-angle monitoring image of the vehicle entering the venue to obtain a brightness enhanced monitoring image; Identify abnormal image noise points on the brightness-enhanced monitoring image and mark multiple abnormal image noise points; Perform non-local mean filtering on multiple image abnormal noise points to obtain a filtered optimized monitoring image; Perform vehicle multi-angle edge feature analysis on the filtered optimized monitoring image and extract vehicle multi-angle edge feature data; Perform detail sharpening and enhancement processing based on multi-angle edge feature data of the vehicle, thereby obtaining multi-angle sharp detail features; The filtered optimized monitoring image is reconstructed by multi-view registration based on the multi-angle sharpening detail features to construct a panoramic reconstructed vehicle monitoring image.

4. The smart parking management method according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: Acquire real-time status information of the three-dimensional parking lot; Step S22: Identify the vacant parking spaces based on the real-time status information of the multi-story parking lot, and extract the vacant parking space status information; Step S23: positioning the idle lifting platform of the three-dimensional parking lot based on the idle parking space status information, and extracting the positioning data of the idle lifting platform; Step S24: performing spatial distribution analysis on the idle lifting platform positioning data to generate idle lifting platform spatial distribution data; Step S25: Render the spatial distribution data of idle lifting platforms in real time according to the parking restriction standards to construct a twin model of the three-dimensional parking lot.

5. The smart parking management method according to claim 4, characterized in that: The specific steps of step S25 are: Conducting a three-dimensional topological structure analysis on the three-dimensional parking lot to generate the three-dimensional topological structure features of the parking lot; Based on the three-dimensional topological structure characteristics of the parking lot, the spatial distribution data of the idle lifting platforms are fitted to construct a three-dimensional distribution network of idle platforms. Perform deactivation type analysis on the idle parking space status information to generate the deactivation vehicle type for each idle parking space; Analyze the parking limit standard of each idle parking space based on the type of parked vehicles, and extract the parking limit standard of each idle parking space; Based on the size limit parking standards of each vacant parking space, the vacant parking space attributes of the 3D distribution network of vacant platforms are rendered in real time to construct a twin model of the 3D parking lot.

6. The smart parking management method according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: performing a parking simulation of an idle parking space on the twin model of the three-dimensional parking lot according to the three-dimensional morphological structure model of the incoming vehicle to generate idle parking space parking simulation data; Step S32: performing optimal deactivatable matching on the idle parking space parking simulation data to obtain the optimal deactivatable matching parking space; Step S33: performing lifting and lowering scheduling of the three-dimensional parking space for the optimal deactivatable matching parking space, and generating lifting and lowering scheduling data of the three-dimensional parking space; Step S34: performing parking path planning based on the parking space scheduling data to generate a parking planning path; Step S35: Make real-time parking management decisions based on the parking planning path and the three-dimensional parking space lifting and dispatching data, and build a real-time parking management strategy.

7. The smart parking management method according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: Obtain parking lot historical operation log; Step S42: Analyze the historical car owner's car-taking behavior in the parking lot's historical operation log to generate dynamic car owner behavior characteristics; Step S43: quantify the vehicle usage time based on the dynamic vehicle owner behavior characteristics to obtain the vehicle owner's multiple vehicle usage timestamps; Step S44: mining personalized car-use behavior features based on the car owner's multiple car-use timestamps to generate the car owner's car-use behavior rules; Step S45: Perform dynamic vehicle use time point prediction based on the vehicle owner's vehicle use behavior pattern, thereby generating a vehicle owner's vehicle use prediction time point.

8. The smart parking management method according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: Calculate the current vehicle position on the three-dimensional parking lot twin model to obtain the parking lot position of the current vehicle; Step S52: performing a planar platform dispatching path analysis according to the current parking lot position of the vehicle to generate a current vehicle platform dispatching path; Step S53: Calculate the path scheduling time for the current vehicle platform scheduling path, so as to obtain the current vehicle path scheduling time; Step S54: performing an optimal scheduling time analysis on the current vehicle path scheduling time based on the predicted time point of the vehicle owner's vehicle use, and generating an optimal vehicle scheduling time; Step S55: Make a pre-vehicle scheduling decision for the optimal vehicle scheduling time and construct a pre-vehicle scheduling strategy.

9. The smart parking management method according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: performing intelligent collaborative parking vehicle transfer learning on the real-time parking management strategy and the pre-use vehicle scheduling strategy to generate an intelligent collaborative scheduling strategy; Step S62: Perform dynamic management efficiency analysis on the intelligent collaborative scheduling strategy to generate collaborative scheduling strategy management efficiency; Step S63: Iteratively strengthen and optimize the intelligent collaborative scheduling strategy according to the management efficiency of the collaborative scheduling strategy to construct a smart parking management model for the multi-story parking lot.

10. A smart parking management system, characterized in that: Used to execute the smart parking management method as claimed in claim 1, comprising: An image registration and reconstruction module is used to collect multi-angle monitoring images of vehicles entering the venue; perform image multi-view registration and reconstruction on the multi-angle monitoring images of vehicles entering the venue, perform vehicle three-dimensional point cloud modeling, and construct a three-dimensional morphological structure model of vehicles entering the venue; The twin model module is used to obtain the real-time status information of the three-dimensional parking lot; identify the vacant parking spaces based on the real-time status information of the three-dimensional parking lot, analyze the spatial distribution of the idle lifting platforms, and build a twin model of the three-dimensional parking lot; The deactivatable matching module is used to perform optimal deactivatable matching on the twin model of the three-dimensional parking lot according to the three-dimensional morphological structure model of the incoming vehicle, obtain the optimal deactivatable matching parking space, and make real-time parking management decisions to build a real-time parking management strategy; The vehicle use time prediction module is used to obtain the parking lot's historical operation logs; analyze the owner's historical vehicle use behavior in the parking lot's historical operation logs, and then predict the dynamic vehicle use time points, thereby generating the owner's vehicle use prediction time points; The pre-use vehicle scheduling module is used to analyze the optimal vehicle scheduling time and pre-use vehicle scheduling decisions for the three-dimensional parking lot twin model based on the predicted time point of the car owner's car use, and to build a pre-use vehicle scheduling strategy; The smart parking management module is used to conduct intelligent collaborative parking vehicle transfer learning for real-time parking management strategies and pre-use vehicle scheduling strategies, and to perform iterative reinforcement optimization to build a smart parking management model for three-dimensional parking lots.

Images