Parking lot parking space management method and system based on parking space detection model

Abstract

The invention provides a parking lot parking space management method and system based on a parking space detection model. The method comprises the following steps: constructing a parking space detection model; collecting parking lot image data, and inputting the parking lot image data to the parking space detection model for parking space analysis to obtain a parking space state; parking spaces in the parking lot are managed according to the parking space state, parking lot image data are acquired through a camera or a sensor, and the parking space state is analyzed through a parking space detection model; corresponding appropriate management is carried out based on the parking space state; the management efficiency of the parking lot is improved, and the use of parking resources is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of parking space management, and in particular to a parking space management method and system based on a parking space detection model. Background Art

[0002] At present, with the continuous increase in the number of cars, the management of parking spaces in parking lots has become increasingly complex. Therefore, effective management of parking spaces in parking lots can greatly improve the parking efficiency for users.

[0003] However, traditional parking lot management methods have many problems. It is difficult to accurately and in real time grasp the occupancy of parking spaces, resulting in difficulties for vehicles to find parking spaces, inefficient utilization of parking resources, and high management costs. Outdoor parking lots are especially susceptible to interference from weather and the surrounding environment, resulting in inaccurate detection of parking space usage, which greatly reduces the effectiveness of parking lot management.

[0004] Therefore, in order to overcome the above-mentioned defects, the present invention provides a parking lot parking space management method and system based on a parking space detection model. Summary of the Invention

[0005] The present invention provides a parking space management method and system for a parking lot based on a parking space detection model, which is used to obtain parking lot image data through a camera or sensor, and analyze the parking space status through the parking space detection model; perform appropriate management on the parking spaces based on the parking space status; improve the management efficiency of the parking lot, and optimize the use of parking resources.

[0006] A parking space management method based on a parking space detection model, comprising:

[0007] Build a parking space detection model;

[0008] Collect parking lot image data and input the parking lot image data into the parking space detection model to perform parking space analysis to obtain parking space status;

[0009] Manage parking spaces in the parking lot according to their status.

[0010] Preferably, a parking lot parking space management method based on a parking space detection model constructs a parking space detection model, including:

[0011] Extract different types of sample image sets from the preset database based on the parking environment characteristics, and add parking space information labels to each sample image;

[0012] The first detection model for each type of sample image set is determined by learning the parking space information labels of each sample image.

[0013] The first detection models under each type are integrated to obtain a second detection model, wherein the second detection model is a parking space detection model.

[0014] Preferably, a parking lot management method based on a parking space detection model is provided, wherein the parking lot data is parking lot image data acquired by a camera or a sensor, and the camera or the sensor is connected to a local network.

[0015] Preferably, a parking lot parking space management method based on a parking space detection model manages parking spaces in the parking lot according to the parking space status, including:

[0016] Real-time monitoring of vehicles entering the parking lot entrance and their attributes, and allocation of target available parking spaces to each entering vehicle based on vehicle attributes and the parking status of each parking space in the parking lot; based on the allocation results, each entering vehicle is dynamically guided to park in a single target available parking space to complete parking management.

[0017] Preferably, a parking lot parking space management method based on a parking space detection model, after collecting parking lot image data and before performing parking space analysis based on the parking space detection model, also needs to correct the parking lot image data according to the template image, stitching results and parking space detection model respectively.

[0018] A parking lot parking space management system based on a parking space detection model includes: a model building module, an analysis module and a parking space management module;

[0019] Among them, the model construction module is used to build a parking space detection model; the analysis module is used to collect parking lot image data and input the parking lot image data into the parking space detection model for parking space analysis to obtain the parking space status; the parking space management module is used to manage the parking spaces in the parking lot according to the parking space status.

[0020] Preferably, a parking lot parking space management system based on a parking space detection model, the model building module includes: a sample image extraction and labeling unit, a learning unit and a model determination unit; wherein, the sample image extraction and labeling unit extracts different types of sample image sets from a preset database according to the parking lot environment characteristics, and adds parking space information labels to each sample image; the learning unit learns different types of sample image sets and adds parking space information labels to each sample image to determine the first detection model under each type; the model determination unit synthesizes the first detection model under each type to obtain a second detection model, wherein the second detection model is a parking space detection model.

[0021] Preferably, a parking lot management system based on a parking space detection model, the parking lot data is parking lot image data acquired based on a camera or sensor, and the camera or sensor is connected to a local network.

[0022] Preferably, a parking lot parking space management system based on a parking space detection model, the parking space management module includes: a parking space allocation unit and a guidance unit, wherein the parking space allocation unit monitors the incoming vehicles and vehicle attributes at the parking lot entrance in real time, and allocates a target available parking space to each incoming vehicle according to the vehicle attributes and the parking space status of each parking space in the parking lot; the guidance unit dynamically guides each incoming vehicle to park in a single target available parking space based on the allocation result, thereby completing parking management.

[0023] Preferably, in a parking lot parking space management system based on a parking space detection model, before performing parking space analysis based on the parking space detection model, the analysis module also corrects the parking lot image data according to the template image, the splicing result and the parking space detection model.

[0024] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in this application document.

[0025] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0027] Figure 1 This is a flow chart of a parking space management method based on a parking space detection model in an embodiment of the present invention;

[0028] Figure 2 This is a structural diagram of a parking lot parking space management system based on a parking space detection model in an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0030] In one embodiment, the present invention provides a parking space management method based on a parking space detection model, such as Figure 1 As shown, including:

[0031] Step 1: Build a parking space detection model;

[0032] Constructing a parking space detection model includes: extracting different types of sample image sets from a preset database according to parking lot environmental characteristics, wherein the sample image sets are divided into training sample sets and verification sample sets, and the division standard is based on a preset ratio; adding parking space information labels to each sample image; training different types of corresponding training sample sets and parking space information labels of training sample images under each type, and determining a first detection model under each type; synthesizing the first detection models under each type to obtain a second detection model; wherein the parking lot is an outdoor parking lot, and the parking lot environmental characteristics are determined based on the lighting environment corresponding to the different external climate environments of the parking lot, including but not limited to external scene information under different lighting conditions such as sunny days, cloudy days, snowy days, rainy days, daytime, and nighttime; wherein the first detection model corresponds to the type one by one and includes multiple, the second detection model is one, and the second detection model is the parking space detection model, wherein the label information details the parking space boundary lines and occupancy status in the sample image.

[0033] The first detection models under each type are integrated. The specific process includes: inputting the verification sample set under each type into the corresponding first detection model for analysis, outputting the first output result of each first detection model, learning each first output result, and constructing a second detection model based on the learning result, wherein the learning of each first output result is the result of integrating multiple first output results based on linear regression.

[0034] Step 2: Collect parking lot image data and input the parking lot image data into the parking space detection model to perform parking space analysis to obtain parking space status;

[0035] Parking lot image data is acquired by a camera or sensor. Connect the camera or sensor to a local network, such as an Ethernet or Wi-Fi connection. To ensure that a computer can access the camera or sensor, configure the camera or sensor's IP address and subnet mask network parameters as follows:

[0036] Step 3: Manage parking spaces in the parking lot according to their status;

[0037] Real-time monitoring of vehicles entering the parking lot entrance and their attributes, and allocation of target available parking spaces to each entering vehicle based on vehicle attributes and the parking status of each parking space in the parking lot; based on the allocation results, each entering vehicle is dynamically guided to park in a single target available parking space to complete parking management.

[0038] In order to further improve the accuracy, adaptability and reliability of allocating target available parking spaces to incoming vehicles, it is proposed that: the real-time parking space status of each parking space in the parking lot is obtained, and the parking space number of each parking space in the parking lot is extracted, and the real-time parking space status is bound to the corresponding parking space number; based on the binding result, the real-time parking space status and parking space number of each parking space are visualized on the large screen at the entrance of the parking lot, and at the same time, during the visualization, available parking spaces and unavailable parking spaces are marked in a differentiated manner; the dynamics of vehicle entry into the parking lot entrance are continuously monitored, and when a vehicle entry is detected, the appearance of the entering vehicle is scanned and special identification is recognized, wherein the special identification recognition is set on the vehicle, including license plates or special marks; the vehicle attributes of the entering vehicle are determined based on the appearance scanning and special identification recognition results, wherein the vehicle attributes include temporary ordinary vehicles, fixed ordinary vehicles, etc. vehicles, vehicles for disabled persons and VIP vehicles; when the entering vehicle is a fixed ordinary vehicle, a fixed available parking space is automatically allocated to the fixed ordinary vehicle based on the background management terminal; otherwise, the entering vehicle is classified based on the vehicle attributes, and the optional parking space set for each category of entering vehicle is determined based on the classification result; based on the optional parking space set, the target available parking space selected by the vehicle user based on the touch management screen at the entrance of the parking lot is preferentially received, and when the vehicle user does not make an independent selection within a preset time period, the position of each optional parking space in the optional parking space set is traversed; based on the traversal result, the target available parking space is automatically allocated to the entering vehicle according to the proximity index, and the parking space blocking device of the corresponding parking space in the parking lot is activated based on the allocation result, wherein the activation of the parking space blocking device is to prevent vehicles that have entered the parking lot and have not parked according to the allocated target available parking space from parking at will.

[0039] In order to further provide parking guidance for vehicles entering the parking lot and ensure that each vehicle can quickly reach the corresponding parking space and will not be occupied by other vehicles at will, the following measures are proposed:

[0040] Based on the allocation results, the display screens at each intersection in the parking lot are controlled in a coordinated manner, and based on the coordinated control results, the direction of the target available parking space allocated to each incoming vehicle is dynamically displayed. The dynamic display guidance includes displaying the parking space number of the target available parking space on the display screen and simultaneously using arrows to indicate the direction of the parking space number;

[0041] When the incoming vehicle arrives at the target available parking space, the vehicle identity is verified based on the camera at the target available parking space, and the parking device is automatically closed after the vehicle identity verification is passed, completing the entry management of the vehicle in the parking lot.

[0042] The present invention provides a parking space management method for a parking lot based on a parking space detection model. The method acquires parking lot image data through a camera or sensor and analyzes the parking space status through the parking space detection model. The method manages the parking spaces accordingly based on the parking space status. The method improves the management efficiency of the parking lot and optimizes the use of parking resources.

[0043] In order to ensure that the parking space detection model can accurately analyze the parking lot image data, in one embodiment, after the parking lot image data is collected and before the parking space analysis is performed based on the parking space detection model, the parking lot image data needs to be corrected;

[0044] The specific steps for correcting parking lot image data are as follows: obtaining the position distribution status of cameras or sensors in the parking lot, and marking position parameters on the parking lot image data based on the position distribution status; retrieving a template image from a preset image database based on the position parameters, matching the template image with the corresponding parking lot image data, determining a correction coefficient for the parking lot image data based on the matching result, and performing a first image correction on the collected parking lot image data based on the correction coefficient to obtain first standard parking lot image data; splicing the parking lot image data based on the position distribution status, and performing a second image correction on the first standard parking lot image data after successful splicing to obtain second standard parking lot image data; obtaining a standard image format for image analysis by a parking space detection model, and performing a third image correction on the second standard parking lot image data based on the standard image format to obtain target standard parking lot image data; wherein the position parameters include the direction and distance of the parking space relative to the parking lot entrance and exit, and the relative position relationship between the parking space and the internal passage of the parking lot; the preset image database is pre-set and used to store template images corresponding to different position parameters; wherein the second image correction of the first standard parking lot image data after successful splicing includes: geometric correction, image noise reduction, etc.

[0045] Among them, the template image is matched with the corresponding parking lot image data, and the correction coefficient of the parking lot image data is determined according to the matching result; first, the template image and the corresponding parking lot image data are overlapped and mapped to determine the overlap degree between the template image and the corresponding parking lot image data; secondly, the correction coefficient is determined according to the overlap degree, that is, when the overlap degree is 100%, the correction coefficient of the parking lot image data is 0; when the overlap degree is less than 100%, the difference position between the parking lot image data and the template image is obtained, and the target pixel value of the template image and the real-time pixel value of the parking lot image data are determined according to the difference position, and the target difference between the target pixel value and the real-time pixel value is obtained, and the target difference is used as the correction coefficient of the parking lot image data; the accuracy and effectiveness of the determination of the difference coefficient are effectively guaranteed.

[0046] When splicing parking lot image data according to the location distribution status, the following steps are performed: the splicing order of the parking lot image data is determined according to a preset location sequence (for example, the sequential numbering of parking spaces in the parking lot), and vector labels are added to the parking lot image data according to the splicing order. At the same time, the parking lot image data is sorted according to the added vector labels to obtain a parking lot image data sequence, and the parking lot image data sequence is spliced.

[0047] In order to comprehensively and effectively determine the parking fees for each vehicle in the parking lot, and to bind the completion of billing and payment to the parking space status, thereby improving the management of parking spaces in the parking lot, a billing joint management model is proposed. In one embodiment, a parking space management method based on a parking space detection model further includes:

[0048] Build an online parking lot for parking spaces and launch the online parking lot based on the cloud server;

[0049] Based on the online results, users' parking reservation requests are received in real time, and the parking reservation requests are analyzed to determine the reservation start time point;

[0050] Prepare an available parking space for the online booking user within a preset time threshold before the scheduled start time, and bind the available parking space to the booking user's account at the scheduled start time;

[0051] Enter the preset starting time point based on the binding result, and perform online billing for the reserved available parking space according to the billing rules based on the entered result;

[0052] At the same time, the system accepts the billing method selected by offline parking users based on the touch screen at the entrance of the parking lot, where the billing methods include hourly billing, pay per time, daily billing, monthly billing and annual billing;

[0053] Call the billing rules based on the selected billing method, track the parking user's vehicle information based on the call results, and determine the corresponding real-time offline billing based on the information tracking results;

[0054] Synchronize online and offline billing updates, and guide vehicle users to pay when the vehicle leaves the parking lot;

[0055] After the vehicle user completes the fee payment, the status of the parking space used by the vehicle user is updated.

[0056] The present invention also provides a parking lot management system based on a parking space detection model, such as Figure 2 As shown, it includes: model building module 1, analysis module 2 and parking space management module 3;

[0057] Among them, the model construction module 1 is used to construct a parking space detection model; the analysis module 2 is used to collect parking lot image data and input the parking lot image data into the parking space detection model for parking space analysis to obtain the parking space status; the parking space management module 3 is used to manage the parking spaces in the parking lot according to the parking space status.

[0058] Among them, the model construction module includes: a sample image extraction and labeling unit, a learning unit and a model determination unit; wherein, the sample image extraction and labeling unit extracts different types of sample image sets from a preset database according to the parking lot environment characteristics, and adds parking space information labels to each sample image respectively; the learning unit learns different types of sample image sets and adds parking space information labels to each sample image to determine the first detection model under each type; the model determination unit synthesizes the first detection model under each type to obtain a second detection model, wherein the second detection model is a parking space detection model.

[0059] The parking lot data is parking lot image data acquired by a camera or sensor, and the camera or sensor is connected to a local network.

[0060] Among them, the parking space management module includes: a parking space allocation unit and a guidance unit. Among them, the parking space allocation unit monitors the vehicles entering the parking lot entrance and the vehicle attributes in real time, and allocates a target available parking space to each entering vehicle according to the vehicle attributes and the parking space status of each parking space in the parking lot; the guidance unit dynamically guides each entering vehicle to park in a single target available parking space based on the allocation results to complete parking management.

[0061] Before performing parking space analysis based on the parking space detection model, the analysis module also corrects the parking lot image data according to the template image, the splicing result and the parking space detection model.

[0062] Constructing a parking space detection model includes: extracting different types of sample image sets from a preset database according to parking lot environmental characteristics, wherein the sample image sets are divided into training sample sets and verification sample sets, and the division standard is based on a preset ratio; adding parking space information labels to each sample image; training different types of corresponding training sample sets and parking space information labels of training sample images under each type, and determining a first detection model under each type; synthesizing the first detection models under each type to obtain a second detection model; wherein the parking lot is an outdoor parking lot, and the parking lot environmental characteristics are determined based on the lighting environment corresponding to the different external climate environments of the parking lot, including but not limited to external scene information under different lighting conditions such as sunny days, cloudy days, snowy days, rainy days, daytime, and nighttime; wherein the first detection model corresponds to the type one by one and includes multiple, the second detection model is one, and the second detection model is the parking space detection model, wherein the label information details the parking space boundary lines and occupancy status in the sample image.

[0063] The first detection models under each type are integrated. The specific process includes: inputting the verification sample set under each type into the corresponding first detection model for analysis, outputting the first output result of each first detection model, learning each first output result, and constructing a second detection model based on the learning result, wherein the learning of each first output result is the result of integrating multiple first output results based on linear regression.

[0064] Parking lot image data is acquired by a camera or sensor. Connect the camera or sensor to a local network, such as an Ethernet or Wi-Fi connection. To ensure that a computer can access the camera or sensor, configure the camera or sensor's IP address and subnet mask network parameters as follows:

[0065] In order to further improve the accuracy, adaptability and reliability of allocating target available parking spaces to incoming vehicles, the parking space allocation unit includes: obtaining the real-time parking space status of each parking space in the parking lot, extracting the parking space number of each parking space in the parking lot, and binding the real-time parking space status with the corresponding parking space number; visually displaying the real-time parking space status and parking space number of each parking space on the large screen at the entrance of the parking lot based on the binding result, and at the same time, when visually displaying, marking the available parking spaces and unavailable parking spaces with differentiated markings; continuously monitoring the dynamics of vehicle entry at the parking lot entrance, and when detecting the presence of a vehicle entering, performing an appearance scan and special identification recognition on the incoming vehicle, wherein the special identification recognition is set on the vehicle, including license plates or special marks; determining the vehicle attributes of the incoming vehicle based on the appearance scan and special identification recognition results, wherein the vehicle attributes include temporary ordinary vehicles, fixed Ordinary vehicles, vehicles for disabled persons and VIP vehicles; when the entering vehicle is a fixed ordinary vehicle, a fixed available parking space is automatically allocated to the fixed ordinary vehicle based on the background management terminal; otherwise, the entering vehicle is classified based on the vehicle attributes, and the set of optional parking spaces for each category of entering vehicles is determined based on the classification results; based on the set of optional parking spaces, the target available parking spaces selected by the vehicle user based on the touch-sensitive management screen at the entrance of the parking lot are given priority, and when the vehicle user does not make an independent selection within a preset time period, the positions of each optional parking space in the set of optional parking spaces are traversed; based on the traversal results, the target available parking space is automatically allocated to the entering vehicle according to the proximity index, and the parking space blocking device of the corresponding parking space in the parking lot is activated based on the allocation result, wherein the activation of the parking space blocking device is to prevent vehicles that have entered the parking lot and have not parked according to the allocated target available parking space from parking at will.

[0066] To further guide vehicles entering the parking lot and ensure that each vehicle can quickly reach the corresponding parking space and is not occupied by other vehicles, the guidance unit includes:

[0067] Based on the allocation results, the display screens at each intersection in the parking lot are controlled in a coordinated manner, and based on the coordinated control results, the direction of the target available parking space allocated to each incoming vehicle is dynamically displayed. The dynamic display guidance includes displaying the parking space number of the target available parking space on the display screen and simultaneously using arrows to indicate the direction of the parking space number;

[0068] When the incoming vehicle arrives at the target available parking space, the vehicle identity is verified based on the camera at the target available parking space, and the parking device is automatically closed after the vehicle identity verification is passed, completing the entry management of the vehicle in the parking lot.

[0069] Capture parking lot image data through cameras or sensors, and analyze parking space status through parking space detection models; manage parking spaces accordingly based on their status; improve parking lot management efficiency and optimize the use of parking resources.

[0070] To ensure that the parking space detection model can accurately analyze the parking lot image data, in one embodiment, after collecting the parking lot image data and before performing parking space analysis based on the parking space detection model, the analysis module also needs to correct the parking lot image data;

[0071] The specific steps for correcting parking lot image data are as follows: obtaining the position distribution status of cameras or sensors in the parking lot, and marking position parameters on the parking lot image data based on the position distribution status; retrieving a template image from a preset image database based on the position parameters, matching the template image with the corresponding parking lot image data, determining a correction coefficient for the parking lot image data based on the matching result, and performing a first image correction on the collected parking lot image data based on the correction coefficient to obtain first standard parking lot image data; splicing the parking lot image data based on the position distribution status, and performing a second image correction on the first standard parking lot image data after successful splicing to obtain second standard parking lot image data; obtaining a standard image format for image analysis by a parking space detection model, and performing a third image correction on the second standard parking lot image data based on the standard image format to obtain target standard parking lot image data; wherein the position parameters include the direction and distance of the parking space relative to the parking lot entrance and exit, and the relative position relationship between the parking space and the internal passage of the parking lot; the preset image database is pre-set and used to store template images corresponding to different position parameters; wherein the second image correction of the first standard parking lot image data after successful splicing includes: geometric correction, image noise reduction, etc.

[0072] Among them, the template image is matched with the corresponding parking lot image data, and the correction coefficient of the parking lot image data is determined according to the matching result; first, the template image and the corresponding parking lot image data are overlapped and mapped to determine the overlap degree between the template image and the corresponding parking lot image data; secondly, the correction coefficient is determined according to the overlap degree, that is, when the overlap degree is 100%, the correction coefficient of the parking lot image data is 0; when the overlap degree is less than 100%, the difference position between the parking lot image data and the template image is obtained, and the target pixel value of the template image and the real-time pixel value of the parking lot image data are determined according to the difference position, and the target difference between the target pixel value and the real-time pixel value is obtained, and the target difference is used as the correction coefficient of the parking lot image data; the accuracy and effectiveness of the determination of the difference coefficient are effectively guaranteed.

[0073] When splicing parking lot image data according to the location distribution status, the following steps are performed: the splicing order of the parking lot image data is determined according to a preset location sequence (for example, the sequential numbering of parking spaces in the parking lot), and vector labels are added to the parking lot image data according to the splicing order. At the same time, the parking lot image data is sorted according to the added vector labels to obtain a parking lot image data sequence, and the parking lot image data sequence is spliced.

[0074] In order to improve the comprehensive and effective determination of parking fees for each vehicle in the parking lot, and at the same time, bind the completion of billing and payment to the parking space status, thereby improving the management effect of parking spaces in the parking lot, a billing joint management model is proposed. In one embodiment, it also includes: a billing management module:

[0075] The billing management process performs the following operations:

[0076] Build an online parking lot for parking spaces and launch the online parking lot based on the cloud server;

[0077] Based on the online results, users' parking reservation requests are received in real time, and the parking reservation requests are analyzed to determine the reservation start time point;

[0078] Prepare an available parking space for the online booking user within a preset time threshold before the scheduled start time, and bind the available parking space to the booking user's account at the scheduled start time;

[0079] Enter the preset starting time point based on the binding result, and perform online billing for the reserved available parking space according to the billing rules based on the entered result;

[0080] At the same time, the system accepts the billing method selected by offline parking users based on the touch screen at the entrance of the parking lot, where the billing methods include hourly billing, pay per time, daily billing, monthly billing and annual billing;

[0081] Call the billing rules based on the selected billing method, track the parking user's vehicle information based on the call results, and determine the corresponding real-time offline billing based on the information tracking results;

[0082] Synchronize online and offline billing updates, and guide vehicle users to pay when the vehicle leaves the parking lot;

[0083] After the vehicle user completes the fee payment, the status of the parking space used by the vehicle user is updated.

[0084] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A parking space management method based on a parking space detection model, characterized in that: include: Build a parking space detection model; Collect parking lot image data and input the parking lot image data into the parking space detection model to perform parking space analysis to obtain parking space status; Manage parking spaces in the parking lot according to their status.

2. The parking lot management method based on the parking space detection model according to claim 1 is characterized in that: Build a parking space detection model, including: Extract different types of sample image sets from the preset database based on the parking environment characteristics, and add parking space information labels to each sample image; The first detection model for each type of sample image set is determined by learning the parking space information labels of each sample image. The first detection models under each type are integrated to obtain a second detection model, wherein the second detection model is a parking space detection model.

3. The parking lot management method based on the parking space detection model according to claim 1 is characterized in that: Parking lot data is based on parking lot image data obtained by cameras or sensors, which are connected to the local network.

4. The parking lot management method based on the parking space detection model according to claim 1 is characterized in that: Manage parking spaces in the parking lot according to their status, including: Real-time monitoring of vehicles entering the parking lot entrance and their attributes, and allocation of target available parking spaces to each entering vehicle based on vehicle attributes and the parking status of each parking space in the parking lot; based on the allocation results, each entering vehicle is dynamically guided to park in a single target available parking space to complete parking management.

5. The parking lot management method based on the parking space detection model according to claim 1 is characterized in that: After collecting parking lot image data and before performing parking space analysis based on the parking space detection model, the parking lot image data needs to be corrected based on the template image, the stitching result, and the parking space detection model.

6. A parking lot management system based on a parking space detection model, characterized in that: include: Model building module, analysis module and parking space management module; Among them, the model construction module is used to build a parking space detection model; the analysis module is used to collect parking lot image data and input the parking lot image data into the parking space detection model for parking space analysis to obtain the parking space status; the parking space management module is used to manage the parking spaces in the parking lot according to the parking space status.

7. The parking lot management system based on the parking space detection model according to claim 6 is characterized in that: The model construction module includes: a sample image extraction and labeling unit, a learning unit and a model determination unit; wherein, the sample image extraction and labeling unit extracts different types of sample image sets from a preset database according to the parking lot environment characteristics, and adds parking space information labels to each sample image respectively; the learning unit learns different types of sample image sets and adds parking space information labels to each sample image to determine the first detection model under each type; the model determination unit synthesizes the first detection model under each type to obtain a second detection model, wherein the second detection model is a parking space detection model.

8. The parking lot management system based on the parking space detection model according to claim 6 is characterized in that: Parking lot data is based on parking lot image data obtained by cameras or sensors, which are connected to the local network.

9. The parking lot management system based on the parking space detection model according to claim 6, characterized in that: The parking space management module includes: a parking space allocation unit and a guidance unit. The parking space allocation unit monitors the vehicles entering the parking lot entrance and their attributes in real time, and allocates a target available parking space to each entering vehicle based on the vehicle attributes and the parking space status of each parking space in the parking lot; the guidance unit dynamically guides each entering vehicle to park in a single target available parking space based on the allocation results to complete parking management.

10. The parking lot management system based on the parking space detection model according to claim 6, characterized in that: Before performing parking space analysis based on the parking space detection model, the analysis module also corrects the parking lot image data according to the template image, the stitching result and the parking space detection model.

Images