Parking lot parking space management system and management method based on machine vision

By using machine vision technology and deep learning models in the parking lot management system, combined with image processing and Yolov8 algorithm, the problems of insufficient license plate recognition accuracy and untimely update of information in complex environments are solved, and efficient parking space recognition and real-time information feedback are achieved , improves the user experience.

CN118942273AInactive Publication Date: 2024-11-12YANCHENG INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410941264.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-15
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing parking lot management system has insufficient license plate recognition accuracy in complex environments, resulting in incorrect billing or failure to identify, and the vehicle guidance process after entry is incomplete, resulting in untimely update of parking lot information, and users are wasting time looking for parking spaces.

Method used

Adopt a machine vision-based parking space management system, combining high-resolution in-site cameras, entrance cameras, MySQL databases and deep learning models, identify license plates and detect parking spaces through image processing and deep learning algorithms (such as Yolov8 and the improved SEAttention module) Occupancy status, update parking lot information in real time, and provide personalized parking suggestions.

Benefits of technology

It improves the accuracy of license plate recognition, updates parking lot information in real time, reduces the waste of time users when looking for parking spaces, and improves user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118942273A_ABST
    Figure CN118942273A_ABST
Patent Text Reader

Abstract

The invention provides a parking lot parking space management system and management method based on machine vision, and belongs to the field of vehicle traffic management methods. The system comprises a high-resolution in-field camera used for collecting real-time images of a parking lot; the entrance camera is used for recording user information; the administrator PC end is controlled; a user is at a mobile APP program end; the management method comprises the following steps: acquiring image information of a parking lot through a high-resolution intra-field camera; detecting a parking space occupation state by using an image processing and deep learning model; real-time parking space information and traffic conditions are obtained according to the obtained image information and the parking space occupancy state, and the optimal parking lot selection and navigation path are provided for a user according to the real-time parking space information and the traffic conditions; based on historical parking preferences and habits of the user, the system can provide personalized parking suggestions, and the user satisfaction is improved. Image processing and a depth model are combined, and the problem that license plate recognition is affected by insufficient light and the like is prevented. The depth model can detect and identify multiple targets at the same time, and the identification efficiency and accuracy of the system are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention provides a parking lot parking space management system and a management method based on machine vision, belonging to the field of vehicle traffic management methods. Background Art

[0002] The existing parking lot management system has insufficient accuracy in collecting and identifying information about vehicles entering the lot.

[0003] Currently, license plate recognition technology has a low recognition accuracy in complex environments such as bad weather, low light or strong reflections, which may result in incorrect billing or failure to recognize the vehicle.

[0004] The vehicle guidance process after entering the parking lot is still imperfect. The delay in information feedback will lead to untimely updating of parking lot information (such as the number of remaining parking spaces). Users can only find empty spaces to park their vehicles on their own, which may waste a lot of time. Summary of the invention

[0005] In view of the above problems, the present invention provides a parking lot parking space management system and a management method based on machine vision.

[0006] The present invention adopts the following technical solution:

[0007] The parking lot management system based on machine vision of the present invention comprises a high-resolution on-site camera for collecting real-time images of the parking lot; an entrance camera for recording user information; a control administrator PC; and a user mobile APP program terminal;

[0008] MySQL database for recording parking lot status, user information, and image data;

[0009] The camera at the parking lot entrance is used to read the vehicle license plate and transmit the information to the MySQL database;

[0010] Control the interaction between the administrator's PC and the APP program;

[0011] The parking spaces in the parking lot are divided into different areas, and the MySQL database is used to record user information and parking space status;

[0012] The cameras in the venue are used to detect whether the parking space is vacant and send the information to the database; when the parking space is vacant, the APP program reminds the user of the specific location of the parking space.

[0013] The management method of the parking lot management system based on machine vision of the present invention comprises the following steps:

[0014] S1, obtain the image information of the parking lot through the high-resolution on-site camera;

[0015] S2, using image processing and deep learning models to detect parking space occupancy status;

[0016] S3, obtaining real-time parking space information and traffic conditions based on the image information and parking space occupancy status obtained in step S1 and step S2, and providing the user with the best parking lot selection and navigation path using the real-time parking space information and traffic conditions; the traffic conditions refer to: whether the parking spaces in the parking lot are vacant, and whether a certain area is congested;

[0017] S4. Based on the user's historical parking preferences and habits, the system can provide personalized parking suggestions to improve user satisfaction.

[0018] Personalized parking suggestions: refers to giving priority to personalized solutions when guiding customers to park based on their previous parking habits, such as: liking to park in a certain area, or liking to park close to the entrance.

[0019] In the management method of the parking lot management system based on machine vision of the present invention, the image processing method in step S2 is:

[0020] When detecting license plates at the entrance of parking lots, image segmentation based on HSV color space is used to extract license plates by combining binary morphology, connectivity analysis, backbone extraction, Hough line detection and contour detection.

[0021] The deep learning model adopts Yolov8 algorithm, and the improved SEAttention module is added to Yolov8; the SE module has the advantages of strong versatility and light weight. In the management method of the parking lot management system based on machine vision described in the present invention, the traffic condition in step S3 refers to: whether the parking spaces in the parking lot are vacant, and whether a certain area in the parking lot is congested.

[0022] The module can introduce relatively small parameters and computational complexity, which does not significantly increase the model complexity and computational cost. This makes it very practical in practical applications, especially in resource-constrained scenarios such as parking systems. In addition, the module can be easily embedded into existing convolutional neural network architectures without large-scale modifications to the network architecture. Its design is simple and has low computational overhead.

[0023] However, the SEAttention module only models the relationship between channels and ignores the relationship in the spatial dimension. This means that it may not work well when dealing with certain tasks that require consideration of spatial context. Therefore, in addition to the functions of squeezing and excitation, the improved SEAttention module also adds multi-scale feature aggregation, channel shuffling, layer normalization, and 1x1 convolution to adjust the number of channels.

[0024] Further improve the performance and applicability of the SEAttention module, thereby improving the performance in the visual task of parking space recognition.

[0025] Beneficial Effects

[0026] The parking lot management system and management method based on machine vision provided by the present invention have the following effects:

[0027] 1. Improved recognition accuracy

[0028] The combination of image processing and deep models is used to prevent problems such as insufficient light from affecting license plate recognition. The deep model can detect and identify multiple targets at the same time, improving the recognition efficiency and accuracy of the system.

[0029] 2. User Experience

[0030] With real-time information feedback, parking lot information such as the number of remaining parking spaces and vehicle entry and exit records are updated in a timely manner to prevent users from wasting a lot of time looking for parking spaces. Mobile apps can also be used to guide customers to the correct location to prevent them from getting lost. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a SEAttention module diagram of the management method of the parking lot parking space management system based on machine vision of the present invention;

[0032] Figure 2 It is a SEAttention module diagram after the improvement of the management method of the parking lot parking space management system based on machine vision of the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose and technical solution of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] The parking lot management system based on machine vision described in the present invention includes a high-resolution on-site camera for collecting real-time images of the parking lot; an entrance camera for recording user information; an APP or WeChat applet and a PC terminal, where users register and log in on their mobile phones, and administrators interact with users using the PC terminal. The parking lots are divided into areas, and the MySQL database is used to record user information and parking lot status;

[0035] Cameras at the entrances to parking lots are used to read vehicle license plates and transmit the information to a database.

[0036] Multiple high-resolution cameras in the parking lot are used to detect whether a parking space is free and send the information to the database. When a parking space is free, the APP reminds the user of the specific location of the parking space.

[0037] The management method of the parking lot management system based on machine vision is as follows:

[0038] S1, obtain the image information of the parking lot through a high-resolution camera;

[0039] S2, using image processing and deep learning models to detect parking space occupancy status;

[0040] Image processing: When detecting license plates at the entrance of parking lots, image segmentation based on HSV color space is used to extract license plates by combining binary morphology, connectivity analysis, backbone extraction, Hough line detection and contour detection.

[0041] The deep learning model uses the Yolov8 algorithm, and adds the improved SEAttention module to Yolov8. The SE module has the advantages of strong versatility and light weight. The parameters and computational complexity that can be introduced by this module are relatively small, and will not significantly increase the complexity and computational cost of the model. This makes it very practical in practical applications, especially in resource-constrained scenarios such as parking systems. In addition, the module is easily embedded in the existing convolutional neural network architecture without large-scale modifications to the network architecture. It has a simple design and low computational overhead. However, the SEAttention module only models the relationship between channels, but ignores the relationship in the spatial dimension. Its module diagram is shown in the figure below. Figure 1 This means that it may not work well on certain tasks that require considering spatial context.

[0042] Therefore, in addition to the functions of squeezing and excitation, the improved SEAttention module also adds multi-scale feature aggregation, channel shuffling, layer normalization, and 1x1 convolution to adjust the number of channels. Figure 2 The performance and applicability of the SEAttention module are further improved, thereby improving the performance in the visual task of parking space recognition.

[0043] S3. Use real-time parking space information and traffic conditions to provide drivers with the best parking lot selection and navigation route; traffic conditions refer to whether there are vacant parking spaces in the parking lot and whether a certain area is congested.

[0044] S4. Based on the user's historical parking preferences and habits, the system can provide personalized parking suggestions to improve user satisfaction.

[0045] Personalized parking suggestions: refers to giving priority to personalized solutions when guiding customers to park based on their previous parking habits, such as: liking to park in a certain area, or liking to park close to the entrance.

[0046] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A parking lot management system based on machine vision, which includes a high-resolution on-site camera for collecting real-time images of the parking lot; an entrance camera for recording user information; and a control administrator PC; The user is on the mobile APP program terminal; MySQL database for recording parking lot status, user information, and image data; The camera at the parking lot entrance is used to read the vehicle license plate and transmit the information to the MySQL database; Control the interaction between the administrator's PC and the APP program; The parking spaces in the parking lot are divided into different areas, and the MySQL database is used to record user information and parking space status; The cameras in the venue are used to detect whether the parking space is vacant and send the information to the database; when the parking space is vacant, the APP program reminds the user of the specific location of the parking space.

2. A management method for a parking lot management system based on machine vision, characterized in that: Here are the steps: S1, obtain the image information of the parking lot through the high-resolution on-site camera; S2, using image processing and deep learning models to detect parking space occupancy status; S3, obtaining real-time parking space information and traffic conditions based on the image information and parking space occupancy status obtained in step S1 and step S2, and providing the user with the best parking lot selection and navigation path using the real-time parking space information and traffic conditions; S4. Based on the user's historical parking preferences and habits, the system can provide personalized parking suggestions to improve user satisfaction.

3. The management method of the parking lot management system based on machine vision according to claim 2 is characterized by: The image processing method in step S2 is: When detecting license plates at the entrance of parking lots, image segmentation based on HSV color space is used to extract license plates by combining binary morphology, connectivity analysis, backbone extraction, Hough line detection and contour detection. The deep learning model adopts the Yolov8 algorithm, and adds an improved SEAttention module to Yolov8.

4. The management method of the parking lot management system based on machine vision according to claim 2 is characterized by: The traffic condition in step S3 refers to whether parking spaces in the parking lot are vacant and whether a certain area in the parking lot is congested.