Intelligent parking system

Abstract

The invention discloses an intelligent parking system which comprises a parking lot center control module, a parking lot entrance and exit terminal module and a geomagnetic sensor module. Wherein in the parking lot center control module, a first information communication module is used for performing signal transmission with a geomagnetic sensor module, and a microprocessor module is used for processing and optimizing a signal received by the parking lot center control module; communication of the geomagnetic sensor module is sent to the parking lot center control module through the edge server, the parking lot center control module processes the information and uploads the information to the cloud server through the local server, and the cloud server issues specific information of a garage to the man-machine interaction system of the parking lot entrance and exit terminal module to achieve interaction. The intelligent parking system provided by the invention can quickly and accurately identify license plate information in a complex scene, provides reliable data support, can accurately track and manage the use condition of each parking space, and effectively avoids resource waste and parking congestion conditions.

Description

Technical Field

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

[0002] With the advancement of science and technology, China's automobile market has developed by leaps and bounds, and the number of cars owned by Chinese residents has continued to increase. However, with the continuous growth of my country's car ownership, problems such as insufficient construction of parking spaces have arisen, resulting in a large parking gap. However, traditional parking methods can no longer meet the rapidly growing demand for urban cars. The method of blindly searching for parking spaces with the naked eye greatly reduces parking efficiency, resulting in a large amount of parking time and even serious traffic congestion, affecting the driver's mood and increasing the probability of traffic accidents. The existing urban-level smart parking systems have various design types, but have not brought substantial help to the parking shortage. Most of the existing parking systems are based on a single sensor, with weak system functions and lack of differentiation and processing between new energy vehicles and oil vehicles. Manual charging and manual inspections are prone to human errors and have high management costs. At the same time, the mobility of vehicles in the parking lot is large, and it is difficult for managers to grasp real-time information in a timely manner, resulting in low utilization of the parking lot. Therefore, the existing parking system has limitations in parking efficiency and convenience, which is also a problem that the smart parking system needs to solve urgently. Summary of the invention

[0003] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is that the parking and car-finding speeds in the existing parking system are slow and inconvenient. The present invention provides a smart parking system that can quickly and accurately identify license plate information in complex scenarios, providing reliable data support for subsequent vehicle management and system operation, and can accurately track and manage the use of each parking space in conjunction with the geomagnetic sensor module, effectively avoiding waste of resources and parking congestion.

[0004] To achieve the above-mentioned objectives, the present invention provides a smart parking system, including a parking lot central control module, a parking lot entrance and exit terminal module and a geomagnetic sensor module; wherein the parking lot central control module includes a first information communication module and a microprocessor module, the first information communication module is used to transmit signals with the geomagnetic sensor module, and the microprocessor module is used to process and optimize the signals received by the parking lot central control module; the communication of the geomagnetic sensor module is sent by the edge server to the parking lot central control module, which is connected to the local server after processing, and then uploaded to the cloud server by the local server, and the cloud server sends the specific information of the garage to the human-computer interaction system of the parking lot entrance and exit terminal module to realize interaction with the user.

[0005] Furthermore, the geomagnetic sensor module includes a geomagnetic sensor, a lora wireless communication module, a battery module and an edge server; the geomagnetic sensor is used to detect geomagnetic changes above the parking space to determine whether there is a vehicle parked. A parking garage will be divided into several areas according to the size of the garage, and the information of all geomagnetic sensor modules in an area will be sent to the edge server in the area through the lora wireless communication module; the battery module is used to power the RM3100 geomagnetic sensor and the lora wireless communication module.

[0006] Furthermore, the parking lot entrance and exit terminal module includes a vehicle visual recognition detection module and a second information communication module. When a vehicle enters the parking lot, the parking lot entrance and exit terminal module obtains the vehicle's valid information through the vehicle visual recognition detection module, and transmits the information to the parking lot central control module through the second information communication module.

[0007] Furthermore, the parking lot entrance and exit terminal module also includes a human-computer interaction module. When the driver needs to find a vehicle, he or she inputs the license plate information or parking space information through the human-computer interaction module, and the parking lot central control module will locate the vehicle position and provide a vehicle search route.

[0008] Furthermore, the vehicle visual recognition detection module first improves the image quality under different conditions through an image enhancement algorithm, and then uses the YOLOv8 model to accurately locate the license plate position to obtain the license plate information.

[0009] Furthermore, the vehicle visual recognition detection module adopts an adaptive correction algorithm for uneven illumination images based on a two-dimensional gamma function, and removes the illumination component in the image by utilizing a multi-scale Gaussian function, thereby eliminating the influence of uneven illumination on image quality.

[0010] Furthermore, the vehicle visual recognition detection module removes the illumination component in the image, specifically comprising the following steps:

[0011] Firstly, the multi-scale Gaussian function is used to process the image and separate the illumination characteristics in the image;

[0012] Then, based on the extracted illumination characteristics, the parameters of the two-dimensional gamma function were adjusted accordingly to achieve the recognition of the target area under different illumination conditions.

[0013] Furthermore, the illumination component is represented by G being a Gaussian function, F being the original image, and I being the illumination component.

[0014]

[0015] I(x,y)=F(x,y)G(x,y) (5)

[0016]

[0017] Wherein, γ is an exponential value of brightness enhancement, and the two-dimensional gamma function O(x, y) can be determined by the illumination component and the brightness mean m of the illumination component.

[0018] Another embodiment of the present invention provides a parking method of a smart parking system, comprising the following steps:

[0019] Step 1: When a vehicle enters the parking lot entrance, the parking lot entrance and exit terminal module starts the visual recognition process;

[0020] Step 2: The parking lot central control module uses the geomagnetic sensor module to detect the vacancy status of parking spaces based on the received vehicle information, allocates a suitable parking space for the vehicle, and plans a path from the entrance to the designated parking space, which is then fed back to the human-computer interaction system of the parking lot entrance and exit terminal module;

[0021] Step 3: The human-computer interaction system of the parking lot entrance and exit terminal module conveys the navigation information to the driver, and the driver quickly finds the parking space and parks the vehicle according to the navigation route.

[0022] Step 4: Once the vehicle enters and stops at the designated parking space, the geomagnetic sensor responds immediately and starts the built-in parking timer.

[0023] Another embodiment of the present invention provides a method for finding a car in a smart parking system, comprising the following steps:

[0024] Step 1: The driver inputs the license plate information or parking space information through the human-computer interaction module of the parking lot entrance and exit terminal module, and sends a car search request to the parking lot central control module;

[0025] Step 2: After receiving the request to find the vehicle, the parking lot central control module uses the vehicle visual recognition detection module and the geomagnetic sensor module of the parking lot entrance and exit terminal module to determine the location of the vehicle, plan a route to find the vehicle, and feed it back to the human-computer interaction system of the parking lot entrance and exit terminal module;

[0026] Step 3: The human-computer interaction system of the parking lot entrance and exit terminal module conveys the car-finding route information to the driver, and the driver quickly finds the vehicle according to the navigation route and leaves the parking space;

[0027] Step 4: After the vehicle leaves the parking space, the geomagnetic sensor module detects that the parking space becomes idle, and updates the parking space status information of the parking lot central control module through the first information communication module;

[0028] Step 5: The driver can check the detailed parking time through the parking payment function of the mini program. After confirmation, he / she can complete the online payment according to the garage charging standards.

[0029] Technical Effects

[0030] The present invention provides a smart parking system which utilizes geomagnetic sensors and visual recognition technology to accurately identify and classify license plate information, thereby improving detection accuracy and the ability to resist environmental interference.

[0031] The present invention provides a smart parking system that globally manages and maintains parking space information through a parking lot central control module, thereby improving the utilization rate of parking spaces. The parking lot central control module compares the value uploaded by the geomagnetic sensor with the set threshold value to determine whether there is a vehicle parked above the parking space, and finally outputs 0 or 1 to represent the parking situation of the parking space. 0 represents no car, and 1 represents a car. If a vehicle enters the parking lot, the parking lot entrance and exit terminal module identifies the vehicle license plate and type and uploads it to the parking lot central control module. The specific parking area and parking space of the vehicle can be accurately located through the visual recognition module and geomagnetic sensor module in the area, thereby forming a specific information set for the parking space. In this way, the specific parking space information of each parking space in each area can be obtained, which is generally convenient for subsequent management and calling.

[0032] The concept, specific structure and technical effects of the present invention will be further described below in conjunction with the accompanying drawings to fully understand the purpose, characteristics and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a schematic diagram of a smart parking system according to a preferred embodiment of the present invention;

[0034] Figure 2 It is a functional schematic diagram of a user interface design of a human-computer interaction module of a smart parking system according to a preferred embodiment of the present invention;

[0035] Figure 3 It is a flow chart of a geomagnetic sensor module and a license plate recognition module of a smart parking system according to a preferred embodiment of the present invention;

[0036] Figure 4 It is a schematic diagram of the implementation flow of an algorithm in a license plate recognition module of a smart parking system according to a preferred embodiment of the present invention;

[0037] Figure 5 It is a schematic diagram of a YOLOv8 algorithm fused with a C2f module of a smart parking system according to a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0039] In the following description, specific details such as specific internal procedures and techniques are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0040] like Figure 1 As shown, an embodiment of the present invention provides a smart parking system, which is a smart parking system based on geomagnetism and visual sensing, including a parking lot central control module, a parking lot entrance and exit terminal module, a geomagnetic sensor module, a visual recognition module and a human-computer interaction module.

[0041] Among them, the geomagnetic sensor module mainly includes RM3100 geomagnetic sensor, lora wireless communication module and battery module. RM3100 geomagnetic sensor is used to detect the geomagnetic changes above the parking space to determine whether there is a vehicle parked. A parking garage will be divided into several areas according to the size of the garage, and the information of all geomagnetic sensor modules in an area will be sent to the edge server in the area through the lora wireless communication module; the battery module is used to power the RM3100 geomagnetic sensor and lora wireless communication module of the geomagnetic sensor module. The geomagnetic sensor module completes the collection of parking space information and is the core module for realizing each parking space as a unit. The geomagnetic sensor adopts the independently developed high-precision RM3100 geomagnetic sensor, which is installed in parallel on the ground under the parking space. It can effectively detect the presence and parking time of the vehicle, reduce false detection and missed detection, and improve the reliability of the system.

[0042] In this embodiment, the geomagnetic sensor module is independently developed and can operate accurately and reliably, ensuring the high precision and stability of the entire system. Based on wireless transmission, real-time data transmission can be achieved; it is moisture-proof and waterproof, has strong signal transmission capability, is not affected by extreme environments and weather, and has excellent anti-interference ability; it can distinguish the size of vehicles according to the size of geomagnetic signals. The vehicle license plate is identified through the visual module and the vehicle type is distinguished according to the license plate color, thereby realizing real-time detection of vehicles and parking spaces.

[0043] The parking lot central control module processes and optimizes the signals received and sent by the geomagnetic sensor module and the vehicle visual recognition detection module.

[0044] The parking lot central control module includes a first information communication module and an STM32F103C6 microprocessor module. The first information communication module is used to transmit signals with the edge server, and the microprocessor module is used to process and optimize the signals received by the parking lot central control module. The information of the geomagnetic sensor module is sent by the edge server to the parking lot central control module, which processes the information and connects to the local server, and then uploads it to the cloud server. The cloud server sends the specific information of the garage to the human-computer interaction system to realize interaction with users.

[0045] The parking lot entrance and exit terminal module includes a vehicle visual recognition detection module and a second information communication module. When a vehicle enters the parking lot, the parking lot entrance and exit terminal module obtains the vehicle's valid information through the vehicle visual recognition detection module, and transmits the information to the parking lot central control module through the second information communication module. When the driver needs to find the vehicle, he inputs the license plate information or parking space information through the human-computer interaction module, and the parking lot central control module will locate the vehicle and provide a route to find the vehicle.

[0046] In an embodiment of the present invention, the system displays the status of parking spaces in the parking lot in real time through a mini-program, provides parking space reservation, parking navigation, car search navigation and online payment functions, and significantly improves the driver's parking experience and satisfaction.

[0047] Among them, in the embodiment of the present invention, the vehicle visual recognition detection module first improves the image quality under different conditions through the image enhancement algorithm, and then uses the YOLOv8 model to accurately locate the license plate position, thereby obtaining the license plate information. Specifically, firstly, the license plate recognition problem caused by uneven lighting at night is solved through preprocessing, and the image quality is optimized by using the image enhancement algorithm. Then, the YOLOv8 model is combined with the GAM attention mechanism to accurately locate the license plate, thereby improving the recognition speed and accuracy. Finally, the license plate information is parsed through traditional character recognition technology, and the results are sent to the parking lot central control module via the communication module to complete an efficient license plate recognition process.

[0048] The dataset of the embodiment of the present invention is derived from China-Balanced-License-Plate-Recognition-Dataset-330k. This is a high-quality, balanced Chinese license plate recognition dataset, which contains 330,000 images of various Chinese license plates. The dataset has been carefully designed to ensure excellent image quality and a balanced distribution of most types of license plates, making it suitable for training and evaluating license plate recognition models. Since images in dark environments can cause errors in the detection results, the collected images need to be normalized to eliminate differences.

[0049] Specifically, this paper proposes an adaptive correction algorithm for uneven illumination images based on a two-dimensional gamma function. By utilizing a multi-scale Gaussian function, the illumination component in the image is effectively removed, thereby eliminating the influence of uneven illumination on image quality.

[0050] First, the image is processed using a multi-scale Gaussian function. The Gaussian function is an algorithm widely used in image processing that can smooth and filter images. By processing multiple levels of Gaussian functions, the details and characteristics of the image at each level can be better captured. The purpose of this step is to accurately separate the illumination characteristics in the image and lay a good foundation for subsequent image processing.

[0051] Next, based on the extracted illumination characteristics, the parameters of the two-dimensional gamma function were adjusted accordingly. The two-dimensional gamma function is an effective tool for processing image brightness and contrast, and can flexibly adjust these properties. By adjusting various parameters, the target area can be recognized under different lighting conditions. For example, in dark places, the brightness is increased to improve the details and overall clarity of the picture.

[0052] After completing the adaptive adjustment process, the imaging effect caused by uneven lighting is significantly improved. By adjusting the light and dark contrast of the picture, the details of the picture are more obvious.

[0053] The illumination component can be obtained by convolving the Gaussian function with the original image. G is the Gaussian function, F is the original image, and I is the illumination component.

[0054]

[0055] I(x,y)=F(x,y)G(xy) (5)

[0056] Among them, γ is the exponential value of brightness enhancement, and the two-dimensional gamma function O(x, y) can be determined by the illumination component and the brightness mean m of the illumination component, that is:

[0057]

[0058] To avoid the mutual influence of RGB channels, the entire processing is carried out in the HSV color space. First, the brightness component V is processed, and then converted from the HSV space to the RGB space.

[0059] The two-dimensional gamma algorithm has significant advantages and effects in processing license plate recognition. This method not only enhances the visual clarity and quality of the image, but also improves the accuracy of recognition. Based on these advantages, the algorithm shows great potential and practical value in the application of license plate recognition.

[0060] This project conducts experiments based on the Windows platform, relying on cloud servers to build a deep learning framework software environment based on Pytorch, making full use of the computing power and stable operating environment of cloud server providers to provide support for the training of deep learning models.

[0061] In the present invention, the YOLO algorithm is mainly used based on the v8 version. The various parts of the model work together. Backbone is mainly responsible for feature extraction of the input image; Neck is responsible for multi-scale feature fusion of the feature map and passing these features to the prediction layer; Head performs the final regression prediction. YOLOv8 adopts the idea of ​​CSP, but compared with YOLOv5, the C3 module in its backbone network is replaced by the C2f module. The C2f module has a larger jump and additional Split operation compared to the C3 module in structure, which helps to enhance the feature extraction ability of the network. Neck fuses feature maps of different levels through a series of upsampling, downsampling and splicing operations, thereby obtaining a multi-scale feature representation that contains both global information and local details. These multi-scale features not only improve the model's ability to recognize targets, but also help the model better handle targets of different scales. The Head part is responsible for converting the feature map processed by Backbone and Neck into the final detection result.

[0062] In YOLOv8, the Neck part adopts a structure similar to PA-FPN. YOLOv8 also adopts a dynamic allocation strategy for positive and negative samples. Compared with the IOU matching or unilateral ratio allocation method used in YOLOv5, this method can more accurately determine positive and negative samples, thereby improving the accuracy of detection. The strategy of dynamic allocation of positive and negative samples can be simply summarized as: select positive samples based on the weighted scores of classification and regression.

[0063] t=s α ×u β (1)

[0064] Among them, s is the prediction score, u is the IOU of the prediction box and the target box, α and β are weight hyperparameters, and the multiplication of the two can measure the degree of alignment. When the score of the classification loss Cls is higher and the IOU is higher, the value of t is closer to 1. By training t, the network can be guided to dynamically focus on high-quality positive samples.

[0065] For position loss, YOLOv8 is divided into two parts: CIou_Loss and Distribution Focal Loss. The first part is to calculate the IOU between the predicted box and the target box, where CIOU Loss is used, and the second part uses DFL. The formula of DFL is as follows:

[0066] DFL=-((y i+1 -y)log(p i )+(yy i )log(p i+1 )) (2)

[0067] DFL allows the network to focus on the values ​​near the target y more quickly, thereby accelerating the network's locking of the distribution of the target's neighboring area. For the classification loss Cls, YOLOv8 uses the BCE loss, and the calculation formula is as follows:

[0068] BCE=-(y i log(p(y i )+(1-y i )log(1-p(y i )) (3)

[0069] In order to improve the target detection capability of YOLO, the present invention combines the current attention mechanism with multiple attempts, combining GAM (Global and Aggregative Multi-Context Attention) with YOLO to improve the license plate recognition algorithm.

[0070] YOLO is a real-time target detection algorithm with the advantages of simplicity and high efficiency, but it still has certain limitations in dealing with complex backgrounds and interference. At the same time, GAM, as an attention mechanism, can integrate global and local context information, making the model more focused on key features, thereby improving object detection accuracy.

[0071] The present invention adds a GAM module to the feature extraction stage of YOLO, so that it can pay attention to the extracted features. In this way, YOLO can pay more attention to the key information in the image in subsequent classification and regression tasks, thereby improving the accuracy of target detection. Then, the model is trained using the labeled data set. During the training process, a suitable optimization algorithm and loss function are used to ensure that the model can fully learn the distribution and characteristics of the data. At the same time, the hyperparameters are adjusted in this paper to optimize the performance of the model.

[0072] The original YOLOv8 model uses the C2f module for feature extraction, but the skip-layer connection causes a global grasp of the features. Therefore, this paper improves the Backbone part of the YOLOv8 model and enters the GAM attention mechanism after the Concat connection to improve the performance of the model by retaining the association between spatial and channel information. GAM can well extract the association information between different channels, thereby better distinguishing multiple objects.

[0073] In the GAM operation mechanism, given the input feature map F1 ∈R C×H×W , intermediate state F 2 and output F 3 Defined as formula (7) and (8):

[0074]

[0075] M C and M S They are channel attention map and spatial attention map respectively. Represents an element-wise multiplication operation. Among them, the channel attention submodule uses a three-dimensional arrangement to retain information in three dimensions. Then a two-layer MLP is used to amplify the cross-dimensional channel-spatial dependency. Cross-dimensional channel-spatial dependency refers to the dependency between different channels (i.e., feature dimensions) and spatial dimensions (pixel positions in images or videos). It is very important information in image or video data. This dependency captures the interaction and relationship between different features and spatial positions in order to better understand and process image or video data. Here, in order to focus on spatial information, the spatial attention submodule uses two convolutional layers to fuse spatial information.

[0076] During the training process of the YOLOv8 model after the introduction of GAM, the weight parameter file obtained from the training is saved every 300 trainings to observe the decrease of the loss function curve of the TensorBoard platform in real time and evaluate the algorithm training effect.

[0077] After 9,000 iterations of training, the YOLOv8 algorithm with the introduction of GAM is basically able to recognize and locate the license plate position, which is manifested in the model's error in identifying the type and position of the target gradually decreasing and tending to converge.

[0078] Ultimately, our visual recognition model achieved a recognition accuracy rate of 96%. This accuracy rate ensures that license plate information can be quickly and accurately recognized in various complex scenarios, such as different lighting conditions and high-speed driving, providing reliable data support for subsequent vehicle management and system operation.

[0079] An embodiment of the present invention provides a parking and car-finding method based on the above-mentioned smart parking system.

[0080] 1. Parking

[0081] Step 1: When a vehicle enters the parking lot entrance, the parking lot entrance and exit terminal module starts the vehicle visual recognition detection module to start the visual recognition process. First, the visual recognition module corrects the license plate image through the adaptive correction algorithm for uneven illumination images based on the two-dimensional gamma function, and then uses the YOLO license plate recognition algorithm with integrated GAM attention mechanism to accurately identify the license plate information and vehicle color, thereby determining the vehicle type. Finally, all recognition information will be transmitted to the parking lot central control module in real time through the communication module.

[0082] Step 2: The parking lot central control module uses the geomagnetic sensor module to detect the vacancy status of parking spaces based on the received vehicle information, allocates a suitable parking space for the vehicle, and plans a path from the entrance to the designated parking space, which is then fed back to the parking lot entrance and exit terminal module.

[0083] Step 3: The parking lot entrance and exit terminal module conveys the navigation information to the driver, and the driver quickly finds the parking space and parks the vehicle according to the navigation route.

[0084] Step 4: Once the vehicle enters and stops in the designated parking space, the geomagnetic sensor responds immediately and starts the built-in parking timer. This process is fully automated without manual intervention, ensuring that the parking time is accurately recorded from the moment the vehicle is parked.

[0085] The working principle of geomagnetic sensors is based on the natural phenomenon of the earth's magnetic field. When ferromagnetic objects (such as car bodies) enter the sensor's detection range, they will produce a measurable disturbance in the surrounding geomagnetic field. The sensor uses this physical phenomenon to sense and identify the presence of the vehicle.

[0086] The embodiment of the present invention provides a geomagnetic sensor module, including a geomagnetic sensor circuit board, a geomagnetic sensor housing and a battery module. The geomagnetic sensor circuit board is placed in the geomagnetic sensor housing and fixed by four M3 screws. It adopts a double-panel design, and all modules are arranged in order on the top or bottom of the circuit board to reduce costs and simplify the manufacturing process, ensuring that the modules are compactly and efficiently arranged. Multiple key modules are integrated on the circuit board: the operation indicator module is used to intuitively display the working status; the crystal oscillator module provides a stable clock signal; the download interface module is convenient for program update and maintenance; the power input and filtering and conversion modules together constitute a stable and reliable power supply system; the LoRa wireless module realizes the remote communication function; the RM3100 three-axis geomagnetic sensor module is used as a core sensing element to accurately capture the changes in the geomagnetic field. The operation indicator module, crystal oscillator module, power filter, LoRa wireless module and RM3100 three-axis geomagnetic sensor module are arranged on the top layer of the circuit board, and the download interface module, power input and power conversion module are arranged on the bottom layer of the circuit board. The battery module is embedded in the geomagnetic sensor housing to provide power for the geomagnetic sensor module, and the battery module is closely connected to the power input interface at the bottom layer of the circuit board through the power cord to ensure the stability of voltage and current. The highly sensitive RM3100 geomagnetic sensor inside the sensor continuously monitors and measures subtle changes in the surrounding magnetic field strength. When a vehicle is parked above the sensor, the ferromagnetic material contained in its body will significantly change the local intensity distribution of the geomagnetic field. By capturing and analyzing this change, the sensor can accurately determine whether the vehicle has been parked in the parking space.

[0087] When the vehicle passes over the sensor dynamically, the sensor will record the change of magnetic field strength in three-dimensional space, especially the change on the Z axis (perpendicular to the ground). When the vehicle body is parallel to the sensor, the magnetic field strength on the Z axis will reach a peak due to the concentration effect of ferromagnetic materials in the vehicle body. This characteristic curve provides the sensor with important information about the vehicle's position and posture.

[0088] In order to accurately determine whether the vehicle is parked stably, the geomagnetic sensor first deeply processes the captured magnetic field intensity change data, and uses a smoothing algorithm to eliminate random noise and short-term fluctuations to ensure the stability and reliability of the data. Here we use the mean filtering method. For the signal (a1, a2, ..., an) acquired by the RM3100 geomagnetic sensor at a certain frequency, a window of size L is defined, and the mean of each signal acquired in the window is calculated:

[0089]

[0090] As a corresponding window The output value at the location.

[0091] When the vehicle is about 1 foot above the sensor (or within the preset sensitivity range), and the smoothed Z-axis magnetic field intensity curve shows a clear and continuous peak feature, the sensor determines that the vehicle is stably parked. At this time, the stable peak feature value is vmax, and we will use the difference between vmax and vmean to specifically determine the size and type of the vehicle parked in the parking space.

[0092] D = vmean – vmax

[0093] The larger the D value, the greater the change in the mean magnetic field. Different D value intervals are used to distinguish the size and type of vehicles. The specific D value needs to be determined based on the actual situation.

[0094] This status information is then uploaded to the parking lot central control module in real time, providing solid data support for subsequent parking space management and billing processes.

[0095] By setting scientific and reasonable thresholds, the interference signals from adjacent lanes or distant vehicles can be effectively distinguished and filtered out to ensure the accuracy of parking detection. On this basis, the information communication module responds quickly and sends the verified parking time data accurately to the parking lot central control module, providing real-time and accurate data support for the intelligent management of the entire parking lot, thereby optimizing resource allocation, reducing false alarms and missed alarms, and improving user experience.

[0096] 2. Find a car

[0097] Step 1: The driver inputs the license plate information or parking space information through the human-computer interaction system of the parking lot entrance and exit terminal module and sends a car search request to the parking lot central control module.

[0098] Step 2: After receiving the request to find the vehicle, the central control module of the parking lot uses the integrated data of the visual recognition module and the geomagnetic sensor module to determine the location of the vehicle, and plans a route to find the vehicle, which is then fed back to the human-computer interaction system of the parking lot entrance and exit terminal module.

[0099] Step 3: The human-computer interaction system of the parking lot entrance and exit terminal module conveys the vehicle search route information to the driver, and the driver quickly finds the vehicle according to the navigation route and leaves the parking space.

[0100] Step 4: After the vehicle leaves the parking space, the geomagnetic sensor module detects that the parking space becomes idle and updates the parking space status information of the parking lot central control module through the information communication module.

[0101] Step 5: Drivers can check the detailed parking time through the parking payment function of the mini program. After confirmation, they can complete the online payment according to the garage charging standards.

[0102] In addition, the system also provides parking space reservation services, allowing drivers to reserve empty parking spaces in advance and achieve quick parking. System managers can monitor the parking information of vehicles in the garage in real time through the web page.

[0103] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A smart parking system, characterized in that: It includes a parking lot central control module, a parking lot entrance and exit terminal module and a geomagnetic sensor module; wherein the parking lot central control module includes a first information communication module and a microprocessor module, wherein the first information communication module is used to transmit signals with the parking lot entrance and exit terminal module and the geomagnetic sensor module, and the microprocessor module is used to process and optimize the signals received by the parking lot central control module; the communication of the geomagnetic sensor module is sent by the edge server to the parking lot central control module, which is connected to the local server after processing, and then uploaded to the cloud server by the local server, and the cloud server sends the specific information of the garage to the parking lot entrance and exit terminal module to realize interaction with the user.

2. A smart parking system as claimed in claim 1, characterized in that: The geomagnetic sensor module includes a geomagnetic sensor, a lora wireless communication module, a battery module and an edge server; the geomagnetic sensor is used to detect the geomagnetic changes above the parking space to determine whether there is a vehicle parked. A parking garage will be divided into several areas according to the size of the garage, and the information of all the geomagnetic sensor modules in an area will be sent to the edge server in the area through the lora wireless communication module; The battery module is used to power the RM3100 geomagnetic sensor and lora wireless communication module.

3. A smart parking system as claimed in claim 1, characterized in that: The parking lot entrance and exit terminal module includes a vehicle visual recognition detection module and a second information communication module. When a vehicle enters the parking lot, the parking lot entrance and exit terminal module obtains the vehicle's valid information through the vehicle visual recognition detection module and transmits the information to the parking lot central control module through the second information communication module.

4. A smart parking system as claimed in claim 3, characterized in that: The parking lot entrance and exit terminal module also includes a human-computer interaction module. When the driver needs to find a vehicle, he or she inputs the license plate information or parking space information through the human-computer interaction module, and the parking lot central control module will locate the vehicle position and provide a vehicle search route.

5. The intelligent parking system according to claim 3, characterized in that: The vehicle visual recognition detection module first improves the image quality under different conditions through an image enhancement algorithm, and then uses the YOLOv8 model to accurately locate the license plate position, thereby obtaining the license plate information.

6. A smart parking system as claimed in claim 5, characterized in that: The vehicle visual recognition detection module is based on an adaptive correction algorithm for uneven illumination images using a two-dimensional gamma function. It removes illumination components in an image by using a multi-scale Gaussian function, thereby eliminating the influence of uneven illumination on image quality.

7. A smart parking system as claimed in claim 6, characterized in that: The vehicle visual recognition detection module removes the illumination component in the image. The following steps are involved: Firstly, the multi-scale Gaussian function is used to process the image and separate the illumination characteristics in the image; Then, based on the extracted illumination characteristics, the parameters of the two-dimensional gamma function were adjusted accordingly to achieve the recognition of the target area under different illumination conditions.

8. A smart parking system as claimed in claim 7, characterized in that: The illumination component G is a Gaussian function, F is the original image, and I is the illumination component. I(x,y)=F(x,y)G(x,y) (5) Wherein, γ is an exponential value of brightness enhancement, and the two-dimensional gamma function O(x, y) can be determined by the illumination component and the brightness mean m of the illumination component.

9. A parking method based on a smart parking system according to any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: When a vehicle enters the parking lot entrance, the parking lot entrance and exit terminal module starts the visual recognition process; Step 2: The parking lot central control module uses the geomagnetic sensor module to detect the vacancy status of parking spaces based on the received vehicle information, allocates a suitable parking space for the vehicle, and plans a path from the entrance to the designated parking space, and feeds back to the parking lot entrance and exit terminal module; Step 3: The human-computer interaction module of the parking lot entrance and exit terminal module conveys the navigation information to the driver, and the driver quickly finds the parking space and parks the vehicle according to the navigation route. Step 4: Once the vehicle enters and stops at the designated parking space, the geomagnetic sensor responds immediately and starts the built-in parking timer.

10. A method for finding a car based on a smart parking system according to any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: The driver inputs the license plate information or parking space information through the human-computer interaction module of the parking lot entrance and exit terminal module, and sends a car search request to the parking lot central control module; Step 2: After receiving the request to find the vehicle, the parking lot central control module uses the vehicle visual recognition detection module and the geomagnetic sensor module of the parking lot entrance and exit terminal module to determine the location of the vehicle, plan a route to find the vehicle, and feed it back to the human-computer interaction module of the parking lot entrance and exit terminal module; Step 3: The human-computer interaction module of the parking lot entrance and exit terminal module conveys the vehicle search route information to the driver, and the driver quickly finds the vehicle according to the navigation route and leaves the parking space; Step 4: After the vehicle leaves the parking space, the geomagnetic sensor module detects that the parking space becomes idle, and updates the parking space status information of the parking lot central control module through the first information communication module; Step 5: The driver can check the detailed parking time through the parking payment function of the mini program. After confirmation, he / she can complete the online payment according to the garage charging standards.

Images