Intelligent parking navigation system based on image recognition and Internet of Things

Abstract

The invention discloses an intelligent parking navigation system based on image recognition and the Internet of Things, and relates to the technical field of intelligent parking navigation, the system draws a three-dimensional electronic map of a parking lot, divides the parking lot into monitoring areas, deploys sensors and cameras to collect image data, and transmits the image data to a cloud through the Internet of Things. The intelligent regulation and control model module uses a convolutional neural network CNN to carry out parking space recommendation and path guidance. And the vehicle monitoring module monitors vehicles entering the parking lot in real time, calculates a geometric environment adaptation coefficient and evaluates a traffic risk. And the vehicle guiding analysis module calculates the traffic flow and the parking space idle index, and recommends a proper parking area for the vehicle. The parking space recommendation analysis module evaluates the comfort level of the parking space, ensures the safety of backing and parking, and avoids the risk of traffic jam or traffic accident through strategy adjustment. The system improves the parking efficiency and safety, and optimizes the parking experience.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent parking navigation, and specifically to an intelligent parking navigation system based on image recognition and the Internet of Things. Background Art

[0002] With the acceleration of the urbanization process and the continuous increase in the number of motor vehicles, the problem of parking difficulty has gradually become a major problem in urban management. Especially in large underground parking lots with complex structures, when vehicle owners enter the underground parking lot, they may encounter problems such as height limits, width limits, and chassis scraping on slopes. When looking for available parking spaces, they also face problems such as congestion, time waste, a large distance between the underground location of the parking space and the above-ground destination location, and unreasonable utilization of the parking lot space.

[0003] Traditional parking systems mainly rely on manual management and vehicle owners' independent search for parking spaces. Although this mode has met the parking needs to a certain extent during a certain period, with the acceleration of the urbanization process and the sharp increase in parking demand, the traditional system has exposed multiple drawbacks such as low parking efficiency, low utilization rate of the parking lot space, poor vehicle owner experience, high manual management cost, lack of intelligence and flexibility, inaccurate matching of parking spaces and vehicles, and potential safety hazards. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides an intelligent parking navigation system based on image recognition and the Internet of Things to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: An intelligent parking navigation system based on image recognition and the Internet of Things, comprising a parking area division and deployment module, an intelligent regulation model module, a vehicle monitoring module, a vehicle guidance analysis module, and a parking space recommendation analysis module;

[0006] The parking area division and deployment module is used to draw a three-dimensional electronic map of the commercial area underground parking lot according to the layout of the first basement floor and the second basement floor of the commercial area underground parking lot, and divide the commercial area underground parking lot into several monitoring areas. High-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, ultrasonic sensors, and LED indicators are deployed inside each monitoring sub-area; and vehicle geometric images, license plate images, traffic flow images, environmental terrain images, parking space information images, and path information images of the parking lot are collected; after preprocessing the images, they are converted into image data, and an image data set is established; and the image data set data is sent to cloud storage through Internet of Things transmission technology;

[0007] The intelligent regulation model module is used to construct an initial model using a Convolutional Neural Network (CNN), and train and test it with historical data and a three-dimensional electronic map of the parking lot. The trained model is used as the intelligent regulation model, which performs feature recognition through the feature vectors of the image dataset and further optimizes the model. Finally, the trained intelligent regulation model is used for intelligent parking space recommendation and route guidance;

[0008] The vehicle monitoring module is used to receive the image dataset transmitted by the Internet of Things in real time, monitor the vehicle information of the vehicle to enter the parking lot, calculate the geometric environment adaptation coefficient (PSC), and compare it with the first threshold (Q1) for analysis to determine whether the monitored vehicle meets the parking lot size requirements and whether there is a risk of traffic difficulties, and give strategies to vehicles that do not meet the requirements;

[0009] The vehicle guidance analysis module is used to communicate wirelessly with the in-vehicle navigation system through Internet of Things technology to obtain the destination information of the vehicle in real time, and combine the data of the image dataset to calculate the traffic volume index (TVI), available parking space index (AVI), driving route index (PTI), and historical preference index (HUI). Further calculate the parking area recommendation coefficient (PARI), and compare it with the second threshold (Q2) to determine whether the target area meets the recommendation conditions. If it does not meet the recommendation conditions, calculate the underground second floor recommendation coefficient (PARI2) of the target area. If it still does not meet the conditions, abandon the recommendation of this parking area;

[0010] The parking space recommendation analysis module is used to monitor the parking space information of the target area in real time when the vehicle enters the target area, calculate the parking space comfort coefficient (SSX), and compare it with the third threshold (Q3) for analysis to determine whether there is a risk that the recommended parking space cannot be reverse parked or traffic jams or accidents are caused by the difficulty of reverse parking, and give strategies to vehicles at risk.

[0011] Preferably, the parking area division and deployment module includes a deployment unit, a collection unit, and an image processing unit;

[0012] The deployment unit is used to draw a three-dimensional electronic map of the underground parking lot in the commercial area based on CAD drawing software and the layout of the first basement floor and the second basement floor of the underground parking lot in the commercial area, divide the underground parking lot in the commercial area into several monitoring areas, and mark the several monitoring areas on the three-dimensional electronic map of the underground parking lot in the commercial area as: Jcq1, Jcq2,..., Jcq n ; n represents the number of monitoring areas; deploy high-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, ultrasonic sensors, and LED indicators inside each monitoring sub-area;

[0013] The acquisition unit is used to monitor the vehicle geometric image, license plate image, traffic flow image, environmental terrain image, parking space information image, and path information image of the parking lot in real time based on the deployed high-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, and ultrasonic sensors;

[0014] The image processing unit is used to preprocess the images, including noise removal, grayscale conversion, contrast enhancement, and edge detection operations. After processing, the image data is converted into matrix form, and the numerical value of each pixel represents the brightness or color information of the image, which is converted into image data to establish an image data set; and the data of the image data set is sent to the cloud storage through the Internet of Things transmission technology.

[0015] Preferably, the intelligent regulation model module is used to utilize a convolutional neural network to construct an initial convolutional neural network model, and use the three-dimensional electronic map corresponding to the historical data of the underground parking lot in the commercial area to train and test the initial convolutional neural network model, and use the initial convolutional neural network model after training as the intelligent regulation model. At the same time, the output of the data middle layer of the image data set is used as a feature vector to identify the feature information, and the intelligent regulation model is trained and tested through the obtained feature information, and the intelligent regulation model after training is used for data operation prediction to provide intelligent parking space recommendations and path guidance for the vehicles entering the parking lot.

[0016] Preferably, the vehicle monitoring module includes a first calculation unit and a first analysis unit:

[0017] The first calculation unit is used to receive the data of the image data set transmitted by the Internet of Things in real time and monitor the vehicle information data of the vehicle to enter the parking lot. After dimensionless processing, the geometric environment adaptation coefficient PSC is calculated, and the formula is as follows:

[0018]

[0019] In the formula, W car represents the vehicle width, W max represents the parking lot lane width, H car represents the vehicle height, H max represents the parking lot height limit, L car represents the vehicle length, L max represents the maximum allowable length of the parking space in the parking lot, θ turn represents the minimum turning angle of the parking lot, θ max represents the maximum turning angle of the parking lot, G slope represents the slope of the parking lot entrance, G max represents the maximum allowable slope, K min represents the minimum vehicle chassis height allowed for the parking lot ramp, K carIt represents the ground clearance of the vehicle chassis, r1 represents the width weight coefficient, r2 represents the height weight coefficient, r3 represents the length weight coefficient, R1 represents the turning angle weight coefficient, R2 represents the slope weight coefficient, and R3 represents the chassis weight coefficient.

[0020] Preferably, the first analysis unit is used to preset a first threshold Q1 in advance, and compare and analyze the geometric environment adaptation coefficient PSC with the first threshold Q1 to obtain the first evaluation result, including:

[0021] When the geometric environment adaptation coefficient PSC ≥ the first threshold Q1, it means that the monitored vehicle enters the parking lot with qualified dimensions and there is no risk of traffic difficulty caused by geometric environment factors, and continuous monitoring is carried out;

[0022] When the geometric environment adaptation coefficient PSC < the first threshold Q1, it means that the monitored vehicle enters the parking lot with unqualified dimensions and there is a risk of traffic difficulty caused by geometric environment factors, triggering a first warning instruction and generating a first strategy: according to the real-time data fed back by the Internet of Things, and combining the type, slope, turning radius and parking space size information of the monitored vehicle, recommend a suitable parking lot and special parking spaces for the vehicle.

[0023] Preferably, the vehicle guidance analysis module includes a second calculation unit, a third calculation unit and a second analysis unit;

[0024] The second calculation unit is used to continuously monitor the vehicles entering the parking lot, conduct wireless communication with the in-vehicle navigation system through Internet of Things technology, and obtain the destination information of the vehicles in real time. After dimensionless processing in combination with the image dataset data, calculate and obtain the traffic volume index TVI of the underground parking lot, the available parking space index AVI of the target area, the driving path index PTI and the historical preference index HUI respectively. The formulas are as follows:

[0025]

[0026] In the formula, C jc represents the number of vehicles entering the parking lot in the past five minutes, C lc represents the number of vehicles leaving the parking lot in the past five minutes, C max represents the maximum vehicle capacity of the parking lot, V avg represents the average vehicle speed, V limit represents the speed limit of the parking lot;

[0027]

[0028] In the formula, CWf represents the number of real-time available parking spaces in the target area, CWm represents the total number of parking spaces in the target area, and Cqx jc represents the number of vehicles entering the target area in the past five minutes;

[0029]

[0030] Wherein, Cpath represents the lane flow of the parking lot, Cmax represents the maximum lane flow carrying capacity, Dpath represents the distance to the target area, Ln represents the number of lanes in the parking lot, and a1 and a2 represent weight coefficients;

[0031]

[0032] Wherein, Hqy represents the historical number of times the user enters the target area, and Htcc represents the historical number of times the user enters the parking lot.

[0033] Preferably, the third calculation unit is used to calculate and obtain the traffic volume index TVI of the underground parking lot, the available parking space index AVI of the target area, the driving path index PTI, and the historical preference index HUI through combined calculation. After dimensionless processing, the parking area recommendation coefficient PARI is calculated and obtained, and the formula is as follows:

[0034] PARI = w1 * TVI + w2 * AVI + w3 * PTI + w4 * HUI;

[0035] Wherein, w1, w2, w3, and w4 represent weight coefficients.

[0036] Preferably, the second analysis unit is used to preset a second threshold Q2 in advance, and compare and analyze the parking area recommendation coefficient PARI with the second threshold Q2 to obtain the second evaluation result, including:

[0037] When the parking area recommendation coefficient PARI ≥ the second threshold Q2, it means that the target area meets the recommendation conditions, without the risk of congestion and no available parking spaces. Combining with the A* algorithm, the optimal path to the target area is generated, and the path is guided by an LED light;

[0038] When the parking area recommendation coefficient PARI < the second threshold Q2, it means that the target area does not meet the recommendation conditions, with the risk of congestion and no available parking spaces. The second warning instruction is triggered, and the second strategy is generated: calculate the parking area recommendation coefficient PARI2 of the second basement floor of the parking area, and compare it with the second threshold Q2 again to obtain the third evaluation result, including:

[0039] When the parking area recommendation coefficient PARI2 of the second basement floor of the parking area ≥ the second threshold Q2, it means that the second basement floor of the target area meets the recommendation conditions, without the risk of congestion and no available parking spaces. Combining with the A* algorithm, the optimal path to the second basement floor of the target area is generated, and the path is guided by an LED light;

[0040] When the recommended coefficient PARI2 of the second basement floor of the parking area is less than the second threshold Q2, it indicates that the second basement floor of the target area does not meet the recommended conditions, and there are risks of congestion and no available parking spaces. The third warning instruction is triggered, and the third strategy is generated: abandon the recommended parking area.

[0041] Preferably, the parking space recommendation analysis module includes a fourth calculation unit and a third analysis unit;

[0042] The fourth calculation unit is used to, after the vehicle enters the target area, monitor the parking space information of the target area in real time. After dimensionless processing, the parking space comfort coefficient SSX is calculated and obtained. The formula is as follows:

[0043]

[0044] In the formula, W slot represents the real-time available width of the parking space, W car represents the vehicle width, L slot represents the real-time available length of the parking space, L car represents the vehicle length, θ entry represents the available angle for reverse parking into the garage, θ opt represents the ideal angle for reverse parking into the garage, and w5, w6, and w7 represent weight coefficients;

[0045]

[0046] In the formula, CW represents the lane width, CL represents the lane length, Rmin represents the minimum turning radius of the vehicle, and γ represents the weight value of the turning radius influence.

[0047] Preferably, the third analysis unit is used to preset the third threshold Q3 in advance, and compare and analyze the parking space comfort coefficient SSX with the third threshold Q3. The obtained fourth evaluation results include:

[0048] When the parking space comfort coefficient SSX ≥ the third threshold Q3, it indicates that the comfort of the recommended parking space is qualified, and there is no risk of being unable to reverse park into the garage and traffic jams or accidents caused by the difficulty of reverse parking. Through the A* algorithm, the optimal path to the recommended parking space is generated, and the path is guided by an LED light;

[0049] When the parking space comfort coefficient SSX < the third threshold Q3, it indicates that the comfort of the recommended parking space is unqualified, and there is a risk of being unable to reverse park into the garage and traffic jams or accidents caused by the difficulty of reverse parking. The fourth warning instruction is triggered, and the fourth strategy is generated: re-recommend other parking spaces in the area, recalculate until the parking space comfort coefficient SSX ≥ the third threshold Q3, and then through the A* algorithm, the optimal path to the recommended parking space is generated, and the path is guided by an LED light.

[0050] The present invention provides an intelligent parking navigation system based on image recognition and the Internet of Things. It has the following beneficial effects:

[0051] (1) The intelligent parking navigation system based on image recognition and the Internet of Things can, through the collaborative work of image recognition technology and Internet of Things devices, monitor the occupancy of each parking space in real time, and based on the destination of the vehicle owner and real-time traffic flow information, intelligently recommend the optimal parking space. The AI intelligent algorithm combines image data and the 3D electronic map of the parking lot to reduce the time for vehicle owners to search for parking spaces and significantly improve parking efficiency.

[0052] (2) The intelligent parking navigation system based on image recognition and the Internet of Things, by analyzing the adaptability of the vehicle size to the parking lot environment in real time, ensures that each vehicle can smoothly enter a parking space that meets its size requirements through image recognition technology, avoiding traffic difficulties or safety risks caused by unsuitable parking spaces. In addition, the intelligent guiding path can help vehicle owners find the target parking space more easily and enhance the parking experience.

[0053] (3) The intelligent parking navigation system based on image recognition and the Internet of Things combines image recognition and Internet of Things sensors. By obtaining the vehicle information and parking space usage status in the parking lot in real time, it can accurately identify occupied and unoccupied parking spaces, and thus allocate parking resources more scientifically. The AI intelligent regulation model can dynamically adjust the parking space recommendations in the parking area, avoid waste of parking spaces, and optimize the overall space utilization rate.

[0054] (4) The intelligent parking navigation system based on image recognition and the Internet of Things, by monitoring the environmental changes and vehicle conditions in the parking lot in real time, can timely identify potential safety hazards, such as narrow parking spaces, difficult reversing, or blocked lanes. The AI intelligent analysis and early warning mechanism can trigger risk alerts in advance, provide safety suggestions for vehicle owners, and reduce the accident risk caused by reversing or improper operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a block diagram and flowchart of an intelligent parking navigation system based on image recognition and the Internet of Things according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0057] Embodiment 1

[0058] Please refer to Figure 1, the present invention provides an intelligent parking navigation system based on image recognition and the Internet of Things, including a parking area division and deployment module, an intelligent regulation model module, a vehicle monitoring module, a vehicle guidance analysis module, and a parking space recommendation analysis module;

[0059] The parking area division and deployment module is used to draw a three-dimensional electronic map of the underground parking lot in the commercial area according to the layout of the first basement floor and the second basement floor of the underground parking lot in the commercial area, and divide the underground parking lot in the commercial area into several monitoring areas. High-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, ultrasonic sensors, and LED indicators are deployed inside each monitoring sub-area; and vehicle geometric images, license plate images, traffic flow images, environmental terrain images, parking space information images, and path information images of the parking lot are collected; after preprocessing the images, they are converted into image data, and an image data set is established; and the image data set data is sent to the cloud storage through the Internet of Things transmission technology;

[0060] The intelligent regulation model module is used to construct an initial model using the convolutional neural network CNN, and train and test it with historical data and the three-dimensional electronic map of the parking lot. The trained model is used as the intelligent regulation model, feature recognition is performed through the feature vectors of the image data set, and the model is further optimized. Finally, the trained intelligent regulation model is used for intelligent parking space recommendation and path guidance;

[0061] The vehicle monitoring module is used to receive the image data set transmitted by the Internet of Things in real time, monitor the vehicle information of the vehicle about to enter the parking lot, calculate the geometric environment adaptation coefficient PSC, and compare and analyze it with the first threshold Q1 to determine whether the monitored vehicle meets the parking lot size requirements and whether there is a risk of traffic difficulties, and give strategies to vehicles that do not meet the requirements;

[0062] The vehicle guidance analysis module is used to communicate wirelessly with the in-vehicle navigation system through the Internet of Things technology, obtain the destination information of the vehicle in real time, and combine the image data set data to calculate the traffic flow index TVI, the parking space vacancy index AVI, the driving path index PTI, and the historical preference index HUI. Further calculate the parking area recommendation coefficient PARI, and compare it with the second threshold Q2 to determine whether the target area meets the recommendation conditions. If it does not meet the recommendation conditions, calculate the second basement floor recommendation coefficient PARI2 of the target area. If it still does not meet the conditions, give up the parking area recommendation;

[0063] The parking space recommendation analysis module is used to monitor the parking space information of the target area in real time when the vehicle enters the target area, calculate the parking space comfort coefficient SSX, and compare and analyze it with the third threshold Q3 to determine whether there is a risk that the recommended parking space cannot reverse into the garage and cause traffic jams or accidents due to the difficulty of reverse parking, and give strategies to vehicles at risk.

[0064] In this embodiment, through the collaborative work of the parking area division and deployment module, intelligent regulation model module, vehicle monitoring module, vehicle guidance analysis module, and parking space recommendation analysis module, the system can accurately monitor and manage the use of parking spaces in the parking lot. Image recognition technology and IoT sensors collect vehicle information in the parking lot in real time, and through AI intelligent model analysis, prediction, and recommendation of the most suitable parking spaces and driving routes, it effectively reduces the time for car owners to find parking spaces. The intelligent algorithm dynamically optimizes parking space recommendations based on multiple factors such as vehicle size and environmental adaptability, traffic flow, and historical preferences to ensure that car owners can find parking spaces that meet their requirements in the shortest possible time, improving parking efficiency and car owner satisfaction. At the same time, the real-time monitoring and early warning mechanism can effectively avoid accidents caused by narrow parking spaces or difficult reversing, improving the safety of the parking lot.

[0065] Embodiment 2

[0066] This embodiment is an explanatory note based on Embodiment 1. Please refer to Figure 1 , specifically, the parking area division and deployment module includes a deployment unit, a collection unit, and an image processing unit;

[0067] The deployment unit is used to draw a three-dimensional electronic map of the commercial area underground parking lot based on CAD drawing software and the layout of the first basement floor and the second basement floor of the commercial area underground parking lot, divide the commercial area underground parking lot into several monitoring areas, and mark the several monitoring areas on the three-dimensional electronic map of the commercial area underground parking lot as: Jcq1, Jcq2,..., Jcq n ; n represents the number of monitoring areas; deploy high-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, ultrasonic sensors, and LED indicators inside each monitoring sub-area;

[0068] The collection unit is used to monitor the vehicle geometry image, license plate image, traffic flow image, environmental terrain image, parking space information image, and path information image in the parking lot in real time based on the deployed high-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, and ultrasonic sensors;

[0069] The image processing unit is used to preprocess the image, including noise removal, grayscale conversion, contrast enhancement, and edge detection operations. After processing, the image data is converted into matrix form, and the value of each pixel represents the brightness or color information of the image, which is converted into image data to establish an image data set; and the image data set data is sent to cloud storage through IoT transmission technology.

[0070] In this embodiment, through the combination of the deployment unit, the acquisition unit, and the image processing unit in the parking area division and deployment module, the system can comprehensively and accurately monitor all monitoring areas of the underground parking lot in the commercial area. The deployment of high-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, and ultrasonic sensors ensures real-time monitoring of each parking space and vehicle in the parking lot, providing multi-dimensional image and data acquisition. After being preprocessed by the image processing unit, these data are converted into available data sets and uploaded to the cloud storage in real time through Internet of Things technology, providing accurate basic information for subsequent intelligent decision-making and parking space recommendation.

[0071] Embodiment 3

[0072] This embodiment is an explanatory description based on Embodiment 2. Please refer to Figure 1 , specifically, the intelligent regulation model module is used to utilize a convolutional neural network to construct an initial convolutional neural network model, and use the three-dimensional electronic map corresponding to the historical data of the underground parking lot in the commercial area to train and test the initial convolutional neural network model, and use the trained initial convolutional neural network model as the intelligent regulation model. At the same time, the data intermediate layer output of the image data set is used as a feature vector to identify feature information, and the intelligent regulation model is trained and tested through the obtained feature information. The trained intelligent regulation model is used for data operation prediction to provide intelligent parking space recommendation and path guidance for vehicles entering the parking lot.

[0073] In this embodiment, through the intelligent regulation model module using a convolutional neural network CNN to construct and train an initial model, combined with the three-dimensional electronic map and historical data of the underground parking lot in the commercial area, the system can accurately identify the real-time parking spaces and vehicle status in the parking lot. By extracting the feature information in the image data set, the intelligent regulation model is further optimized to realize intelligent parking space recommendation and path guidance for vehicles entering the parking lot. This optimization method based on deep learning not only improves the accuracy of parking space allocation, but also can automatically adjust the recommendation strategy according to the real-time traffic flow and parking demand, greatly improving the utilization efficiency of parking lot resources and providing a faster and more convenient parking experience for car owners.

[0074] Embodiment 4

[0075] This embodiment is an explanatory description based on Embodiment 3. Please refer to Figure 1 , specifically, the vehicle monitoring module includes a first calculation unit and a first analysis unit:

[0076] The first calculation unit is used to receive the data of the image data set transmitted by the Internet of Things in real time, monitor the vehicle information data of the vehicle to enter the parking lot, and after dimensionless processing, calculate and obtain the geometric environment adaptation coefficient PSC. The formula is as follows:

[0077]

[0078] In the formula, W car represents the vehicle width, and W max represents the width of the parking lot lane, H car represents the vehicle height, and H max represents the height limit of the parking lot, L car represents the vehicle length, and L max represents the maximum allowable length of the parking space in the parking lot, θ turn represents the minimum turning angle of the parking lot, and θ max represents the maximum turning angle of the parking lot, G slope represents the slope of the parking lot entrance, and G max represents the maximum allowable slope, K min represents the minimum vehicle chassis height allowed for the parking lot ramp, and K car represents the clearance between the vehicle chassis and the ground. r1 represents the width weight coefficient, r2 represents the height weight coefficient, r3 represents the length weight coefficient, R1 represents the turning angle weight coefficient, R2 represents the slope weight coefficient, and R3 represents the chassis weight coefficient. Refer to Table 1 and Table 2:

[0079] Vehicle type Vehicle width Vehicle length Vehicle height r1 r2 r3 Compact car 2.2 4.5 1.8 1.0 0.8 0.9 Mid-size car 2.4 4.8 1.9 1.1 1.0 1.0 Full-size car 2.5 5.0 2.0 1.2 1.1 1.0 SUV 2.6 5.2 2.2 1.3 1.2 1.1 Full-size SUV 2.7 5.5 2.3 1.4 1.3 1.2 MPV 2.8 5.6 2.3 1.4 1.3 1.3

[0080] Table 1

[0081]

[0082]

[0083] Table 2.

[0084] In this embodiment, the vehicle monitoring module receives the image data transmitted by the Internet of Things in real time and performs dimensionless processing. Combining with the calculation of the geometric environment adaptation coefficient PSC, the system can accurately evaluate whether the vehicle entering the parking lot meets the parking lot size and traffic conditions. This calculation is based on a comprehensive analysis of various dimensional parameters of the vehicle (such as width, height, length, chassis height, etc.) and the requirements of the parking lot lane width, height limit, turning angle, slope, etc., effectively determining whether the vehicle is suitable for the parking lot environment and avoiding traffic difficulties or safety risks. This intelligent screening mechanism can automatically adjust the matching degree of the vehicles entering the parking lot, provide more reasonable parking recommendations for different vehicle models, improve the intelligent management level of the parking lot, and reduce the risks of congestion and accidents caused by unsuitable vehicles.

[0085] Example 5

[0086] This embodiment is an explanatory description based on Example 4. Please refer to Figure 1, specifically, the first analysis unit is used to preset a first threshold Q1 in advance, and compare and analyze the geometric environment adaptation coefficient PSC with the first threshold Q1 to obtain the first evaluation result, including:

[0087] When the geometric environment adaptation coefficient PSC ≥ the first threshold Q1, it indicates that the monitored vehicle enters the parking lot with qualified dimensions and there is no risk of traffic difficulties caused by geometric environment factors, and continuous monitoring is carried out;

[0088] When the geometric environment adaptation coefficient PSC < the first threshold Q1, it indicates that the monitored vehicle enters the parking lot with unqualified dimensions and there is a risk of traffic difficulties caused by geometric environment factors, triggering a first warning instruction and generating a first strategy: according to the real-time data fed back by the Internet of Things, and combining the type, slope, turning radius and parking space size information of the monitored vehicle, recommend a suitable parking lot and special parking space for the vehicle.

[0089] In this embodiment, by comparing and analyzing the geometric environment adaptation coefficient PSC with the preset first threshold Q1 through the first analysis unit, the system can evaluate in real time whether the vehicle entering the parking lot meets the traffic dimension requirements of the parking lot. When it is detected that the vehicle dimensions are unqualified, the system automatically triggers an early warning mechanism and generates a targeted strategy, and based on the Internet of Things real-time data combined with factors such as vehicle type, slope, turning radius and parking space size, intelligently recommends a suitable parking area or special parking space. This mechanism can effectively avoid jams and accidents caused by traffic incompatibility.

[0090] Embodiment 6

[0091] This embodiment is an explanatory description carried out in Embodiment 5, please refer to Figure 1 , specifically, the vehicle guidance analysis module includes a second calculation unit, a third calculation unit and a second analysis unit;

[0092] The second calculation unit is used to, in real time monitor the vehicle entering the parking lot, conduct wireless communication with the in-vehicle navigation system through the Internet of Things technology, in real time obtain the destination information of the vehicle, and after dimensionless processing in combination with the image dataset data, respectively calculate and obtain the traffic volume index TVI of the underground parking lot, the parking space vacancy index AVI of the target area, the driving path index PTI and the historical preference index HUI, and the formulas are as follows:

[0093]

[0094] In the formula, C jc represents the number of vehicles entering the parking lot in the past five minutes, C lc represents the number of vehicles leaving the parking lot in the past five minutes, C max represents the maximum vehicle capacity of the parking lot, V avg represents the average vehicle speed, V limitIndicates the speed limit of the parking lot;

[0095]

[0096] In the formula, CWf represents the number of real-time available parking spaces in the target area, CWm represents the total number of parking spaces in the target area, and Cqx jc represents the number of vehicles entering the target area in the past five minutes;

[0097]

[0098] In the formula, Cpath represents the lane flow of the parking lot, Cmax represents the maximum lane flow carrying capacity, Dpath represents the distance to the target area, Ln represents the number of lanes in the parking lot, and a1 and a2 represent weight coefficients;

[0099]

[0100] In the formula, Hqy represents the historical number of times a user enters the target area, and Htcc represents the historical number of times a user enters the parking lot.

[0101] In this embodiment, the system can monitor the vehicle information entering the parking lot in real time, and wirelessly communicate with the in-vehicle navigation system through the Internet of Things technology to accurately obtain the destination information of the vehicle. Combining with the image dataset, the system calculates the traffic flow index TVI, the available parking space index AVI, the driving path index PTI, and the historical preference index HUI, and provides personalized parking area recommendations and the best driving paths for each vehicle. By analyzing the real-time traffic flow conditions, the available parking space situation in the target area, the driving path situation, and the historical preference data, the system can efficiently guide vehicles to avoid congested areas, optimize the parking experience, and reduce the time for the vehicle owner to find a parking space.

[0102] Embodiment 7

[0103] This embodiment is an explanatory description based on Embodiment 6. Please refer to Figure 1 , specifically, the third calculation unit is used to calculate and obtain the traffic flow index TVI of the underground parking lot, the available parking space index AVI of the target area, the driving path index PTI, and the historical preference index HUI in combination. After dimensionless processing, the parking area recommendation coefficient PARI is calculated, and the formula is as follows:

[0104] PARI = w1 * TVI + w2 * AVI + w3 * PTI + w4 * HUI;

[0105] In the formula, w1, w2, w3, and w4 represent weight coefficients, 0 < w1 < 1, 0 < w2 < 1, 0 < w3 < 1, and 0 < w4 < 1, and w1 + w2 + w3 + w4 = 1.

[0106] In this embodiment, by calculating the Parking Area Recommendation Index (PARI), the system can intelligently recommend the most suitable parking area for each vehicle based on real-time traffic flow conditions, available parking spaces, the best driving route, and the vehicle's historical preferences. This process effectively reduces the search time of vehicle owners in the parking lot, improves parking efficiency, and avoids the selection of congested areas.

[0107] Embodiment 8

[0108] This embodiment is an explanatory description based on Embodiment 7. Please refer to Figure 1 , specifically, the second analysis unit is used to preset a second threshold Q2 in advance, and compare and analyze the Parking Area Recommendation Index (PARI) with the second threshold Q2 to obtain the second evaluation result, including:

[0109] When the Parking Area Recommendation Index (PARI) ≥ the second threshold Q2, it indicates that the target area meets the recommendation conditions, without the risk of congestion and no available parking spaces. Combining with the A* algorithm, the optimal route to the target area is generated, and the route is guided by an LED light;

[0110] When the Parking Area Recommendation Index (PARI) < the second threshold Q2, it indicates that the target area does not meet the recommendation conditions, with the risk of congestion and no available parking spaces. The second warning instruction is triggered to generate the second strategy: calculate the Parking Area Recommendation Index (PARI2) for the second basement floor of the parking area, and then compare it with the second threshold Q2 to obtain the third evaluation result, including:

[0111] When the Parking Area Recommendation Index (PARI2) for the second basement floor of the parking area ≥ the second threshold Q2, it indicates that the second basement floor of the target area meets the recommendation conditions, without the risk of congestion and no available parking spaces. Combining with the A* algorithm, the optimal route to the second basement floor of the target area is generated, and the route is guided by an LED light;

[0112] When the Parking Area Recommendation Index (PARI2) for the second basement floor of the parking area < the second threshold Q2, it indicates that the second basement floor of the target area does not meet the recommendation conditions, with the risk of congestion and no available parking spaces. The third warning instruction is triggered to generate the third strategy: abandon the recommended parking area.

[0113] In this embodiment, through the comparative analysis of the Parking Area Recommendation Index (PARI) and the second threshold Q2 by the second analysis unit, the system can dynamically evaluate the recommendation conditions of the target parking area, avoiding vehicle owners from selecting congested or understocked areas. When the recommendation index reaches the standard, the system will generate the optimal driving route through the A* algorithm and use the LED light for real-time guidance to ensure that the vehicle can quickly and smoothly reach the available parking space; if the recommendation index does not reach the standard, the system will trigger a warning and recommend the second basement floor parking area according to the situation to further optimize the route guidance. If the second basement floor also does not meet the conditions, the system will abandon the recommendation of this area, thus effectively improving parking efficiency and reducing unnecessary time waste.

[0114] Example 9

[0115] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically, the parking space recommendation analysis module includes a fourth calculation unit and a third analysis unit;

[0116] The fourth calculation unit is used to, after the vehicle enters the target area, monitor the parking space information of the target area in real time. After dimensionless processing, calculate and obtain the parking space comfort coefficient SSX. The formula is as follows:

[0117]

[0118] In the formula, W slot represents the real-time available width of the parking space, W car represents the vehicle width, L slot represents the real-time available length of the parking space, L car represents the vehicle length, θ entry represents the available angle for reverse parking into the garage, θ opt represents the ideal reverse parking angle into the garage, w5, w6, and w7 represent weight coefficients, 0 < w5 < 1, 0 < w6 < 1, and 0 < w7 < 1 and w5 + w6 + w7 = 1;

[0119]

[0120] In the formula, CW represents the lane width, CL represents the lane length, Rmin represents the minimum turning radius of the vehicle, γ represents the weight value affecting the turning radius, as shown in the following table:

[0121] Vehicle type Minimum turning radius Lane length γ Compact car 4.7 5.5 1.2 Mid-size car 5.2 5.8 1.1 Full-size car 5.5 6.0 1.0 SUV 5.9 6.5 0.9 Full-size SUV 6.2 6.8 0.8 MPV 7.2 7.5 0.7

[0122] Table 3.

[0123] In this embodiment, through the real-time monitoring of the parking space information in the target area by the fourth calculation unit and combining dimensionless processing to calculate the parking space comfort coefficient SSX, the system can accurately evaluate the actual adaptation situation of the parking space. This analysis takes into account multiple factors such as the available width, length, and reverse parking angle of the parking space, and adjusts different factors through weight coefficients to ensure that the vehicle owner can obtain more suitable parking space recommendations. By calculating the parking space comfort coefficient in real time, the system can effectively avoid problems such as unsuitable parking spaces for reverse parking or causing traffic jams, improve parking efficiency, and reduce the difficulties and risks during parking.

[0124] Example 10

[0125] This embodiment is an explanatory description based on Example 9. Please refer to Figure 1, specifically, the third analysis unit is used to preset a third threshold Q3 in advance, and compare and analyze the parking space comfort coefficient SSX with the third threshold Q3 to obtain the fourth evaluation result, including:

[0126] When the parking space comfort coefficient SSX ≥ the third threshold Q3, it indicates that the comfort of the recommended parking space is qualified, and there is no risk of being unable to reverse into the parking space or traffic jams or accidents caused by the difficulty of reversing into the parking space. Through the A* algorithm, the optimal path to the recommended parking space is generated, and the path is guided by an LED light;

[0127] When the parking space comfort coefficient SSX < the third threshold Q3, it indicates that the comfort of the recommended parking space is unqualified, and there is a risk of being unable to reverse into the parking space or traffic jams or accidents caused by the difficulty of reversing into the parking space. The fourth warning instruction is triggered, and the fourth strategy is generated: re-recommend other parking spaces in the area, recalculate until the parking space comfort coefficient SSX ≥ the third threshold Q3, and then through the A* algorithm, the optimal path to the recommended parking space is generated, and the path is guided by an LED light.

[0128] In this embodiment, through the real-time evaluation of the parking space comfort coefficient SSX by the third analysis unit, the system can judge the comfort and safety of the recommended parking space. When the parking space meets the comfort standard, the system will generate the optimal path to guide the vehicle to the recommended parking space to ensure a smooth reverse into the parking space without causing the risk of traffic jams or accidents. When the comfort of the parking space does not meet the requirements, the system will trigger a warning and re-recommend other parking spaces in the area until a suitable parking space is found. This intelligent evaluation mechanism effectively improves the safety and efficiency of the parking process and avoids the risks caused by inappropriate parking spaces.

[0129] The setting of the size of the threshold is for the convenience of comparison. Regarding the size of the threshold, it depends on the amount of sample data and the number of base numbers set by those skilled in the art for each group of sample data; as long as it does not affect the proportional relationship between the parameters and the quantified values.

[0130] The above formulas are all obtained by collecting a large amount of data for software simulation and selecting a formula close to the real value. The coefficients in the formula are set by those skilled in the art according to the actual situation. 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 person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. An intelligent parking navigation system based on image recognition and the Internet of Things, characterized in that, It includes a parking area division and deployment module, an intelligent regulation model module, a vehicle monitoring module, a vehicle guidance analysis module, and a parking space recommendation analysis module; The parking area division and deployment module is used to draw a three-dimensional electronic map of the underground parking lot in the commercial area according to the layout of the first basement floor and the second basement floor of the underground parking lot in the commercial area, and divide the underground parking lot in the commercial area into several monitoring areas. High-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, ultrasonic sensors, and LED indicators are deployed inside each monitoring sub-area; and vehicle geometric images, license plate images, traffic flow images, environmental terrain images, parking space information images, and path information images of the parking lot are collected; after preprocessing the images, they are converted into image data to establish an image data set; and the image data set data is sent to the cloud for storage through Internet of Things transmission technology; The intelligent regulation model module is used to construct an initial model using the convolutional neural network CNN, and train and test it with historical data and the three-dimensional electronic map of the parking lot. The trained model is used as an intelligent regulation model to perform feature recognition through the feature vectors of the image data set, and further optimize the model. Finally, the trained intelligent regulation model is used for intelligent parking space recommendation and path guidance; The vehicle monitoring module is used to receive the image data set transmitted by the Internet of Things in real time, monitor the vehicle information of the vehicle about to enter the parking lot, calculate the geometric environment adaptation coefficient PSC, and compare and analyze it with the first threshold Q1 to determine whether the monitored vehicle meets the parking lot size requirements and whether there is a risk of traffic difficulties, and give strategies to vehicles that do not meet the requirements; The vehicle guidance analysis module is used to communicate wirelessly with the in-vehicle navigation system through Internet of Things technology to obtain the destination information of the vehicle in real time, and combine the image data set data to calculate the traffic flow index TVI, the parking space vacancy index AVI, the driving path index PTI, and the historical preference index HUI. Further calculate the parking area recommendation coefficient PARI, and compare it with the second threshold Q2 to determine whether the target area meets the recommendation conditions. If it does not meet the recommendation conditions, calculate the second basement floor recommendation coefficient PARI2 of the target area. If it still does not meet the conditions, give up the recommendation of this parking area; The parking space recommendation analysis module is used to monitor the parking space information of the target area in real time when the vehicle enters the target area, calculate the parking space comfort coefficient SSX, and compare and analyze it with the third threshold Q3 to determine whether there is a risk that the recommended parking space cannot be reversed into the garage or traffic jams or accidents are caused by the difficulty of reversing into the garage, and give strategies to vehicles at risk.

2. The intelligent parking navigation system based on image recognition and Internet of Things according to claim 1, wherein The parking area division and deployment module includes a deployment unit, a collection unit, and an image processing unit; The deployment unit is used to draw a three-dimensional electronic map of the underground parking lot in the commercial area based on the CAD drawing software and the layout of the first basement floor and the second basement floor of the underground parking lot in the commercial area, divide the underground parking lot in the commercial area into several monitoring areas, and mark the several monitoring areas on the three-dimensional electronic map of the underground parking lot in the commercial area as: Jcq1, Jcq2,..., Jcq n ; n represents the number of monitoring areas; high-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, ultrasonic sensors and LED indicators are deployed inside each monitoring sub-area; The collection unit is used to monitor vehicle geometric images, license plate images, traffic flow images, environmental terrain images, parking space information images, and path information images of the parking lot in real time based on the deployed high-definition cameras, license plate recognition systems, geomagnetic sensors, lidar sensors, and ultrasonic sensors; The image processing unit is used to preprocess the image, including denoising, grayscale conversion, contrast enhancement, and edge detection operations. After processing, the image data is converted into matrix form, and the value of each pixel represents the brightness or color information of the image, which is then converted into image data to establish an image data set. The data of the image data set is sent to the cloud storage through Internet of Things transmission technology.

3. An intelligent parking navigation system based on image recognition and the Internet of Things according to claim 2, characterized in that, The intelligent regulation model module is used to build an initial convolutional neural network model using a convolutional neural network, and use the three-dimensional electronic map of the commercial area underground parking lot corresponding to historical data to train and test the initial convolutional neural network model. The trained initial convolutional neural network model is used as the intelligent regulation model. At the same time, the data middle layer output of the image data set is used as the feature vector to identify the feature information, and the intelligent regulation model is trained and tested based on the obtained feature information. The trained intelligent regulation model is used for data operation prediction to provide intelligent parking space recommendations and route guidance for vehicles entering the parking lot.

4. An intelligent parking navigation system based on image recognition and the Internet of Things according to claim 3, characterized in that, The vehicle monitoring module includes a first calculation unit and a first analysis unit: The first calculation unit is used to receive the data of the image data set transmitted by the Internet of Things in real time, monitor the vehicle information data of the vehicle to enter the parking lot, and calculate and obtain the geometric environment adaptation coefficient PSC after dimensionless processing. The formula is as follows: Where, W car represents the vehicle width, W max represents the width of the parking lot lane, H car represents the vehicle height, H max represents the height limit of the parking lot, L car represents the vehicle length, L max represents the maximum allowable length of the parking space in the parking lot, θ turn represents the minimum turning angle of the parking lot, θ max represents the maximum turning angle of the parking lot, G slope represents the slope of the parking lot entrance, G max represents the maximum allowable slope, K min represents the minimum vehicle chassis height allowed for the ramp in the parking lot, K car represents the ground clearance of the vehicle chassis. r1 represents the width weight coefficient, r2 represents the height weight coefficient, r3 represents the length weight coefficient, R1 represents the turning angle weight coefficient, R2 represents the slope weight coefficient, and R3 represents the chassis weight coefficient.

5. An intelligent parking navigation system based on image recognition and the Internet of Things according to claim 4, characterized in that, The first analysis unit is used to preset a first threshold Q1 in advance, and compare and analyze the geometric environment adaptation coefficient PSC with the first threshold Q1 to obtain the first evaluation result, including: When the geometric environment adaptation coefficient PSC ≥ the first threshold Q1, it indicates that the size of the monitored vehicle entering the parking lot is qualified, and there is no risk of traffic difficulty caused by geometric environment factors, and continuous monitoring is carried out; When the geometric environment adaptation coefficient PSC < the first threshold Q1, it indicates that the size of the monitored vehicle entering the parking lot is unqualified, and there is a risk of traffic difficulty caused by geometric environment factors, triggering a first warning instruction and generating a first strategy: according to the real-time data fed back by the Internet of Things, combined with the type, slope, turning radius, and parking space size information of the monitored vehicle, recommend a suitable parking lot and special parking spaces for the vehicle.

6. An intelligent parking navigation system based on image recognition and the Internet of Things according to claim 5, characterized in that, The vehicle guidance analysis module includes a second calculation unit, a third calculation unit, and a second analysis unit; The second calculation unit is used to monitor the vehicle entering the parking lot in real time, communicate wirelessly with the in-vehicle navigation system through Internet of Things technology, obtain the destination information of the vehicle in real time, and calculate and obtain the traffic volume index TVI of the underground parking lot, the available parking space index AVI of the target area, the driving path index PTI, and the historical preference index HUI after dimensionless processing in combination with the image data set data. The formulas are as follows: Where C jc represents the number of vehicles entering the parking lot in the last five minutes, C lc represents the number of vehicles leaving the parking lot in the last five minutes, C max represents the maximum vehicle capacity of the parking lot, V avg represents the average vehicle speed, V limit represents the speed limit of the parking lot; Where CWf represents the real-time number of available parking spaces in the target area, CWm represents the total number of parking spaces in the target area, and Cqx jc represents the number of vehicles entering the target area in the past five minutes; In the formula, Cpath represents the traffic volume of the lane in the parking lot, Cmax represents the maximum traffic volume carrying capacity of the lane, Dpath represents the distance to the target area, Ln represents the number of lanes in the parking lot, and a1 and a2 represent weight coefficients; In the formula, Hqy represents the historical number of times the user enters the target area, and Htcc represents the historical number of times the user enters the parking lot.

7. An intelligent parking navigation system based on image recognition and the Internet of Things according to claim 6, characterized in that, The third calculation unit is used to calculate and obtain the traffic volume index TVI of the underground parking lot, the parking space vacancy index AVI of the target area, the driving path index PTI, and the historical preference index HUI through combined calculations. After dimensionless processing, the parking area recommendation coefficient PARI is calculated and obtained. The formula is as follows: PARI = w1 * TVI + w2 * AVI + w3 * PTI + w4 * HUI; In the formula, w1, w2, w3, and w4 represent weight coefficients.

8. An intelligent parking navigation system based on image recognition and the Internet of Things according to claim 7, characterized in that The second analysis unit is used to preset a second threshold Q2 in advance, and compare and analyze the parking area recommendation coefficient PARI with the second threshold Q2 to obtain the second evaluation result, including: When the parking area recommendation coefficient PARI ≥ the second threshold Q2, it means that the target area meets the recommendation conditions, without the risk of congestion and no available parking spaces. Combining with the A* algorithm, the optimal path to the target area is generated, and the path is guided by an LED light; When the parking area recommendation coefficient PARI < the second threshold Q2, it means that the target area does not meet the recommendation conditions, with the risk of congestion and no available parking spaces. The second warning instruction is triggered, and the second strategy is generated: calculate the parking area recommendation coefficient PARI2 for the second basement floor, and then compare it with the second threshold Q2 to obtain the third evaluation result, including: When the parking area recommendation coefficient PARI2 for the second basement floor ≥ the second threshold Q2, it means that the second basement floor of the target area meets the recommendation conditions, without the risk of congestion and no available parking spaces. Combining with the A* algorithm, the optimal path to the second basement floor of the target area is generated, and the path is guided by an LED light; When the parking area recommendation coefficient PARI2 for the second basement floor < the second threshold Q2, it means that the second basement floor of the target area does not meet the recommendation conditions, with the risk of congestion and no available parking spaces. The third warning instruction is triggered, and the third strategy is generated: abandon the recommended parking area.

9. The intelligent parking navigation system based on image recognition and the Internet of Things according to claim 8, characterized in that, The parking space recommendation analysis module includes a fourth calculation unit and a third analysis unit; The fourth calculation unit is used to, when the vehicle enters the target area, real-time monitor the parking space information of the target area. After dimensionless processing, the parking space comfort coefficient SSX is calculated and obtained. The formula is as follows: Where, W slot represents the real-time available width of the parking space, W car represents the width of the vehicle, L slot represents the real-time available length of the parking space, L car represents the length of the vehicle, θ entry represents the available angle for reverse parking into the space, θ opt represents the ideal angle for reverse parking into the space, and w5, w6, and w7 represent the weighting coefficients; In the formula, CW represents the lane width, CL represents the lane length, Rmin represents the minimum turning radius of the vehicle, and γ represents the weight value of the turning radius influence.

10. An intelligent parking navigation system based on image recognition and the Internet of Things according to claim 9, characterized in that, The third analysis unit is used to preset a third threshold Q3 in advance, and compare and analyze the parking space comfort coefficient SSX with the third threshold Q3 to obtain the fourth evaluation result, including: When the parking space comfort coefficient SSX ≥ the third threshold Q3, it means that the comfort level of the recommended parking space is qualified, without the risk of being unable to reverse into the parking space and causing traffic jams or accidents due to the difficulty of reverse parking. Through the A* algorithm, the optimal path to the recommended parking space is generated, and the path is guided by an LED light; When the parking space comfort coefficient SSX < the third threshold Q3, it means that the comfort level of the recommended parking space is unqualified, with the risk of being unable to reverse into the parking space and causing traffic jams or accidents due to the difficulty of reverse parking. The fourth warning instruction is triggered, and the fourth strategy is generated: re-recommend other parking spaces in the area, recalculate until the parking space comfort coefficient SSX ≥ the third threshold Q3, and then through the A* algorithm, the optimal path to the recommended parking space is generated, and the path is guided by an LED light.

Images