Intelligent parking space guiding system for hospital parking lot

Abstract

The invention discloses an intelligent parking space guiding system for a hospital parking lot, relates to the technical field of intelligent traffic, and is used for solving the problems that the hospital parking lot is seriously jammed in peak hours and special vehicles lack exclusive passing authority. The system comprises a parking space monitoring module, a vehicle identification module, an intelligent guiding module and a user interaction module. The parking space state is monitored in real time through the multi-mode sensor array; recognizing entering and exiting vehicles through a deep learning recognition algorithm based on cameras deployed at entrances and exits, and obtaining corresponding vehicle information; based on parking space state information and vehicle identification information, a route optimal value is dynamically calculated, and a whole-course navigation service from an entrance to a target parking space is provided for a driver, so that accurate detection of the parking space state, special vehicle priority distribution, high-low peak dynamic route planning and whole-process intelligent guidance are realized, the hospital parking efficiency is remarkably improved, and the hospital parking experience is improved. And the congestion condition in the peak period is relieved.

Description

Technical Field

[0001] The present invention relates to the field of smart transportation technology, and specifically to a smart parking guidance system for a hospital parking lot. Background Art

[0002] With the acceleration of urbanization and the continued growth of car ownership, parking difficulties have become a serious problem plaguing public transportation, especially in high-mobility public service venues such as hospitals. Hospitals face concentrated traffic peaks and a diverse user base, placing high demands on traffic efficiency and priority guarantees. While existing parking management technologies have made some progress in parking space monitoring and vehicle guidance, they still face numerous challenges:

[0003] Existing technologies rely on a single type of sensor to monitor parking space status. This single type of sensor is susceptible to interference. At the same time, existing guidance systems generally use static or single optimal indicator guidance methods, which cannot adapt to the unique traffic peaks and valleys of hospitals, resulting in low traffic efficiency in hospital parking lots.

[0004] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem of low traffic efficiency in hospital parking lots and to propose an intelligent parking guidance system for hospital parking lots.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A smart parking guidance system for hospital parking lots includes a parking monitoring module that uses a multimodal sensor array to monitor parking status data in real time. The specific implementation process includes:

[0008] The multimodal sensor array includes a geomagnetic sensor group and a camera array. The geomagnetic sensor continuously collects the change in magnetic field strength. When the change in magnetic field strength exceeds a preset threshold, the parking space is determined to be occupied. Otherwise, it is determined to be vacant.

[0009] The vehicle identification module uses cameras deployed at each entrance and exit of the hospital parking lot to identify and record vehicle information at the entrance and exit;

[0010] The intelligent guidance module dynamically calculates the optimal route based on parking space status information and vehicle identification information, and provides drivers with full navigation services from the entrance to the target parking space through various guidance methods;

[0011] The user interaction module is used to provide multi-channel information interaction services for hospital visitors, patients and medical staff, and realize intelligent management of the entire parking process.

[0012] Furthermore, the specific implementation process of the camera array monitoring the parking space status includes:

[0013] By presetting physical marking points at the four corners of the parking space, a mapping relationship between pixel coordinates and actual space coordinates is established based on a calibration algorithm, generating a virtual network of rectangular parking spaces bound with unique numbers;

[0014] The vehicle outline is extracted and calculated based on the improved YOLOv5 target detection network. When the outline area occupies more than 40% of the rectangular area, the occupancy value P1 is output;

[0015] In the idle state, the surface texture features of the parking space are pre-collected as a reference template, and the similarity between the current texture features and the reference template is compared in real time; when the texture feature similarity is lower than the preset threshold, it is determined that the surface texture of the parking space has changed significantly and the texture change value P2 is output;

[0016] An improved KCF tracker is used to track vehicle motion trajectories and generate motion trajectory sequences. When a vehicle is detected entering a parking area and its speed drops from an initial value to near zero within 20 consecutive frames, it is considered a parking behavior. When a vehicle is detected starting to move from a stationary state and leaving the parking area within 15 consecutive frames, it is considered a leaving behavior. The occupancy probability value P3 is output.

[0017] After normalizing the occupancy value P1, texture change value P2, and occupancy probability value P3, enter them into the formula S=a1·P1+a2·P2+a 3· P3 obtains the decision value, where a1, a2, and a3 are the influence weight factors of occupancy value, texture change value, and occupancy probability value, respectively;

[0018] When the decision value is greater than the preset threshold and the vehicle outline is detected for 5 consecutive frames, the texture feature matching is successful, and the target stays in the parking space for more than 2 seconds, it is judged to be in an occupied state; when the decision value is less than the preset threshold and no valid vehicle target is detected in the rectangular area for 30 consecutive frames or the vehicle is shown to have left the parking space, it is judged to be in an idle state.

[0019] Furthermore, the specific implementation process of the parking space monitoring module also includes:

[0020] The data from the geomagnetic sensor and the camera are fused based on the data conflict handling mechanism;

[0021] When the results of the geomagnetic sensor and the camera monitoring are consistent, the current parking space status is directly determined and updated; otherwise, a secondary judgment is made and the time window secondary verification mechanism is activated;

[0022] Trigger data collection three times at 5-second intervals within 15 seconds to synchronously obtain the magnetic field intensity change sequence and the camera's parking area video frame;

[0023] The dynamic time warping algorithm compares the magnetic field change curve with the appearance sequence of the vehicle outline in the video frame. If the time difference between the sudden change of the geomagnetic signal and the vehicle entering the frame in the video is less than 2 seconds in at least two collected data, the parking space is determined to be occupied.

[0024] If all three acquisitions show that the magnetic field change is continuously lower than the preset threshold and there is no vehicle outline in the video frame, it is determined to be in an idle state;

[0025] If the status of the parking space is still unclear after the second verification, the parking space will be marked as pending calibration and the administrator will be notified for manual review.

[0026] Furthermore, the specific implementation of the vehicle identification module is as follows:

[0027] When a vehicle enters the sensing area, the deployed camera group collects all-round images of the vehicle and transmits them to the vehicle recognition server for processing. The overall morphological characteristics of the vehicle are analyzed and matched with the pre-built vehicle type knowledge base, and the vehicle type is judged in combination with visual identification. OCR technology is used to extract license plate characters, and the license plate type is determined based on the background color of the license plate. The obtained vehicle type, license plate characters and license plate type are integrated into the vehicle information of the current vehicle.

[0028] Furthermore, the specific implementation process of the intelligent guidance module includes:

[0029] When the target vehicle is at the entrance of the parking lot, the number of vacant parking spaces in the parking lot is obtained, a complete parking space status matrix is ​​constructed, and the matrix is ​​compared with the multi-dimensional preset threshold. If the number of vacant parking spaces is greater than the preset number of spaces, the target vehicle is allowed to enter the parking lot. Otherwise, the priority assessment mechanism is activated.

[0030] By the formula YX=10+up1·V t +up2·U f +up3·C i Get the priority coefficient, where V t is the vehicle type coefficient, which refers to the inherent benchmark value of the vehicle, including ambulance = 1.0, hospital service vehicle = 0.9, ordinary private car = 0.5 and large passenger car = 0.4; U f C is the emergency state coefficient, which is linked to the hospital HIS system, including critical illness = 1.0, emergency = 0.8, general outpatient = 0.5, and accompanying visit = 0.3; i is the reservation status coefficient, advance reservation = 0.8, no reservation = 0.4; up1, up2, up3 are the influencing weight factors of vehicle type coefficient, emergency status coefficient and reservation status coefficient respectively;

[0031] Based on the location of the recommended parking space and the user's vehicle location, multiple candidate routes are automatically generated, and the route preference value LJ of each candidate route is calculated. The optimal route value is obtained, where CD represents the actual route distance from the parking lot entrance to the recommended parking space, TX refers to the estimated time for the vehicle to travel along the route; YX refers to the vehicle priority coefficient, and YD refers to the vehicle density of the current candidate route. The number of vehicles per unit length is captured in real time by cameras deployed along the route; α, β, γ, and δ are the influence weight factors of CD, TX, YX, and YD, respectively.

[0032] The route preference values ​​of each candidate route are compared with each other, and the candidate route with the highest route preference value is marked as the preferred route.

[0033] Furthermore, the intelligent guidance module also includes:

[0034] Automatically determine whether the current time is peak or off-peak through the time period recognition mechanism, and provide differentiated guidance routes accordingly;

[0035] During peak hours, the weight of YD in the route preference calculation formula is increased by 30%. If the route distance is greater than the shortest route distance but the vehicle density is lower than the preset threshold, the channel will also be given priority recommendation. The entire traffic flow is reassessed every 90 seconds. When the vehicle density in a certain area exceeds the warning threshold, an alternative route is automatically planned for subsequent vehicles.

[0036] During off-peak hours, the shortest route strategy is adopted, and the distance coefficient weight is increased by 25%, providing users with the shortest physical route to the target parking space; combined with the patient's medical information, parking spaces within 100 meters of the elevator or entrance and exit are recommended.

[0037] Furthermore, the multiple guidance methods include:

[0038] Deploy LED displays at the intersection of main roads and elevator waiting areas in the parking lot to display the number of available parking spaces in each area in real time;

[0039] Ground guidance uses an array of embedded LED indicators installed on the roadway to form a dynamic light strip guidance. Based on the location of the recommended parking space, the LED lights on the corresponding route are lit;

[0040] Voice guidance: A voice broadcast device is set up at the turning node, and navigation information is automatically broadcast when the vehicle approaches;

[0041] Mobile navigation is achieved through a mobile phone APP. You can check real-time parking spaces through the APP. After selecting a parking space, the APP will display real-time navigation information and show different routes based on peak and off-peak hours.

[0042] Furthermore, the specific implementation of the user interaction module is as follows:

[0043] Users reserve parking spaces through mobile terminals or self-service devices. The system automatically recommends available spaces within 100 meters based on the location of the registered department, and dynamically adjusts the strategy based on the reservation period: during peak hours, it recommends routes that avoid congestion, while during off-peak hours, it defaults to the shortest route. After the reservation is completed, the parking space status is locked, triggering the intelligent lifting bollard to rise and the LED to display the reservation sign and the last license plate number. If the vehicle is not parked within the preset reserved time, the parking space will be released.

[0044] Users can view the parking space saturation heat map and predictions for the next 30 minutes in real time, and can select an area independently to automatically guide vehicles when they enter. When leaving, routes to parking spaces are planned based on real-time passenger flow: congestion-avoiding routes are recommended during peak hours, and the shortest route is the default during off-peak hours. Medical staff use their work badges to activate exclusive channel priority, and special patients can be provided with one-click reservations for barrier-free parking spaces.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] (1) The present invention uses a multimodal sensor array to collect parking space status in real time, and fuses the data of the geomagnetic sensor and the camera based on a data conflict processing mechanism. When a data conflict occurs, a secondary judgment is initiated. This effectively solves the problem of a single sensor being susceptible to interference and causing misjudgment, and significantly improves the accuracy and real-time performance of basic parking space status data.

[0047] (2) The present invention dynamically calculates the optimal route based on parking space status and vehicle information, triggers automatic adjustment of route strategies through real-time traffic evaluation, increases the congestion coefficient weight during peak hours to generate congestion-avoiding routes, and increases the distance weight coefficient during off-peak hours to generate direct routes; the user end achieves full-process optimization and guidance through a reservation mechanism, a visual parking space heat map, and a variety of guidance methods, significantly improving traffic efficiency in complex traffic environments and increasing the convenience of special vehicles and users. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0049] Figure 1 This is the overall system block diagram of the present invention. DETAILED DESCRIPTION

[0050] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0051] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0052] It should also be understood that the terminology used in this disclosure is for the purpose of describing specific embodiments only and is not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should be further understood that the term "and / or" as used in this disclosure and the claims refers to any and all possible combinations of one or more of the associated listed items, including and including these combinations.

[0053] like Figure 1 As shown, a smart parking guidance system for hospital parking lots includes a parking space monitoring module, a vehicle identification module, an intelligent guidance module and a user interaction module.

[0054] The parking space monitoring module uses a multimodal sensor array to monitor the parking space status data in the hospital parking lot in real time. The specific monitoring process includes:

[0055] The multimodal sensor array includes a geomagnetic sensor group and a camera array. The geomagnetic sensor is installed at the center of each parking space. It continuously collects changes in magnetic field strength to generate dynamic monitoring signals, providing real-time information on the current parking space status, including occupied, free, and reserved.

[0056] When a vehicle enters a parking space, the metal structure of the vehicle will change the value of the magnetic field strength in the parking space area. When the magnetic sensor detects that the magnetic field strength change exceeds the preset threshold, the parking space is determined to be occupied; when the magnetic sensor detects that the magnetic field change is lower than the preset threshold, the parking space is determined to be vacant.

[0057] As an auxiliary means of monitoring parking space status, wide-angle cameras are deployed in various functional areas of the hospital parking lot. Each camera can cover a bird's-eye view of 6 to 8 parking spaces. They are used to collect video streams and perform real-time analysis. The specific analysis process is as follows:

[0058] By pre-setting physical markers at the four corners of the parking space, the camera automatically establishes a mapping relationship between pixel coordinates and real-world coordinates through a calibration algorithm, and generates a virtual parking network. Each parking space is divided into a rectangular area of ​​standard pixels and associated with a unique number. Each parking space is accurately located through real-world coordinate mapping.

[0059] Based on the improved YOLOv5 target detection network, the detected vehicle outline is subjected to edge extraction and contour fitting. When the vehicle outline area occupies more than 40% of the rectangular area, the occupancy value P1 is output;

[0060] In the idle state, the parking space surface texture features are pre-collected as a reference template, and the similarity between the current texture features and the reference template is compared in real time. When the texture feature similarity is lower than the preset threshold, it is determined that the parking space surface texture has changed significantly and the texture change value P2 is output;

[0061] An improved KCF tracker is used to track vehicle motion trajectories. The target vehicle's center of mass position change is calculated at a frame rate of 30fps to generate a motion trajectory sequence. A parking area activity window is set. When a vehicle is detected entering the parking area and its speed drops from an initial value to near zero within 20 consecutive frames, it is considered a parking behavior. When a vehicle is detected starting to move from a stationary state and leaving the parking area within 15 consecutive frames, it is considered a departure behavior and the occupancy probability value P3 is output.

[0062] After normalizing the occupancy value P1, texture change value P2, and occupancy probability value P3, enter them into the formula S=a1·P1+a2·P2+a 3· P3 calculates the decision value, where a1, a2, and a3 are the influence weight factors of occupancy value, texture change value, and occupancy probability value respectively;

[0063] When the decision value is greater than the preset threshold and the vehicle outline is detected in 5 consecutive frames, the texture feature matching is successful, and the target stays in the parking space for more than 2 seconds, it is determined to be in an occupied state; when the decision value is less than the preset threshold and no valid vehicle target is detected in the rectangular area for 30 consecutive frames or the target motion tracking shows that the vehicle has left the parking space, it is determined to be in an idle state;

[0064] The data from the geomagnetic sensor and the camera are fused based on the data conflict handling mechanism;

[0065] When the results of the geomagnetic sensor and the camera monitoring are consistent, the current parking space status is directly determined and updated;

[0066] When the judgment results of the geomagnetic sensor and the camera monitoring are inconsistent, a secondary judgment is performed and the time window secondary verification mechanism is activated;

[0067] Trigger three data acquisitions at 5-second intervals within 15 seconds, synchronously acquiring the magnetic field intensity change sequence of the geomagnetic sensor and the video frame of the parking area from the camera. Use a dynamic time warping algorithm to compare the magnetic field change curve with the timing of the appearance of the vehicle outline in the video frame. If the time difference between the sudden change of the geomagnetic signal and the vehicle entering the scene in the video is less than 2 seconds in at least two acquisitions, the parking space is determined to be occupied. If all three acquisitions show that the magnetic field change is continuously below the preset threshold and there is no vehicle outline in the video frame, the parking space is determined to be idle.

[0068] If a clear conclusion about the parking space status cannot be drawn after the second verification, the parking space will be automatically marked as pending calibration and the administrator will be notified for manual review.

[0069] The vehicle identification module deploys cameras at each entrance and exit of the hospital parking lot to identify and record vehicle information at the entrance and exit. The specific process is as follows:

[0070] When a vehicle enters the sensing area, all-round images of the vehicle are collected by cameras deployed at the entrances and exits. The images are then transmitted to the vehicle identification server for processing. The overall morphological features of the vehicle are analyzed and matched with a pre-built vehicle type knowledge base. The vehicle type is determined by combining visual identification. Vehicle types include general vehicles and special vehicles. General vehicles include ordinary private cars, taxis, online ride-hailing vehicles, and barrier-free vehicles. Special vehicles include ambulances, fire trucks, police cars, and hospital service vehicles. OCR technology is used to extract license plate characters, and the license plate type is determined based on the background color of the license plate. The license plate types include ordinary blue plates, yellow plates, and new energy green plates. The vehicle type, license plate characters, and license plate type obtained at the same time are integrated into the vehicle information of the current vehicle.

[0071] The intelligent guidance module dynamically calculates the route preference value based on parking space status information and vehicle identification information, obtains the optimal route, and provides the driver with full navigation services from the entrance to the target parking space through various guidance methods. The specific process includes:

[0072] When the target vehicle is at the entrance of the parking lot, the target vehicle refers to the vehicle whose license plate number has been scanned. The number of vacant parking spaces in the parking lot is obtained. Based on the real-time collection and aggregation of sensor data from all areas of the entire site, a complete parking space status matrix is ​​constructed to accurately count the number and distribution of vacant parking spaces of various types, and perform real-time comparison and analysis with multi-dimensional preset thresholds. When the number of vacant parking spaces is greater than the preset number of parking spaces, the target vehicle is allowed to enter the parking lot, and the vehicle information of the target vehicle is received. Based on the vehicle information and the location of the vacant parking spaces, a recommended parking space is generated. When the number of vacant parking spaces is less than the preset number of parking spaces, the priority of the vehicle is determined through the priority assessment mechanism to ensure that special vehicles obtain priority right of way and exclusive parking space resources.

[0073] By the formula YX=10+up1·V t +up2·U f +up3·C i Get the priority coefficient, where V t is the vehicle type coefficient, which refers to the inherent benchmark value of the vehicle, including ambulance = 1.0, hospital service vehicle = 0.9, ordinary private car = 0.5 and large passenger car = 0.4; U f C is the emergency state coefficient, which is linked to the hospital HIS system, including critical illness = 1.0, emergency = 0.8, general outpatient = 0.5, and accompanying visit = 0.3; i is the reservation status coefficient, with pre-reserved parking spaces = 0.8 and unreserved = 0.4; up1, up2, and up3 are the influence weight factors of vehicle type coefficient, emergency status coefficient, and reservation status coefficient, respectively; when the priority coefficient is lower than the preset threshold, the vehicle enters the waiting queue; when the priority coefficient is higher than the preset threshold, the exclusive channel right of way is activated, and the nearest vacant parking space to the vehicle's terminal position is locked;

[0074] Based on the location of the recommended parking space and the user's vehicle location, multiple candidate routes are automatically generated, and the route preference value LJ of each candidate route is calculated. The route optimization value is calculated, where CD represents the actual route distance from the parking lot entrance to the recommended parking space, TX refers to the estimated time it takes for a vehicle to travel along the route, YX refers to the vehicle priority coefficient, and YD refers to the vehicle density of the current candidate route. The number of vehicles per unit length is captured in real time by cameras deployed along the route. α, β, γ, and δ are the influence weight factors of CD, TX, YX, and YD, respectively.

[0075] Compare the route preference values ​​of each candidate route with each other, mark the candidate route with the highest route preference value as the optimal route based on the comparison results, and recommend it to the user;

[0076] Automatically determine whether the current time is peak or off-peak through the time period recognition mechanism, and provide differentiated guidance routes accordingly;

[0077] During peak hours (8:00-10:30 a.m. and 1:30-4:00 p.m. on weekdays, and 9:30-11:30 a.m. on holidays), priority is given to diversion factors and congestion prevention strategies are implemented. The YD weight in the route optimization calculation formula is increased by 30%. Channels with route distances greater than the shortest route distance but vehicle density below a preset threshold are also prioritized. Regional heat maps on electronic maps visually display congestion conditions on each floor. The entire venue is reassessed every 90 seconds. When the vehicle density in a certain area exceeds the warning threshold, alternative routes are automatically planned for subsequent vehicles to prevent the spread of congestion.

[0078] During off-peak hours, such as 10:30 AM to 1:30 PM on weekdays, after 4:00 PM, after 11:30 AM on holidays, and before 8:00 AM on weekdays, we prioritize convenience and adopt a shortest route strategy, increasing the distance coefficient by 25% to provide users with the shortest physical route to their target parking space. Furthermore, based on patient information, we recommend parking spaces less than 100 meters from elevators or entrances to improve traffic efficiency.

[0079] Intelligent guidance adopts a combination of multiple guidance methods, including:

[0080] High-brightness LED information displays are deployed at the intersections of major roads and in elevator waiting areas of the parking lot, displaying the number of available parking spaces in each area in real time. These displays include floor information, area number, and the number of available spaces. Each display integrates three layers of data visualization: the top layer displays the current floor's numerical code and direction identifier; the middle layer dynamically updates the number of available spaces in the corresponding area by area number, using a red and green background to distinguish between occupied and idle spaces; and the bottom layer displays arrows indicating the optimal route direction and estimated time. Using an industrial-grade communication protocol, all data is synchronously updated every five seconds to ensure that the displayed content at each node is consistent with the parking lot status database.

[0081] Ground guidance uses an array of embedded LED indicators installed on the roadway to form a dynamic light guide strip. When a vehicle enters the parking lot, the LED lights on the corresponding route are illuminated based on the recommended parking space, forming a flowing light strip effect to guide the vehicle forward. The LED lights use RGB full-color design and can display different colors: green indicates the direction of travel, red indicates no passage, and yellow indicates slow down.

[0082] Voice guidance: A voice broadcast device is set up at the turning node, and navigation information is automatically broadcast when the vehicle approaches. The voice content includes turn prompts, floor prompts and parking location prompts.

[0083] Mobile navigation is provided through a mobile phone app. Before entering the parking lot, drivers can use the app to view real-time parking information and select their target parking space. Once inside, the app displays a real-time navigation map, providing turn and distance prompts. The app also supports AR navigation, which uses the phone's camera to scan the surrounding environment and overlay navigation arrows and prompts on the screen. The app interface automatically switches navigation strategy display modes based on peak and off-peak times, highlighting less congested routes during peak times and the shortest, optimal routes during off-peak times, and providing estimated arrival time comparison information.

[0084] The user interaction module provides multi-channel information interaction services for hospital visitors, patients, and medical staff, realizing intelligent management of the entire parking process, including:

[0085] Users reserve parking spaces through mobile terminals or self-service devices in the hospital, and the hospital will automatically recommend parking spaces close to the clinic based on the location of the registration department; the recommendation strategy will be automatically adjusted based on whether the user's appointment time is peak period. During peak periods, priority will be given to parking spaces in dispersed areas with smooth traffic, and during off-peak periods, priority will be given to convenient parking spaces with the closest distance; when the user completes the appointment, the status of the reserved parking space will be automatically changed to the reservation locked state, and this information will be synchronized to the parking space monitoring module to form a reservation locked mark, and the pre-installed smart lifting bollard will be raised, and the parking space LED will display the reservation mark and the last license plate number; if the user fails to park the vehicle within the preset reserved time, when the user's appointment time arrives, the user will be reminded to complete the parking within the reserved time through APP or SMS. If the user fails to park the vehicle within the preset reserved time, the user's appointment will be cancelled, and the parking space status will be changed from the reservation locked state to the idle state, and the smart lifting bollard will be lowered at the same time;

[0086] Users can use mobile devices to view the parking space saturation and traffic conditions in each area of ​​the parking lot in real time. A heat map is used to visually display the current parking space distribution during peak and off-peak periods, and a prediction of parking space changes within the next 30 minutes is provided. Users can independently select relatively empty parking areas based on the heat map information. The user's selection is recorded and integrated into the vehicle identification module, allowing vehicles to receive guidance services when they arrive at the entrance without having to make a second selection.

[0087] After completing their medical consultation, users can check their vehicle's location via mobile devices or self-service devices within the hospital. Based on the user's current location and real-time traffic data, the optimal route to the parking space is automatically planned, avoiding construction areas and congested passages in real time. During peak hours, users are prompted to choose a slightly longer route that avoids crowded areas. During off-peak hours, a direct route with the shortest physical distance is recommended by default. Users can view a comparison of the two routes on their mobile devices and switch between them freely.

[0088] Supporting multi-terminal collaboration, users can complete parking reservations on the mobile app and synchronize real-time guidance information to the in-car navigation or smartwatch. After entering the parking lot, users can issue interactive instructions through voice interaction, including "Find My Parking Space" and "Go to the Nearest Exit." Differentiated voice guidance is provided based on the current peak and low traffic conditions.

[0089] To improve the work efficiency of medical staff, an exclusive service channel for medical staff is set up. Medical staff can automatically obtain exclusive parking area information through work badge recognition and obtain priority access during peak hours. For special outpatient patients and people with mobility difficulties, a one-click reservation function for barrier-free parking spaces is provided, and relevant information is automatically pushed to the parking space monitoring module and intelligent guidance module to ensure that these special parking spaces can be reasonably reserved even during peak hours.

[0090] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A smart parking guidance system for hospital parking lots, characterized by: include: The parking space monitoring module uses a multimodal sensor array to monitor the parking space status data in the hospital parking lot in real time; The specific implementation process includes: The multimodal sensor array includes a geomagnetic sensor group and a camera array. The geomagnetic sensor continuously collects the change in magnetic field strength. When the change in magnetic field strength exceeds a preset threshold, the parking space is determined to be occupied. Otherwise, it is determined to be vacant. The vehicle identification module uses cameras deployed at each entrance and exit of the hospital parking lot to identify and record vehicle information at the entrance and exit; The intelligent guidance module dynamically calculates the optimal route based on parking space status information and vehicle identification information, and provides drivers with full navigation services from the entrance to the target parking space through various guidance methods; The user interaction module is used to provide multi-channel information interaction services for hospital visitors, patients and medical staff, and realize intelligent management of the entire parking process.

2. The intelligent parking guidance system for hospital parking lots according to claim 1 is characterized in that: The specific implementation process of the camera array monitoring the parking space status includes: By presetting physical marking points at the four corners of the parking space, a mapping relationship between pixel coordinates and actual space coordinates is established based on a calibration algorithm, generating a virtual network of rectangular parking spaces bound with unique numbers; The vehicle outline is extracted and calculated based on the improved YOLOv5 target detection network. When the outline area occupies more than 40% of the rectangular area, the occupancy value P1 is output; In the idle state, the surface texture features of the parking space are pre-collected as a reference template, and the similarity between the current texture features and the reference template is compared in real time; when the texture feature similarity is lower than the preset threshold, it is determined that the surface texture of the parking space has changed significantly and the texture change value P2 is output; An improved KCF tracker is used to track vehicle motion trajectories and generate motion trajectory sequences. When a vehicle is detected entering a parking area and its speed drops from an initial value to near zero within 20 consecutive frames, it is considered a parking behavior. When a vehicle is detected starting to move from a stationary state and leaving the parking area within 15 consecutive frames, it is considered a leaving behavior. The occupancy probability value P3 is output. After normalizing the occupancy value P1, texture change value P2, and occupancy probability value P3, enter them into the formula S=a1·P1+a2·P2+a 3· P3 obtains the decision value, where a1, a2, and a3 are the influence weight factors of occupancy value, texture change value, and occupancy probability value, respectively; When the decision value is greater than the preset threshold and the vehicle outline is detected for 5 consecutive frames, the texture feature matching is successful, and the target stays in the parking space for more than 2 seconds, it is judged to be in an occupied state; when the decision value is less than the preset threshold and no valid vehicle target is detected in the rectangular area for 30 consecutive frames or the vehicle is shown to have left the parking space, it is judged to be in an idle state.

3. The intelligent parking guidance system for hospital parking lots according to claim 2 is characterized in that: The specific implementation process of the parking space monitoring module also includes: The data from the geomagnetic sensor and the camera are fused based on the data conflict handling mechanism; When the results of the geomagnetic sensor and the camera monitoring are consistent, the current parking space status is directly determined and updated; otherwise, a secondary judgment is made and the time window secondary verification mechanism is activated; Trigger data collection three times at 5-second intervals within 15 seconds to synchronously obtain the magnetic field intensity change sequence and the camera's parking area video frame; The dynamic time warping algorithm compares the magnetic field change curve with the appearance sequence of the vehicle outline in the video frame. If the time difference between the sudden change of the geomagnetic signal and the vehicle entering the frame in the video is less than 2 seconds in at least two collected data, the parking space is determined to be occupied. If all three acquisitions show that the magnetic field change is continuously lower than the preset threshold and there is no vehicle outline in the video frame, it is determined to be in an idle state; If the status of the parking space is still unclear after the second verification, the parking space will be marked as pending calibration and the administrator will be notified for manual review.

4. The intelligent parking guidance system for hospital parking lots according to claim 1 is characterized in that: The specific implementation of the vehicle identification module is as follows: When a vehicle enters the sensing area, the deployed camera group collects all-round images of the vehicle and transmits them to the vehicle recognition server for processing. The overall morphological characteristics of the vehicle are analyzed and matched with the pre-built vehicle type knowledge base, and the vehicle type is judged in combination with visual identification. OCR technology is used to extract license plate characters, and the license plate type is determined based on the background color of the license plate. The obtained vehicle type, license plate characters and license plate type are integrated into the vehicle information of the current vehicle.

5. The intelligent parking guidance system for hospital parking lots according to claim 1 is characterized in that: The specific implementation process of the intelligent guidance module includes: When the target vehicle is at the entrance of the parking lot, the number of vacant parking spaces in the parking lot is obtained, a complete parking space status matrix is ​​constructed, and the matrix is ​​compared with the multi-dimensional preset threshold. If the number of vacant parking spaces is greater than the preset number of spaces, the target vehicle is allowed to enter the parking lot. Otherwise, the priority assessment mechanism is activated. By the formula YX=10+up1·V t +up2·U f +up3·C i Get the priority coefficient, where V t is the vehicle type coefficient, which refers to the inherent benchmark value of the vehicle, including ambulance = 1.0, hospital service vehicle = 0.9, ordinary private car = 0.5 and large passenger car = 0.4; U f C is the emergency state coefficient, which is linked to the hospital HIS system, including critical illness = 1.0, emergency = 0.8, general outpatient = 0.5, and accompanying visit = 0.3; i is the reservation status coefficient, advance reservation = 0.8, no reservation = 0.4; up1, up2, up3 are the influencing weight factors of vehicle type coefficient, emergency status coefficient and reservation status coefficient respectively; Based on the location of the recommended parking space and the user's vehicle location, multiple candidate routes are automatically generated, and the route preference value LJ of each candidate route is calculated. The optimal route value is obtained, where CD represents the actual route distance from the parking lot entrance to the recommended parking space, TX refers to the estimated time for the vehicle to travel along the route; YX refers to the vehicle priority coefficient, and YD refers to the vehicle density of the current candidate route. The number of vehicles per unit length is captured in real time by cameras deployed along the route; α, β, γ, and δ are the influence weight factors of CD, TX, YX, and YD, respectively. The route preference values ​​of each candidate route are compared with each other, and the candidate route with the highest route preference value is marked as the preferred route.

6. The intelligent parking guidance system for hospital parking lots according to claim 5 is characterized in that: The intelligent guidance module also includes: Automatically determine whether the current time is peak or off-peak through the time period recognition mechanism, and provide differentiated guidance routes accordingly; During peak hours, the weight of YD in the route preference calculation formula is increased by 30%. If the route distance is greater than the shortest route distance but the vehicle density is lower than the preset threshold, the channel will also be given priority recommendation. The entire traffic flow is reassessed every 90 seconds. When the vehicle density in a certain area exceeds the warning threshold, an alternative route is automatically planned for subsequent vehicles. During off-peak hours, the shortest route strategy is adopted, and the distance coefficient weight is increased by 25%, providing users with the shortest physical route to the target parking space; combined with the patient's medical information, parking spaces within 100 meters of the elevator or entrance and exit are recommended.

7. The intelligent parking guidance system for hospital parking lots according to claim 1 is characterized in that: The multiple guidance methods include: Deploy LED displays at the intersection of main roads and elevator waiting areas in the parking lot to display the number of available parking spaces in each area in real time; Ground guidance uses an array of embedded LED indicators installed on the roadway to form a dynamic light strip guidance. Based on the location of the recommended parking space, the LED lights on the corresponding route are lit; Voice guidance: A voice broadcast device is set up at the turning node, and navigation information is automatically broadcast when the vehicle approaches; Mobile navigation is achieved through a mobile phone APP. You can check real-time parking spaces through the APP. After selecting a parking space, the APP will display real-time navigation information and show different routes based on peak and off-peak hours.

8. The intelligent parking guidance system for hospital parking lots according to claim 1 is characterized in that: The specific implementation of the user interaction module is as follows: Users reserve parking spaces through mobile terminals or self-service devices. The system automatically recommends available parking spaces within 100 meters based on the location of the department where they are registered. The system also dynamically adjusts the parking strategy based on the reservation period: during peak hours, it recommends routes that avoid congestion, while during off-peak hours, it defaults to the shortest route. After the reservation is completed, the parking space status is locked, the intelligent lifting column is triggered to rise, and the LED display of the reservation logo and the license plate number is displayed; If the vehicle is not parked within the preset reserved time, the parking space will be released; Users can view a real-time heat map of parking space saturation and a 30-minute forecast. They can then select their own area and be automatically guided when entering the parking lot. When leaving the parking lot, routes to parking spaces are planned based on real-time passenger flow: during peak hours, congestion-avoiding routes are recommended, while during off-peak hours, the shortest route is the default. Medical staff use their ID cards to activate exclusive channel priority, and provide one-click reservations for barrier-free parking spaces for patients with special needs.

Images