A method and system for detecting the presence of a vehicle in situ

Abstract

This invention discloses an intelligent vehicle presence detection method and system, specifically relating to the field of intelligent traffic control technology. It addresses the problem that existing parking lot vehicle presence detection technologies fail to recognize continuous behaviors such as vehicles briefly leaving and returning, leading to interruptions in parking timekeeping and errors in billing and management rule execution. The method is achieved through the following technical solution: acquiring and analyzing vehicle presence detection data to determine departure events; analyzing and distinguishing the types of departure events; initiating short-term analysis timing with different durations based on the type; monitoring entry events during the timing period and analyzing whether the status changes of preset adjacent parking spaces before and after departure and entry events meet correlation conditions; when the correlation conditions and vehicle identity conditions are met, associating the two events and correcting the vehicle presence duration of the parking space; thus achieving intelligent identification and accurate timing of continuous parking behavior, improving the accuracy and intelligence level of the parking lot management system.

Description

Technical Field

[0001] This invention relates to the field of intelligent traffic control technology, and more specifically, to an intelligent detection method and system for vehicle presence. Background Technology

[0002] In the field of intelligent parking management, vehicle presence detection based on visual recognition or various sensors can achieve refined management of vehicle occupancy status in different functional areas, such as fixed parking spaces, temporary loading and unloading areas, and time-limited parking zones. By deploying detection nodes near parking spaces or areas, the arrival and departure events of vehicles can be sensed in real time, and the judgment results can be uploaded to the management platform. This provides basic status information for services such as parking guidance, billing and timing, and violation monitoring. Existing technical solutions mainly focus on improving the accuracy and real-time performance of determining the instantaneous physical state of a vehicle's presence or departure by optimizing sensor performance, improving recognition algorithms, or fusing multi-source data.

[0003] However, existing detection technologies generally focus on the instantaneous and isolated judgment of a vehicle's physical state. When a vehicle briefly leaves an area with clear time management rules, such as a loading / unloading area or a temporary parking area, and then returns to its original or nearby location due to yielding or repositioning, the system records this as an independent departure event and a new presence event based on continuous instantaneous state judgments. This detection logic based on purely physical state switching cannot technically understand and associate the brief departure behavior as an inherent component of the same continuous parking process. This leads to logical execution errors in subsequent core management rules such as billing and timeout determination based on accumulated state time, which are inconsistent with the actual parking behavior, constituting a disconnect between the detection results and high-level management semantics. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent detection method and system for vehicle presence to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for intelligent detection of vehicle presence includes the following steps:

[0007] S1. Obtain vehicle presence detection data that is sensed and uploaded in real time by detection nodes deployed in the parking area;

[0008] S2. Based on the vehicle presence detection data, determine whether a departure event targeting the same parking space has occurred;

[0009] S3. When it is determined that a departure event has occurred, analyze and distinguish the type of departure event, which includes temporary movement or final departure.

[0010] S4. When the departure event is classified as temporary relocation, a first short-term analysis timer is started for the parking space; when it is classified as final departure, a second short-term analysis timer is started, and the duration of the first short-term analysis timer is shorter than that of the second short-term analysis timer.

[0011] S5. Monitor whether a parking space entry event occurs within the corresponding duration. If an entry event is detected, analyze the state change information of at least one preset adjacent parking space corresponding to the parking space recorded when the exit event occurs, and whether the state change information of at least one adjacent parking space before and after the entry event meets the preset correlation conditions.

[0012] S6. When the association conditions are met and the vehicle corresponding to the entry event and the vehicle corresponding to the exit event meet the vehicle identity determination conditions, the entry event and the exit event are associated and the vehicle in the parking space is corrected for the duration of the vehicle in the parking space.

[0013] Furthermore, S1 includes:

[0014] A geomagnetic sensor is independently deployed in each parking space to sense changes in the magnetic field of the corresponding parking space in real time and generate magnetic field change data.

[0015] Deploy at least one camera in a public area covering multiple parking spaces to collect video image data containing the corresponding parking spaces;

[0016] Vehicle in-situ detection data includes magnetic field change data and video image data.

[0017] Furthermore, S2 includes:

[0018] Based on the magnetic field change data of the same parking space, determine whether the duration of the magnetic field strength recovering from the disturbance state representing the presence of the vehicle to the baseline state exceeds the first threshold.

[0019] When the first threshold is exceeded, the analysis of the parking space video image data is triggered;

[0020] Analyze video image data to identify whether the vehicle outline disappears from the corresponding parking space area and remains for a second threshold duration;

[0021] When the vehicle outline disappears for a duration that exceeds the second threshold, it is determined that a vehicle has left the parking space.

[0022] Furthermore, S3 includes:

[0023] Acquire video image data within a first preset time period before the departure event occurs and within a second preset time period after the event occurs;

[0024] Based on video image data, analyze whether the vehicle's departure speed exceeds a speed threshold during the departure process;

[0025] At the same time, it was analyzed whether the parking characteristics of the vehicle before the departure incident met the complete parking conditions;

[0026] When the departure speed exceeds the speed threshold and the parking characteristics do not meet the complete parking conditions, the departure event is classified as a temporary movement.

[0027] Otherwise, the departure event will be classified as a final departure.

[0028] Furthermore, whether the parking characteristics meet the complete parking conditions is determined by the following methods: extracting the time period from when the vehicle comes to a complete stop to when the door is opened from the video image data within a first preset time period before the departure event occurs; determining whether the time period lasts for a third preset time period; and determining whether the vehicle outline remains stationary during the time period.

[0029] Furthermore, S4 includes:

[0030] Configure a timer with a first time window as the first short-time analysis timer for temporary relocation type departure events;

[0031] Configure a timer with a second time window as a second short-time analysis timer for departure events of the final departure type;

[0032] The duration of the first time window is shorter than the duration of the second time window;

[0033] After distinguishing the type of departure event, the corresponding timer is immediately activated to start counting.

[0034] Furthermore, S5 includes:

[0035] Based on the geographical location of the parking spaces, at least one adjacent parking space is pre-assigned;

[0036] When a departure event occurs, record the first state change sequence of at least one adjacent parking space changing from occupied to vacant within a first preset time period, starting from the corresponding time.

[0037] When an entry event is detected, a second state change sequence is recorded, which is the sequence of at least one adjacent parking space changing from an idle state to an occupied state within a second preset time period, ending at the time of the entry event.

[0038] Determine whether the sequence of changes in the first state and the sequence of changes in the second state are opposite in chronological order and logically correspond to each other, in order to determine whether the association condition is met.

[0039] Furthermore, based on the geographical location of the parking spaces, at least one adjacent parking space is pre-defined in the following way: based on the driving direction of the parking lane, an adjacent parking space located in front of the target parking space in the direction of exit, and another adjacent parking space located to the side of the target parking space and sharing the same lane exit are jointly defined as at least one adjacent parking space.

[0040] Furthermore, S6 includes:

[0041] When both the departure event and the entry event meet the association conditions and the vehicle identity determination conditions, the departure event is marked as an invalid departure event;

[0042] Mark the entry event as an invalid in-place event;

[0043] Based on the actual occurrence time of the departure event and the actual occurrence time of the entry event, calculate the actual continuous occupancy time of the parking space between the marked invalid departure event and invalid in-space event;

[0044] The actual continuous occupancy time is added to the duration of the vehicle that was in the parking space first.

[0045] On the other hand, the present invention provides an intelligent vehicle presence detection system, comprising the following modules:

[0046] The data acquisition module is used to acquire vehicle presence detection data that is sensed and uploaded in real time by detection nodes deployed in the parking area;

[0047] The event judgment module is used to determine whether a departure event targeting the same parking space has occurred based on vehicle presence detection data.

[0048] The type differentiation module is used to analyze and differentiate the type of departure event when it is determined that a departure event has occurred. The types include temporary movement or final departure.

[0049] The timing start module is used to start a first short-term analysis timer for the parking space when the departure event is classified as temporary movement; and to start a second short-term analysis timer when it is classified as final departure. The duration of the first short-term analysis timer is shorter than that of the second short-term analysis timer.

[0050] The correlation analysis module is used to monitor whether a parking space entry event occurs within the corresponding duration. If an entry event is detected, the module analyzes the state change information of at least one preset adjacent parking space corresponding to the parking space recorded when the exit event occurs, and whether the state change information of at least one adjacent parking space before and after the entry event meets the preset correlation conditions.

[0051] The association correction module is used to associate the entry event and the exit event and correct the vehicle's in-place duration in the parking space when the association conditions are met and the vehicle corresponding to the entry event and the vehicle corresponding to the exit event meet the vehicle identity determination conditions.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. By introducing behavioral intent analysis of departure events, differentiated monitoring strategies, and multi-dimensional correlation verification mechanisms, intelligent understanding of vehicle presence status and continuous behavior determination are achieved. This effectively identifies the behavior of vehicles briefly leaving and returning for purposes such as yielding or repositioning, and distinguishes it from true final departures. By logically linking physically separate departure and entry events as the same continuous parking process, the problems of parking time interruption and management logic errors caused by misjudgment of status are solved. This enables management functions such as parking space occupancy time statistics, billing, and overtime determination to truly reflect the actual usage behavior of vehicles, significantly improving the intelligence level and decision-making accuracy of the parking management system.

[0054] 2. By predicting the type of departure events (temporary movement or final departure), key behavioral semantic information is provided for subsequent processing. Based on this prediction result, differentiated short-term analysis timing is initiated, achieving optimized allocation of monitoring resources. This ensures rapid response to vehicles that may return while avoiding unnecessary waiting for long-term departures. The sequence of adjacent parking space status changes is introduced as evidence of behavioral correlation. By analyzing the symmetrical changes in the occupancy status of adjacent parking spaces before and after departure and entry, strong spatial logical evidence is provided that departure-return is the same continuous interactive behavior. This forms a complete logical closed loop from perception, understanding, prediction to verification, enabling the system to have behavioral cognitive capabilities beyond instantaneous state detection. Ultimately, it achieves accurate and coherent calculation of the duration of vehicle presence, ensuring the fair and correct execution of parking management rules. Attached Figure Description

[0055] Figure 1 This is a flowchart of an intelligent vehicle presence detection method according to the present invention;

[0056] Figure 2 This is a schematic diagram of the structure of an intelligent vehicle presence detection system according to the present invention. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0058] Example 1: Figure 1 The present invention provides an intelligent detection method for vehicle presence, which includes the following steps:

[0059] S1. Obtain vehicle presence detection data that is sensed and uploaded in real time by detection nodes deployed in the parking area;

[0060] S2. Based on the vehicle presence detection data, determine whether a departure event targeting the same parking space has occurred;

[0061] S3. When it is determined that a departure event has occurred, analyze and distinguish the type of departure event, which includes temporary movement or final departure.

[0062] S4. When the departure event is classified as temporary relocation, a first short-term analysis timer is started for the parking space; when it is classified as final departure, a second short-term analysis timer is started, and the duration of the first short-term analysis timer is shorter than that of the second short-term analysis timer.

[0063] S5. Monitor whether a parking space entry event occurs within the corresponding duration. If an entry event is detected, analyze the state change information of at least one preset adjacent parking space corresponding to the parking space recorded when the exit event occurs, and whether the state change information of at least one adjacent parking space before and after the entry event meets the preset correlation conditions.

[0064] S6. When the association conditions are met and the vehicle corresponding to the entry event and the vehicle corresponding to the exit event meet the vehicle identity determination conditions, the entry event and the exit event are associated and the vehicle in the parking space is corrected for the duration of the vehicle in the parking space.

[0065] S1. Obtain vehicle presence detection data that is sensed and uploaded in real time by detection nodes deployed in the parking area. The specific implementation is as follows:

[0066] A geomagnetic sensor is fixedly installed at the center of the ground in each parking space. The sensor's sensitive axis is perpendicular to the ground to sense the local Earth's magnetic field strength at that location. The geomagnetic sensor continuously measures the magnetic field strength data at a preset sampling frequency, for example, 10 times per second. When the parking space is vacant (no vehicle is parked), the magnetic field strength measured by the sensor fluctuates within a stable range; this fluctuation range is recorded and set as the magnetic field reference state for that parking space. Magnetic field change data refers to the real-time, timestamped sequence of magnetic field strength values ​​measured and output by the geomagnetic sensor. When a vehicle enters and parks in the space, the vehicle's metal body disturbs the local Earth's magnetic field, causing a significant and continuous change in the magnetic field strength value measured by the sensor. This changed state is defined as the disturbance state characterizing the vehicle's presence. The geomagnetic sensor transmits its generated, timestamped sequence of magnetic field strength values ​​as magnetic field change data via wired or wireless communication links.

[0067] In the public area of ​​the parking lot, installation locations are chosen such that a single camera's field of view can simultaneously cover at least two and no more than six parking spaces. The camera is installed 3 to 5 meters above the ground, with its tilt angle adjusted so that its center line of view covers the central area of ​​each parking space it is assigned to monitor. The camera uses a fixed-focus lens with a resolution of 1920 x 1080 pixels and a video frame rate of 25 frames per second. The camera continuously captures a video stream of the monitored area, which is encoded into video image data containing consecutive image frames. Each image frame is accompanied by a timestamp generated by the system clock. The video image data is transmitted in real-time via a wired network and then processed. Using camera calibration technology and the geometric markings of the parking spaces on the ground, a polygonal area corresponding to each actual parking space is delineated in the image frame; this polygonal area is the designated parking space area for that space.

[0068] Vehicle presence detection data comprises the aforementioned magnetic field change data and video image data. It receives magnetic field change data streams from various magnetic sensors and video image data streams from various cameras. To ensure data time consistency, a time synchronization protocol is used to synchronize the clocks of all connected data sources, ensuring that the timestamps carried by the magnetic field change data and video image data from the same or related parking spaces are based on a unified clock reference. Magnetic field change data from the same physical parking space is associated with the image data of the corresponding parking space area in the video image data through a preset parking space identification code. This parking space identification code is written into the configuration of the geomagnetic sensor during installation and corresponds to the unique number of the parking space area designated for that parking space in the video analysis system. Therefore, for any specific parking space, its vehicle presence detection data refers to the time-synchronized sequence of magnetic field change data and video image data bound to that parking space identification code.

[0069] S2. Based on vehicle presence detection data, determine whether a departure event targeting the same parking space has occurred. The specific implementation is as follows:

[0070] Vehicle presence detection data from a specific parking space is continuously received. This data includes magnetic field change data and video image data associated with that parking space. The process of determining a departure event based on the magnetic field change data of the same parking space is as follows: The magnetic field strength numerical sequence in the magnetic field change data is analyzed in real time; this sequence represents the measured magnetic field strength values ​​that change over time. It is necessary to identify whether the magnetic field strength has transitioned from a disturbance state representing the presence of a vehicle to a magnetic field reference state. This identification is achieved by calculating the difference between the magnetic field strength value and the pre-learned and stored average value of the magnetic field reference state for that parking space. When the absolute value of the difference between multiple consecutive sampling points is less than a preset disturbance judgment threshold (e.g., the absolute value of the difference between five consecutive sampling points is less than 200 nanoteslas), the magnetic field strength is determined to have entered the magnetic field reference state. The timestamp corresponding to the first sampling point that indicates entry into the magnetic field reference state is recorded. Next, it is necessary to determine the moment when the state transition begins, i.e., the moment when the magnetic field strength leaves the disturbance state representing the presence of a vehicle. By tracing back the magnetic field strength numerical sequence, the last sampling point with an absolute difference value greater than or equal to the disturbance judgment threshold is found; the timestamp of this point is recorded as the start time of state recovery. The time interval between the start of the recovery from the state to the moment of entering the magnetic field reference state is the duration of the magnetic field strength recovering from the disturbance state representing the presence of the vehicle to the reference state.

[0071] The first threshold is set to distinguish between actual vehicle departure and transient magnetic field fluctuations caused by occupant activity. This threshold is derived from statistical analysis of a large number of historical vehicle departure samples. Specifically, the first threshold is obtained by collecting multiple sets of instances where vehicles have indeed left their parking spaces, calculating the duration for the magnetic field strength to recover from the disturbance state representing the vehicle's presence to a baseline state in each instance, and then taking the minimum value or a percentile close to the minimum (e.g., the fifth percentile) as the reference benchmark for the first threshold. Considering system fault tolerance, this reference benchmark can be multiplied by a safety factor, such as 1.2, to obtain the specific value of the first threshold. In this way, the first threshold is set to a time length that can reliably distinguish between actual departure and transient fluctuations, for example, 2 seconds. The calculated duration for the magnetic field strength to recover from the disturbance state representing the vehicle's presence to the baseline state is compared with the first threshold. If this duration exceeds the first threshold, a trigger signal is generated.

[0072] When a trigger signal is generated, the analysis process for video image data of the same parking space is initiated. The video image data is analyzed to identify whether the vehicle outline disappears from the corresponding parking space's defined area. The video image data frame associated with the current moment is accessed, and the parking space's defined area is located. Moving target detection is performed using background subtraction. Specifically, a background image model of the parking space's defined area in a vehicle-free state is pre-established and maintained. The pixel grayscale values ​​of the parking space's defined area in the current image frame are compared with the corresponding pixel grayscale values ​​in the background image model. For each pixel location, the absolute value of the difference between the current grayscale value and the background grayscale value is calculated. If this absolute value is greater than a preset image difference threshold, the pixel is marked as a foreground pixel; otherwise, it is marked as a background pixel. After performing the above operations on all pixels, a binarized foreground mask image is obtained. Morphological processing is performed on the foreground mask image, including erosion to eliminate small noise points and dilation to connect adjacent foreground regions. The total pixel area of ​​the morphologically processed foreground region is then calculated. Simultaneously, an edge detection algorithm is used to extract all significant edge contours within the designated parking space area. The edge detection algorithm calculates the image grayscale gradient and applies high and low thresholds to filter out strong edge pixels, connecting adjacent strong edge pixels to form contours. The spatial position of the foreground region obtained from motion detection is compared with the contours obtained from edge detection, identifying edge sets surrounded by the foreground region and possessing closed or semi-closed features. These edge sets are identified as potentially representing the overall contour of the vehicle. It is then determined whether this overall contour is moving out of the designated parking space area. By calculating the centroid position of this overall contour in multiple consecutive frames, if the centroid position sequence shows that it is continuously moving towards the boundary of the designated parking space area, it is determined that the vehicle contour is leaving. The timestamp corresponding to the image frame where the centroid of the vehicle contour first completely moves out of the boundary of the designated parking space area is recorded as the starting time of the vehicle contour's disappearance.

[0073] The second threshold duration is set to confirm that the vehicle has left the parking space and is not merely a brief stop at the parking space boundary or to detect jitter. The second threshold duration is obtained similarly to the first threshold, based on statistical analysis of typical video clips where the vehicle does not return after leaving. The second threshold duration, for example, is determined by analyzing the average time interval between the vehicle outline disappearing from the designated parking space area and no longer reappearing, taking into account a certain amount of redundancy. This means that it is necessary to confirm that the vehicle outline's disappearance within the designated parking space area lasts for at least the second threshold duration. In the video image data, starting from the aforementioned disappearance start time, the designated parking space area is continuously analyzed for subsequent consecutive image frames. If, within the second threshold duration (e.g., 1 second), in each subsequent frame, the same background subtraction and outline analysis fails to detect a complete outline of the same vehicle within the designated parking space area, then the vehicle outline disappearance is determined to have lasted for the second threshold duration. If, within the second threshold duration, a clear and continuous vehicle outline is detected again within the designated parking space area, then the disappearance is determined to have not lasted for the second threshold duration, and the departure judgment process terminates.

[0074] When two conditions are met simultaneously—that is, the duration for which the magnetic field strength recovers from the disturbance state indicating the presence of a vehicle to the baseline state exceeds a first threshold, and the vehicle outline is identified as disappearing from the designated area of ​​the corresponding parking space based on video image data for a duration exceeding a second threshold—a departure event is ultimately determined to have occurred for that parking space. This departure event is recorded as a data object containing the event type, parking space identification code, and the time of determination, for use in subsequent processing steps. The determination time can be the start time of the vehicle outline disappearing, the time when the magnetic field state recovers to the baseline state, or a weighted average of both, such as choosing the later of the two times as the determination time to ensure robustness of the judgment.

[0075] S3. When a departure event is determined to have occurred, analyze and distinguish the type of departure event, which includes temporary movement or final departure. The specific implementation is as follows:

[0076] Once a departure event is determined, its type is immediately analyzed and differentiated. First, based on the determined time of the departure event, video clips covering a specific time range before and after that moment are retrieved from the stored video image data. Specifically, using the determined time as the time origin, video image data of a first preset duration is extracted forward, and video image data of a second preset duration is extracted backward. The first preset duration is set to capture a complete record of the vehicle's parking behavior before departure; its value should be sufficient to include the complete process of the vehicle coming to a complete stop and any possible passengers getting in or out, for example, it could be set to 30 seconds. The second preset duration is set to completely capture the process of the vehicle moving from the start of its movement to leaving the designated parking area; its value must cover the longest time from start to departure, for example, it could be set to 15 seconds. This acquired continuous video image data will be used for subsequent departure speed analysis and parking feature analysis.

[0077] Based on the acquired video image data, the analysis examines whether the vehicle's departure speed exceeds a speed threshold. The specific analysis process is as follows: In the acquired video image data, the sequence of image frames within a second preset time period following the determination time is located. Within these consecutive image frames, the identified vehicle contour is continuously tracked. For each pair of adjacent image frames, the position change of the vehicle contour's centroid in the image pixel coordinate system is calculated. The centroid of the vehicle contour is obtained by calculating the average of the horizontal and vertical coordinates of all pixels constituting the contour. According to the camera's calibration parameters, the position change in the image pixel coordinate system is converted into displacement in the real-world ground coordinate system. This conversion is based on the camera's intrinsic parameter matrix and its extrinsic parameter matrix relative to the ground, achieved by solving the homography matrix, the mapping relationship between image coordinate points and ground coordinate points. The homography matrix is ​​obtained through the camera calibration process, which includes capturing images of a calibration board of known size and calculating transformation parameters. Simultaneously, the actual time interval between two adjacent frames is obtained from the timestamps of the image frames. For each pair of adjacent frames, the calculated ground position is subtracted by the time interval to obtain an instantaneous speed value. The analysis covers all calculated instantaneous speed values ​​from the moment the vehicle's outline begins to move noticeably (i.e., the center of gravity displacement exceeds a set threshold) until the vehicle's center of gravity completely leaves the designated parking space area. The departure speed can be the average of these instantaneous speeds, or the maximum value can be used as a representative value. The speed threshold is set based on the difference between common parking maneuvering behavior and normal departure behavior. Through statistical analysis of a large number of samples, the average speed of normally departing vehicles within the parking space area is usually below a specific value, such as 1.5 meters per second; while the speed of short-term maneuvering for reasons such as yielding is often faster. Therefore, the speed threshold can be set between 1.5 meters per second and 2.5 meters per second, for example, 2 meters per second. The analyzed departure speed is compared with this speed threshold to determine whether it exceeds the threshold.

[0078] Simultaneously, based on the acquired video image data, the system analyzes whether the vehicle's parking characteristics before the departure event meet the complete parking conditions. The parking characteristic satisfaction condition is determined through two sub-judgments. First, from the video image data within a first preset time period before the departure event, the time period from when the vehicle comes to a complete stop to when the door opens is extracted. The criterion for determining when the vehicle is stopped is that, in multiple consecutive frames (e.g., approximately 0.4 seconds for 10 consecutive frames), the change in the centroid of the vehicle's outline is less than a minimum threshold, such as 5 pixels. When the change in the centroid's position is consistently less than this minimum threshold, that moment is recorded as the moment the vehicle comes to a complete stop. The determination of when the door is open is based on image change analysis of a specific region of the vehicle's outline, such as the area near the side of the vehicle. Using background subtraction, the pixels of the vehicle outline region in the image after the vehicle has come to a complete stop are compared with the pixels of the same region in the image at the initial stage of stopping. When a continuous and significant change in pixel grayscale values ​​is detected in the normally open area of ​​the door, and this change region conforms to the morphological characteristics of an open door, it is determined that the door is open. The moment when this type of change is first detected is recorded as the moment the door opens. The interval between the moment the vehicle comes to a complete stop and the moment the door opens is the time period from when the vehicle comes to a complete stop until the door opens. If no door opening event is detected within the first preset time period, the end of this time period is considered the moment when the departure event is determined to have occurred.

[0079] Next, it is determined whether the time period continuously reaches the third preset duration. The third preset duration is set to distinguish between short-term parking (stop and go) and complete parking with a clear purpose. Complete parking usually involves actions such as people getting out of the vehicle and turning off the engine, requiring a certain minimum time. The third preset duration can be determined based on time statistics of typical parking behaviors. By collecting samples of people getting out of the vehicle after it has come to a stop, the shortest time distribution can be analyzed, for example, taking the fifth percentile to obtain a reference value such as 5 seconds. If the time period from when the vehicle comes to a stop to when the door opens is less than the third preset duration, the parking feature is considered not to meet the complete parking condition. In addition, it is also necessary to determine whether the vehicle outline remains stationary during this time period. The criterion for remaining stationary is that in each frame of the image during this time period, the change in the centroid position of the vehicle outline is less than the aforementioned minimum threshold, such as 5 pixels. If the vehicle outline moves in any position exceeding the threshold during this time period, it is also determined that the complete parking condition is not met. Only when the time period from when the vehicle comes to a stop to when the door opens continuously reaches the third preset duration, and the vehicle outline remains stationary during this time period, is the parking feature considered to meet the complete parking condition.

[0080] The departure event type is determined by combining the results of the two analyses above. The first analysis result is whether the departure speed exceeds a speed threshold. The second analysis result is whether the parking characteristics meet the complete parking conditions. When the departure speed exceeds the speed threshold and the parking characteristics do not meet the complete parking conditions, the departure event is classified as a temporary movement. Temporary movement represents a situation where the vehicle moves rapidly before completing a complete parking action, such as a short-distance position adjustment to make way for a passage. In all other cases, i.e., when the departure speed does not exceed the speed threshold, or the parking characteristics meet the complete parking conditions, or both, the departure event is classified as a terminated departure. A terminated departure represents the end of a complete parking action, with the vehicle leaving the parking space and not expected to return immediately. This classification result will be linked to the departure event to form a departure event record with a type label, which will be used to guide different monitoring strategies in subsequent steps.

[0081] S4. When a departure event is classified as temporary relocation, a first short-term analysis timer is initiated for the parking space; when it is classified as a final departure, a second short-term analysis timer is initiated. The duration of the first short-term analysis timer is shorter than that of the second short-term analysis timer. Specifically, the implementation is as follows:

[0082] Based on the classification of departure events, a differentiated timing strategy is implemented. For departure events classified as temporary relocation, a timer with a first time window is configured as the first short-term analysis time. For departure events classified as final departures, a timer with a second time window is configured as the second short-term analysis time. The timer is a logical functional entity that records the elapsed time and is implemented through software programming. This timer is uniquely associated with a specific parking space identifier code and this departure event.

[0083] The duration of the first and second time windows is set based on statistical analysis of vehicle behavior patterns after different types of departure events in historical data. The analysis process requires collecting correctly labeled departure event samples, including known temporary movement events and final departure events. For each temporary movement event sample, the time interval from when the vehicle leaves to when it returns to its original or adjacent parking space is recorded. These time intervals are statistically analyzed to calculate their numerical distribution, such as calculating the 90th percentile. The duration of the first time window can be set slightly larger than this 90th percentile value; for example, if the percentile is 60 seconds, the duration of the first time window can be set to 70 seconds. For final departure event samples, the general pattern of vehicles not returning for extended periods after leaving is analyzed. The duration of the second time window can be set based on parking lot operation experience, such as the typical maximum time required for a vehicle to complete payment and leave the parking lot, for example, 300 seconds. The principle is that the duration of the first time window configured for a temporary relocation event is shorter than the duration of the second time window configured for a final departure event.

[0084] The specific operation for configuring a first short-term analysis timer for a temporary relocation type departure event is to create a logical object. This logical object contains the following information: the associated parking space identifier code, the associated departure event identifier, a timer type marked as first short-term analysis timer, a time window duration value equal to the duration of the first time window determined above, a timer status marked as pending activation, and the timer start time information initialized to empty. The operation for configuring a second short-term analysis timer for a final departure type departure event is to create another logical object whose timer type is marked as second short-term analysis timer, and whose time window duration value equals the duration of the second time window determined above.

[0085] After classifying departure events and generating departure event records with type labels, the corresponding timer for each type is immediately activated. The activation process involves selecting the configured corresponding timer logical object based on the type label in the departure event record. Then, the current system time is written to the timer's start time information location within the timer logical object. Simultaneously, the timer's status is marked as running. This current system time is the timer's start time. After the timer starts counting, the system periodically checks its status. This check involves reading the current system time, calculating the difference between the current system time and the timer's start time recorded in the timer logical object, and obtaining the elapsed time. Then, the elapsed time is compared with the duration of the time window stored in the timer logical object.

[0086] The first and second short-term analysis timers operate independently. Each activated timer independently determines its timeout based on its own start time and the duration of a preset time window. While the timer is running and the elapsed time has not exceeded its time window duration, the system executes specific monitoring logic for the associated parking space. Once the elapsed time exceeds the time window duration, the timer is considered to have timed out, its status is marked as timed out, and the associated specific monitoring logic terminates. In this way, the first short-term analysis timer provides a shorter monitoring window for temporary relocation events, while the second short-term analysis timer provides a longer observation window for final departure events. The activation and operation of the timers are crucial links connecting the event type analysis step with subsequent event monitoring steps.

[0087] S5. Monitor whether an entry event occurs in the parking space within the corresponding duration; if an entry event is detected, analyze the state change information of at least one preset adjacent parking space corresponding to the parking space recorded at the time of the exit event, and whether it meets the preset correlation conditions with the state change information of at least one adjacent parking space before and after the entry event. Specifically, the implementation is as follows:

[0088] During the corresponding time window duration of the activated and running first or second short-term analysis timer, the associated parking space is continuously monitored to determine whether an entry event has occurred. The method for monitoring entry events corresponds logically to, but in the opposite direction to, the method for determining departure events. The monitoring process continuously checks the vehicle presence detection data of the associated parking space. Specifically, the magnetic field change data of the parking space is analyzed to determine whether the magnetic field strength has changed from the magnetic field baseline state to a disturbance state indicating the presence of a vehicle, and whether this disturbance state continues to exceed a preset dwelling judgment threshold. The dwelling judgment threshold is used to confirm that the vehicle has come to a stable stop rather than passing by briefly; its value is determined by statistically analyzing the shortest time required for a vehicle to come to a stable stop, for example, it can be set to 3 seconds. At the same time, the video image data of the parking space is analyzed to identify whether a vehicle silhouette has entered and remained stably within the designated area of ​​the parking space for more than a preset dwelling confirmation threshold. The dwelling confirmation threshold is used to visually confirm that the vehicle has come to a stable stop; its value can be set in coordination with the dwelling judgment threshold, for example, it can be set to 2 seconds. When both the magnetic field change and video image analysis confirm that a vehicle has entered and stopped stably, it is determined that an entry event has occurred in that parking space, and the time of the entry event is recorded.

[0089] Based on the geographical relationships of parking spaces, at least one adjacent parking space is pre-configured statically for each parking space. The configuration process is based on the digital layout map of the parking lot. First, the location of the target parking space in the layout map and its adjacent lanes are determined. Based on the preset one-way travel direction of the lane, the vehicle exit direction of the target parking space is determined. The next parking space immediately preceding the target parking space in the exit direction, i.e., the first space the vehicle passes after exiting the target parking space, is designated as the first adjacent parking space. Next, other parking spaces adjacent to the target parking space on the side and sharing the same lane exit are identified. Sharing the same lane exit means that vehicles exiting from these side parking spaces must use the same lane segment as the target parking space to leave the area. One side adjacent parking space that meets this condition is designated as the second adjacent parking space. The set of adjacent parking spaces configured for each target parking space, such as including the two parking spaces mentioned above, is determined during system initialization and stored as a fixed mapping table. The key is the target parking space identifier code, and the value is a list of identifier codes for its adjacent parking spaces.

[0090] Upon the occurrence of a vehicle departure event, status monitoring of at least one pre-defined adjacent parking space is immediately initiated, and status change information is recorded. A first state change sequence is recorded, starting from the time the departure event is determined, within a first pre-defined time period, where at least one adjacent parking space changes from occupied to vacant. The length of the first pre-defined time period should be sufficient to cover the time required for the chain of yielding actions that may be triggered by the departure of the target vehicle to complete. Its value is estimated based on the typical length of the parking lot aisle and the vehicle movement speed; for example, it can be set to 30 seconds. The monitoring process involves acquiring vehicle presence detection data for each adjacent parking space in real time, starting from the time the departure event is determined. For each adjacent parking space, it is determined whether its status changes from occupied to vacant. The criteria for determining occupied status are that the magnetic field strength of the adjacent parking space is in a disturbed state, indicating the presence of a vehicle, and video image analysis confirms that there is a stable vehicle outline within the designated area of ​​the parking space. The criteria for vacant status are that the magnetic field strength returns to the magnetic field baseline state, and video image analysis confirms that there is no stable vehicle outline within the designated area of ​​the parking space. When a nearby parking space is detected to change from occupied to vacant, a status change event is recorded. This event includes the nearby parking space's identifier, the change type (from occupied to vacant), and the precise time the change occurred. Within a first preset time period, all such status change events are recorded in chronological order, forming a first status change sequence. If a nearby parking space does not experience a status change within the first preset time period, or if its status change is not from occupied to vacant, then that nearby parking space has no corresponding record in the first status change sequence.

[0091] Upon detecting an entry event, a backtracking analysis of the status of at least one pre-defined adjacent parking space is initiated, recording status change information. A second state change sequence is recorded, ending at the time of the entry event, within a second pre-defined time period, showing at least one adjacent parking space changing from an vacant to an occupied state. The length of the second pre-defined time period is set similarly to the first pre-defined time period, for example, 30 seconds. The backtracking analysis process involves tracing back the historical data within the second pre-defined time period from the time of the entry event. For each adjacent parking space, its status change history within that time period is analyzed. It is determined whether its status has changed from vacant to occupied. Specifically, this is done by querying historical data to find the last time the adjacent parking space changed from vacant to occupied within that time period. This event is recorded, including the adjacent parking space identifier code, the change type (vacant to occupied), and the precise time of the change. Within the second pre-defined time period, all such state change events are recorded in chronological order, forming a second state change sequence. Similarly, if a adjacent parking space does not experience a vacant-to-occupied change within that time period, it will not have a corresponding record in the second state change sequence.

[0092] To determine if the association condition is met, the first and second state change sequences are determined to be opposite in chronological order and logically corresponding. This determination process consists of two sub-steps. First, check event pairing. Iterate through each state change event in the first state change sequence that changes from occupied to empty. For each such event, check in the second state change sequence whether there exists a corresponding state change event in the second state change sequence that changes from empty to occupied for the same adjacent parking space. The event pairing requirement is met only when every adjacent parking space state change event recorded in the first state change sequence has a corresponding reverse state change event in the second state change sequence for the same adjacent parking space. Second, check temporal reversal. For all paired state change events, extract their chronological order in the first state change sequence. Simultaneously, extract the chronological order of these corresponding events in the second state change sequence. Compare these two chronological orders. If the order of events in the second state change sequence is substantially opposite to the order of corresponding events in the first state change sequence, then the chronological order is reversed. One specific checking method is to assign a sequential number to each event in the first state change sequence, starting from 1 according to the order of occurrence. Then, based on the order of occurrence of the corresponding pairs of events in the second state change sequence, they are also assigned sequential numbers. The rank correlation coefficient between these two sets of numbers is calculated. If the calculated rank correlation coefficient is negative and its absolute value is greater than a preset correlation coefficient threshold, such as -0.7, then the two sequences are determined to have opposite occurrence times. Alternatively, a simpler rule can be used: if the adjacent parking space that experienced the earliest state change in the first sequence is the last to be occupied in the second sequence, and the adjacent parking space that experienced the latest state change in the first sequence is the first to be occupied in the second sequence, then the sequence is directly determined to be opposite. When the oppositeness of the sequence is satisfied, the two sequences are considered to correspond logically. Only when the results of both the event pairing and the oppositeness of the sequence are satisfied is it finally determined that the first state change sequence and the second state change sequence satisfy the preset association condition. The satisfaction of the association condition provides spatial logical evidence that the departure event and the arrival event belong to the same continuous behavior.

[0093] S6. When the association conditions are met and the vehicle corresponding to the entry event and the vehicle corresponding to the exit event meet the vehicle identity determination conditions, the entry event and the exit event are associated and the vehicle's in-place duration in the parking space is adjusted. Specifically, the implementation is as follows:

[0094] The satisfaction of the vehicle identity determination condition means confirming that the vehicle feature information recorded at the time of departure and the vehicle feature information recorded at the time of arrival correspond to the same physical vehicle. Vehicle feature information includes the vehicle's exterior color, brand and model outline features extracted from video image data, and the vehicle's license plate number characters. If the comparison process uses license plate number recognition, the identified license plate number characters from the departure event and the arrival event must be completely identical. If the comparison process uses vehicle exterior features, the similarity score between the two sets of feature vectors from the departure and arrival events is calculated. The feature vectors are obtained through feature extraction algorithms on the vehicle images. The similarity score is obtained by calculating the cosine similarity between the two feature vectors. If the similarity score exceeds a preset feature matching threshold, the vehicle identity determination condition is satisfied. The feature matching threshold is determined based on statistical analysis of numerous vehicle image matching experiments. In the experiments, the distribution of feature vector similarity scores between different vehicles and between images of the same vehicle at different times is calculated, and the threshold is set above the minimum similarity required to distinguish different vehicles; for example, it can be set to 0.95. Subsequent association and correction operations are only triggered when both the association conditions and the vehicle identity determination conditions are determined to be met.

[0095] Departure events that meet the specified conditions are marked as invalid departure events. Specifically, in the data structure storing departure event records, a dedicated status field indicating the validity of the event is identified; this field's default value is valid. The value of this status field is modified to a specific value representing invalidity, for example, changing the field value from 0 to 1, where 0 represents valid and 1 represents invalid. Simultaneously, a unique identifier for the associated entry event is entered into a field specifically used for recording logical associations. This marking operation means that in all subsequent statistical, billing, and parking space status deduction logic, this departure event will not be considered a valid vehicle departure.

[0096] Mark eligible entry events as invalid in-situ events. The implementation is similar to marking departure events: access the data structure storing the entry event record, locate its event validity status field, and modify the value of this field to a specific value indicating invalidity, such as changing it from 0 to 1. Simultaneously, fill in the corresponding departure event's unique identifier in its associated field. This marking operation means that in subsequent logic, this entry event will not be considered a new, independent vehicle arrival and parking commencement behavior.

[0097] Based on the actual occurrence times of the departure and arrival events, the actual continuous occupancy time of the parking space between the marked invalid departure and invalid presence events is calculated. The actual occurrence time of the departure event is the final determination time identified during the departure event judgment step; this time is a timestamp with specific precision, such as a Unix timestamp format accurate to milliseconds. The actual occurrence time of the arrival event is also a timestamp with the same format and precision, determined by the time it was judged to occur. The method for calculating the actual continuous occupancy time is to subtract the actual occurrence time of the departure event from the actual occurrence time of the arrival event; the difference is the actual continuous occupancy time measured in timestamp units. If the timestamp unit is milliseconds, the difference unit is milliseconds. This time difference needs to be converted to a more easily understood unit, such as seconds, by dividing the millisecond value by 1000. The calculated actual continuous occupancy time is a scalar value representing the actual length of time the parking space is continuously occupied by the vehicle despite a brief physical movement.

[0098] The calculated actual continuous occupancy time is added to the parking space's first-occupancy duration. The parking space's first-occupancy duration is a continuously updated state variable used to accumulate the total duration of continuous vehicle occupancy for that parking space. Its initial value begins accumulating from the moment the parking space was most recently determined to be a valid entry event (i.e., not an invalid occupancy event). Before performing the accumulation operation, the currently stored parking space's first-occupancy duration value needs to be obtained. The accumulation operation involves arithmetically adding this value to the actual continuous occupancy time value. The addition operation follows basic mathematical addition rules and ensures that the units are consistent, for example, both are in seconds. The sum is a new total value. Subsequently, this new total value is used to update the stored parking space's first-occupancy duration, overwriting the original value. This update operation directly corrects the accumulation process interrupted by misjudging a brief departure as a final departure, ensuring that the accumulated duration accurately reflects the continuous occupancy of the parking space.

[0099] After completing all the above operations, the processing of this brief departure and return sequence is finished. When the system subsequently performs billing or timeout checks, it will calculate based on the corrected, longer vehicle-in-position duration in the parking space, thus avoiding billing or management errors caused by incorrectly segmenting a single continuous parking sequence. This entire step is the final integration and output stage of the whole methodology. Based on the analysis results and judgments provided in the preceding steps, it executes explicit and reversible data modification operations, achieving intelligent understanding and accurate recording of vehicle-in-position status. All data operations are completed by reading and writing specific fields in storage.

[0100] Example 2: Figure 2A schematic diagram of a vehicle-in-situ intelligent detection system according to the present invention is provided. The vehicle-in-situ intelligent detection system includes the following modules:

[0101] The data acquisition module is used to acquire vehicle presence detection data that is sensed and uploaded in real time by detection nodes deployed in the parking area;

[0102] The event judgment module is used to determine whether a departure event targeting the same parking space has occurred based on vehicle presence detection data.

[0103] The type differentiation module is used to analyze and differentiate the type of departure event when it is determined that a departure event has occurred. The types include temporary movement or final departure.

[0104] The timing start module is used to start a first short-term analysis timer for the parking space when the departure event is classified as temporary movement; and to start a second short-term analysis timer when it is classified as final departure. The duration of the first short-term analysis timer is shorter than that of the second short-term analysis timer.

[0105] The correlation analysis module is used to monitor whether a parking space entry event occurs within the corresponding duration. If an entry event is detected, the module analyzes the state change information of at least one preset adjacent parking space corresponding to the parking space recorded when the exit event occurs, and whether the state change information of at least one adjacent parking space before and after the entry event meets the preset correlation conditions.

[0106] The association correction module is used to associate the entry event and the exit event and correct the vehicle's in-place duration in the parking space when the association conditions are met and the vehicle corresponding to the entry event and the vehicle corresponding to the exit event meet the vehicle identity determination conditions.

[0107] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0108] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0109] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0110] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0111] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0113] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligently detecting the presence of a vehicle, comprising: Includes the following steps: S1. Obtain vehicle presence detection data that is sensed and uploaded in real time by detection nodes deployed in the parking area; S2. Based on the vehicle presence detection data, determine whether a departure event targeting the same parking space has occurred; S3. When it is determined that a departure event has occurred, analyze and distinguish the type of departure event, which includes temporary movement or final departure. S4. When the departure event is classified as temporary relocation, a first short-term analysis timer is started for the parking space; when it is classified as final departure, a second short-term analysis timer is started, and the duration of the first short-term analysis timer is shorter than that of the second short-term analysis timer. S5. Monitor whether a parking space entry event occurs within the corresponding duration; If an entry event is detected, analyze the state change information of at least one preset adjacent parking space corresponding to the parking space recorded when the exit event occurs, and whether the state change information of at least one adjacent parking space before and after the entry event meets the preset correlation conditions. S6. When the association conditions are met and the vehicle corresponding to the entry event and the vehicle corresponding to the exit event meet the vehicle identity determination conditions, the entry event and the exit event are associated and the vehicle in the parking space is corrected for the duration of the vehicle in the parking space.

2. The method of claim 1, wherein, S1 includes: A geomagnetic sensor is independently deployed in each parking space to sense changes in the magnetic field of the corresponding parking space in real time and generate magnetic field change data. Deploy at least one camera in a public area covering multiple parking spaces to collect video image data containing the corresponding parking spaces; Vehicle in-situ detection data includes magnetic field change data and video image data.

3. The method of claim 1, wherein S2 include: Based on the magnetic field change data of the same parking space, determine whether the duration of the magnetic field strength recovering from the disturbance state representing the presence of the vehicle to the baseline state exceeds the first threshold. When the first threshold is exceeded, the analysis of the parking space video image data is triggered; Analyze video image data to identify whether the vehicle outline disappears from the corresponding parking space area and remains for a second threshold duration; When the vehicle outline disappears for a duration that exceeds the second threshold, it is determined that a vehicle has left the parking space.

4. The method of claim 1, wherein S3 include: Acquire video image data within a first preset time period before the departure event occurs and within a second preset time period after the event occurs; Based on video image data, analyze whether the vehicle's departure speed exceeds a speed threshold during the departure process; At the same time, it was analyzed whether the parking characteristics of the vehicle before the departure incident met the complete parking conditions; When the departure speed exceeds the speed threshold and the parking characteristics do not meet the complete parking conditions, the departure event is classified as a temporary movement. Otherwise, the departure event will be classified as a final departure.

5. The method of claim 4, wherein, Whether the parking features meet the complete parking conditions is determined by the following methods: extracting the time period from when the vehicle comes to a complete stop to when the door is opened from the video image data within the first preset time period before the departure event occurs; determining whether the time period lasts for a third preset time period; and determining whether the vehicle outline remains stationary during the time period.

6. The method of claim 1, wherein S4 include: Configure a timer with a first time window as the first short-time analysis timer for temporary relocation type departure events; Configure a timer with a second time window as a second short-time analysis timer for departure events of the final departure type; The duration of the first time window is shorter than the duration of the second time window; After distinguishing the type of departure event, the corresponding timer is immediately activated to start counting.

7. The method of claim 1, wherein S5 include: Based on the geographical location of the parking spaces, at least one adjacent parking space is pre-assigned; When a departure event occurs, record the first state change sequence of at least one adjacent parking space changing from occupied to vacant within a first preset time period, starting from the corresponding time. When an entry event is detected, a second state change sequence is recorded, which is the sequence of at least one adjacent parking space changing from an idle state to an occupied state within a second preset time period, ending at the time of the entry event. Determine whether the sequence of changes in the first state and the sequence of changes in the second state are opposite in chronological order and logically correspond to each other, in order to determine whether the association condition is met.

8. The method of claim 7, wherein, Based on the geographical relationship of parking spaces, at least one adjacent parking space is pre-defined in the following way: based on the driving direction of the parking lane, an adjacent parking space located in front of the target parking space in the direction of exit, and another adjacent parking space located to the side of the target parking space and sharing the same lane exit are jointly defined as at least one adjacent parking space.

9. The method of claim 1, wherein S6 include: When both the departure event and the entry event meet the association conditions and the vehicle identity determination conditions, the departure event is marked as an invalid departure event; Mark the entry event as an invalid in-place event; Based on the actual occurrence time of the departure event and the actual occurrence time of the entry event, calculate the actual continuous occupancy time of the parking space between the marked invalid departure event and invalid in-space event; The actual continuous occupancy time is added to the duration of the vehicle that was in the parking space first.

10. A vehicle-in-place intelligent detection system for implementing the vehicle-in-place intelligent detection method of any one of claims 1-9, characterized in that, Includes the following modules: The data acquisition module is used to acquire vehicle presence detection data that is sensed and uploaded in real time by detection nodes deployed in the parking area; The event judgment module is used to determine whether a departure event targeting the same parking space has occurred based on vehicle presence detection data. The type differentiation module is used to analyze and differentiate the type of departure event when it is determined that a departure event has occurred. The types include temporary movement or final departure. The timing start module is used to start a first short-term analysis timer for the parking space when the departure event is classified as temporary movement; and to start a second short-term analysis timer when it is classified as final departure. The duration of the first short-term analysis timer is shorter than that of the second short-term analysis timer. The correlation analysis module is used to monitor whether a parking space entry event occurs within the corresponding duration. If an entry event is detected, analyze the state change information of at least one preset adjacent parking space corresponding to the parking space recorded when the exit event occurs, and whether the state change information of at least one adjacent parking space before and after the entry event meets the preset correlation conditions. The association correction module is used to associate the entry event and the exit event and correct the vehicle's in-place duration in the parking space when the association conditions are met and the vehicle corresponding to the entry event and the vehicle corresponding to the exit event meet the vehicle identity determination conditions.

Images