CN-121980504-A - Parking lot management method, device and system for checking authenticity of vehicle entrance and exit

CN121980504ACN 121980504 ACN121980504 ACN 121980504ACN-121980504-A

Abstract

The invention discloses a parking lot management method, a device and a system for checking the authenticity of vehicle entrance and exit, and particularly relates to the technical field of intelligent parking lot management, wherein when a vehicle motion track is rebuilt, surrounding interference vectors and behavior discontinuous marks are coupled to form an extended track data set; the method comprises the steps of obtaining a pure vector by analyzing and stripping a neighboring vehicle interference signal through a principal component, integrating a Markov chain with gradient descent optimization, dynamically generating and calibrating an intended branch path, carrying out weighted fusion correction on nonstandard behaviors, introducing Monte Carlo avoidance simulation prediction interference amplification effect to isolate incomplete behaviors, adaptively adjusting an isolation condition threshold value and a dynamic cooperation threshold value by using the predicted interference amplification effect probability, generating an access record through a dual threshold value judging mechanism, and carrying out iterative optimization on system parameters through association rule mining, thereby solving the misjudgment and delay linkage problems caused by multi-vehicle interference and nonstandard behaviors in a high-density parking lot.

Inventors

ZHANG JUN
LIANG YEHUI
HUANG JIANLI
HU PENGFEI
SHI JIAJUN
Gu bokai
Qian Zelei
GAO FENG

Assignees

浙江嘉广信息科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260121

Claims (10)

1. The parking lot management method for checking the authenticity of the vehicle entering and exiting is characterized by comprising the following steps: S1, capturing peripheral interference vectors and behavior discontinuous marks through a space grid dividing and time sequence analyzing mechanism when reconstructing a vehicle motion track to form an extended track data set, wherein the interference vectors comprise overlapping contour point clouds, and the behavior discontinuous marks comprise position shaking or direction reversal; S2, separating adjacent vehicle wave and signals from target vehicle signals through a vector interference stripping algorithm based on an extended track data set, and taking an interference vector as a behavior extension component part to perform principal component analysis so as to extract a pure vector, integrating behavior probability calculation, dynamically simulating discontinuous transition of vehicle behavior by using a state transition probability matrix based on the pure vector and a behavior discontinuous mark to generate a plurality of intention branch paths with corresponding probability weights, and performing weight optimization and behavior mark feedback correction based on distance measurement on the intention branch paths; S3, correcting the nonstandard behavior according to the generated intention branch path, introducing avoidance simulation operation when judging the in-out intention, predicting the interference amplification effect to isolate the unfinished behavior, and outputting a corrected intention vector; And S4, performing record generation operation by using the corrected intention vectors, adjusting isolation conditions of the intention vectors according to the prediction effect, screening effective intention vectors from the intention vectors, calculating the duty ratio of the number of the effective intention vectors in all corrected intention vectors, and comparing the duty ratio with a dynamic cooperation threshold value to prevent delay linkage caused by collective misjudgment.
2. The parking lot management method for checking the authenticity of the entrance and exit of the vehicle according to claim 1, wherein the parking lot management method is characterized in that time sequence analysis is carried out on position data, position coordinate difference values between continuous frames are calculated, jitter or inversion phenomena are judged by combining the position coordinate difference values with direction vectors, behavior discontinuous marks are formed, interference vectors are connected with the behavior discontinuous marks through grid time sequence coupling calculation, position jitter or direction inversion amplitude of discontinuous points is marked by adopting a vector norm formula, and low-efficiency identification caused by fragmentation independent processing is avoided.
3. The parking lot management method for checking the authenticity of the entrance and exit of the vehicle according to claim 1, wherein the interference subspace is subjected to dimension reduction operation by utilizing principal component analysis, eigenvalues are obtained and ordered by calculating covariance matrices of interference vectors in a data set, eigenvectors with the accumulated contribution rate of the eigenvalues reaching a preset proportion are selected as principal components, original interference vectors are projected onto the principal components to extract pure vectors, elements of the covariance matrices are calculation results of variances and covariances among dimensions of the interference vectors, the eigenvalues are obtained through matrix diagonalization, and the projection operation is realized by adopting matrix multiplication.
4. The parking lot management method for checking the authenticity of the entrance and exit of the vehicle according to claim 1, wherein the process of generating the intended branch paths and carrying out weight optimization and feedback correction on the intended branch paths comprises the steps of simulating behavior discontinuous transition by using a Markov chain based on the extracted pure vector and the behavior discontinuous marks, dynamically generating a plurality of intended branch paths, simultaneously iteratively optimizing the engagement weights of the branch paths by comparing the Euclidean distance between the pure vector and each branch path by using a gradient descent method to minimize the distance error, and correcting the confidence of the corresponding state transition in real time by taking the amplitude of the behavior discontinuous marks as a feedback factor, thereby forming a complementary purification chain integrating signal stripping, path generation and feedback correction.
5. The parking lot management method for checking the authenticity of the entrance and exit of the vehicle according to claim 1 is characterized in that correcting nonstandard behaviors and predicting interference amplification effects comprises the steps of carrying out weighted fusion on corresponding vectors of a plurality of intention branch paths by utilizing branch probability to form a single calibration intention vector, adopting Monte Carlo sampling, generating multiple simulated tracks by repeatedly and randomly adding adjacent vehicle interference variables subjected to preset distribution to the calibration intention vector, determining a dynamic boundary threshold value for judging whether the simulated tracks cross the boundary or not in each sampling by multiplying the sum of a basic boundary threshold value and a preamble interference amplification effect probability by a dynamic offset coefficient, counting the proportion that the deviation amplitude exceeds the dynamic boundary threshold value in all the sampling as the current interference amplification effect probability, and marking the corresponding intention vector as the unfinished behaviors and rejecting when the interference amplification effect probability is higher than a preset boundary value.
6. The parking lot management method for vehicle entrance/exit authenticity verification according to claim 1, wherein the intention vectors are calibrated by using a branch probability weighted fusion operation, the corresponding intention vectors in the paths are weighted and summed by calculating the probability value of each branch path as a weight, the weighted and summed formula is that the corrected intention vector is equal to the sum of the products of the respective branch intention vectors and the corresponding probabilities, thereby fusing a plurality of branches to form a single calibrated intention vector, and the intention vector weights of the branches with low probability are multiplied by the inverse of the interference amplification effect probability to perform penalty adjustment, the inverse is calculated as 1 divided by the interference amplification effect probability, and the penalty amplitude is increased when the interference amplification effect probability approaches 1, thereby realizing the confusion correction for nonstandard behaviors.
7. The parking lot management method for checking the authenticity of the entrance and exit of the vehicle according to claim 1 is characterized in that the isolation condition for adjusting the intention vector according to the predicted effect is that the probability of the interference amplification effect calculated in the step S3 is calculated, the probability of the interference amplification effect is differentiated from a preset interference effect reference value, nonlinear mapping is conducted through a hyperbolic tangent saturation function to obtain an adjustment coefficient, the adjustment coefficient is multiplied by a preset isolation threshold amplification coefficient and then added with 1 and then multiplied by an original isolation threshold to obtain an adjusted isolation condition threshold, and geometric distance screening is conducted on the corrected intention vector by utilizing the adjusted isolation condition threshold to obtain an effective intention vector.
8. The parking lot management method for vehicle entrance/exit authenticity verification according to claim 1, wherein the game theory balancing optimization in step S4 includes the steps of: establishing a double game model between a threshold adjustment strategy and an interference environment state, wherein the pure strategy of the threshold adjustment strategy comprises a relaxation isolation threshold and a tightening isolation threshold, and the pure strategy of the interference environment state comprises a high-interference scene and a low-interference scene; Determining a profit value corresponding to each strategy combination in the profit matrix through historical data statistics or online learning, wherein the profit value is calculated by the identification accuracy, the false alarm rate and the passing delay time according to preset weight combinations; And mapping the equalization solution into a theoretical upper limit value of threshold adjustment of an isolation condition, and taking the theoretical upper limit value as an actually adopted adjustment result if a threshold adjustment suggested value calculated according to the real-time interference amplification effect probability exceeds the theoretical upper limit value during operation.
9. Parking lot management device for vehicle entrance/exit authenticity verification, comprising a method according to any of the preceding claims 1-8, characterized by comprising a processor, a memory, a communication interface, and one or more programs stored in said memory and configured to be executed by said processor, said programs comprising steps for performing the method according to any of the claims 1-8.
10. Parking lot management system for verification of the authenticity of vehicle entrance/exit, for implementing the method according to any one of the preceding claims 1 to 8, comprising: The track expansion module is used for capturing peripheral interference vectors and behavior discontinuous marks through a space grid dividing and time sequence analyzing mechanism when the motion track of the vehicle is rebuilt, so as to form an expanded track data set, wherein the interference vectors comprise overlapping contour point clouds, and the behavior discontinuous marks comprise position shaking or direction reversal; The intention purifying module is used for separating adjacent car wave and signals by utilizing a vector interference stripping algorithm based on the extended track data set, and dynamically generating an intention branch path by integrating behavior probability calculation; the risk prediction module corrects the nonstandard behavior according to the generated intention branch path, introduces avoidance simulation operation when judging the intention of the user, predicts the interference amplification effect to isolate incomplete behavior, and outputs corrected intention vectors; the decision generation module is used for executing record generation operation by utilizing the corrected intention vectors, adjusting the isolation conditions of the intention vectors according to the prediction effect, screening effective intention vectors from the intention vectors, calculating the duty ratio of the number of the effective intention vectors in all corrected intention vectors, and comparing the duty ratio with a dynamic cooperation threshold value to prevent delay linkage caused by collective misjudgment.

Description

Parking lot management method, device and system for checking authenticity of vehicle entrance and exit Technical Field The invention relates to the technical field of intelligent parking lot management, in particular to a parking lot management method, device and system for checking the authenticity of vehicles entering and exiting. Background The parking lot is used as an urban transportation junction, the authenticity verification management of the vehicle entrance and exit is directly related to the optimization of space utilization and user experience, and in the actual scene of compact layout, the motion track is rebuilt through multi-sensor fusion and the intention is judged, so that the record is generated only when the vehicle completely crosses the boundary, and the charging omission caused by false entrance and exit is prevented. As vehicle types diversify and traffic increases, uncertainty in driver behavior and surrounding disturbances have become a common challenge, for example temporary reversing or hesitation pauses often occur due to avoidance, and signal crossings of adjacent vehicles can confound identification, making it difficult for the system to accurately isolate individual behaviors. The vehicle access checking system judges the intention of a driver to nonstandard behaviors such as hesitation and reversing can be further fragile due to multi-vehicle signal intersection, so that the recognition specificity is damaged, because the traditional track reconstruction only depends on a speed direction vector, but the behavior is easy to be affected by adjacent vehicles when discontinuous in reality, and the electric vehicle weak signals are mixed with the traditional vehicles to further exacerbate confusion. Specifically, the compounding problem occurs by overlapping the interference intention vectors through the contour point cloud, physical changes such as position jitter and adjacent vehicle directions cannot be accurately captured, collective misjudgment is finally caused, incomplete behaviors and interference are mixed into a whole, recording dislocation, capacity distortion and delay linkage are caused, and particularly in a crowding period, the effect can enlarge management confusion and safety risks, reduce adaptation of the system to behavior variability, and if the effect is not solved, prevent expansion to high-density intelligent application, increase intervention requirements and influence overall benefits. Disclosure of Invention In order to overcome the defects in the prior art, the invention provides a parking lot management method, device and system for checking the authenticity of the entrance and exit of a vehicle, which are used for coupling and purifying by taking an interference vector as a behavior extension and driving the self-adaptive adjustment of a double threshold value and the balanced constraint of a game theory by utilizing the probability of an interference amplification effect so as to realize the specificity and the robustness recognition of the entrance and exit intention of the vehicle under a composite interference environment, so as to solve the problems in the background technology and solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the parking lot management method for checking the authenticity of the entrance and exit of the vehicle comprises the following steps: Step S1, in the process of reconstructing a vehicle motion track, carrying out joint processing on vehicle periphery sensing data through a space grid dividing and time sequence analyzing mechanism so as to capture an interference vector caused by an adjacent vehicle or environment and a behavior discontinuous sign reflecting vehicle behavior abnormality, and forming an extended track data set containing target vehicle track information, a periphery interference vector and the behavior discontinuous sign, wherein the interference vector at least comprises contour point cloud overlapping information, and the behavior discontinuous sign at least comprises position shaking or direction reversal information; S2, based on the extended track data set, vector interference stripping processing is carried out by taking the interference vector as a behavior extension component part, and dimension reduction analysis is carried out on an interference subspace to separate neighboring vehicle wave signals and extract pure vectors which only reflect the motion characteristics of the target vehicle; After the pure vector is obtained, combining the behavior discontinuous mark, carrying out probabilistic simulation on discontinuous behavior transition possibly occurring in the vehicle based on a behavior state transition probability model, dynamically generating a plurality of intention branch paths with different probability weights, carrying out weight optimization based on distance measurement on the intention b