CN-122020075-A - Two-wheel vehicle illegal charging detection method and system based on pattern recognition

CN122020075ACN 122020075 ACN122020075 ACN 122020075ACN-122020075-A

Abstract

The application discloses a method and a system for detecting illegal charging of a two-wheeled vehicle based on pattern recognition, wherein the method comprises the steps of determining a charging pattern of the two-wheeled vehicle to be detected according to positioning information and basic charging data in charging data of a battery end, extracting a plurality of output frequency domain energy characteristics of a charger corresponding to the two-wheeled vehicle to be detected based on the charging data of the charging end, generating charging characteristic vectors of the charger based on the plurality of output frequency domain energy characteristics, calculating matching degrees of the charging characteristic vectors and candidate charger model characteristic vectors in a preset charger model characteristic library, acquiring historical charging time sequence distribution characteristics of the two-wheeled vehicle to be detected, determining current charging time sequence distribution characteristics of the two-wheeled vehicle to be detected according to battery state data in the charging data of the battery end, and determining an illegal charging detection result of the two-wheeled vehicle to be detected based on the matching degrees, the historical charging time sequence distribution characteristics and the current charging time sequence distribution characteristics, so that comprehensiveness of a detection mode and accuracy of a detection result can be improved.

Inventors

DONG GUANGMING
HU REN
WANG XINLIAN

Assignees

深圳聚瑞云控科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. The method for detecting the illegal charging of the two-wheeled vehicle based on the pattern recognition is characterized by comprising the following steps of: acquiring battery end charging data, charging end charging data and positioning information of a two-wheel vehicle to be detected in real time, and determining a charging mode of the two-wheel vehicle to be detected according to the positioning information and basic charging data in the battery end charging data; Extracting a plurality of output frequency domain energy characteristics of a charger corresponding to the two-wheel vehicle to be detected based on charging data of the charging end under the condition that the charging mode is a charged mode; Generating a charging feature vector of the charger based on the plurality of output frequency domain energy features, and calculating the matching degree of the charging feature vector and each candidate charger model feature vector in a preset charger model feature library; Acquiring historical charging time sequence distribution characteristics of the two-wheel vehicle to be detected, determining current charging time sequence distribution characteristics of the two-wheel vehicle to be detected according to battery state data in battery end charging data, and determining illegal charging detection results of the two-wheel vehicle to be detected based on the matching degree, the historical charging time sequence distribution characteristics and the current charging time sequence distribution characteristics.
2. The method for detecting the illegal charging of the two-wheeled vehicle based on the pattern recognition according to claim 1, wherein the extracting the plurality of output frequency domain energy features of the charger corresponding to the two-wheeled vehicle to be detected based on the charging end charging data comprises: generating an output high-frequency current waveform of a charger corresponding to the two-wheel vehicle to be detected based on the charging data of the charging end, and preprocessing the output high-frequency current waveform; Performing wavelet packet transformation on the preprocessing result of the output high-frequency current waveform according to the preset frequency band number to obtain a plurality of output frequency band signals, and calculating the output signal energy of each output frequency band signal based on the signal amplitude value of each output frequency band signal; And normalizing the output signal energy to obtain a plurality of output frequency domain energy characteristics of the charger.
3. The pattern recognition-based two-wheeled vehicle illicit charge detection method of claim 2, wherein the generating a charge feature vector of the charger based on the plurality of output frequency domain energy features comprises: determining the energy duty ratio of the output signals corresponding to the output frequency band signals according to the normalization processing result of the output signal energy; Filtering noise signals in the plurality of output frequency band signals based on a preset noise signal energy duty ratio threshold and the signal energy duty ratio to obtain a plurality of effective frequency band signals, and calculating an effective signal energy duty ratio corresponding to each effective frequency band signal; And determining a second frequency band splitting sequence corresponding to the effective frequency band signals according to the first frequency band splitting sequence corresponding to the output frequency band signals, and integrating the effective signal energy duty ratio according to the second frequency band splitting sequence to obtain the charging feature vector of the charger.
4. The pattern recognition-based two-wheeled vehicle illegal charging detection method according to claim 1, wherein the charging terminal charging data includes a charging terminal current and a charging terminal voltage; the calculating the matching degree of the charging feature vector and each candidate charger model feature vector in the preset charger model feature library comprises the following steps: calculating a charging end load according to the charging end current and the charging end voltage; Determining a target load interval to which the charging end load belongs and the deviation degree of the target load interval according to load thresholds corresponding to preset load intervals; And calculating the similarity of the charging feature vector and each candidate charger model feature vector in a preset charger model feature library, and correcting the similarity based on the deviation degree to obtain the matching degree of the charging feature vector and each candidate charger model feature vector in the preset charger model feature library.
5. The method for detecting the illegal charging of the two-wheeled vehicle based on the pattern recognition according to claim 4, wherein the determining the target load interval to which the charging end load belongs and the deviation from the target load interval according to the load threshold corresponding to each preset load interval comprises: Comparing the magnitude relation between the charging end load and the load threshold corresponding to each preset load interval, and determining a target load interval to which the charging end load belongs according to the magnitude relation; and determining the central load and the interval length of the target load interval, calculating the absolute value of the difference between the charging end load and the central load, and calculating the ratio of the absolute value of the difference to the interval length and the reciprocal of the ratio to obtain the deviation degree of the charging end load and the target load interval.
6. The pattern recognition-based two-wheeled vehicle illegal charge detection method according to claim 1, wherein the battery state data includes voltage time series data, current time series data, real-time temperature time series data, and remaining power time series data of the battery in a current time window; the determining the current charging time sequence distribution characteristic of the two-wheel vehicle to be detected according to the battery state data in the battery end charging data comprises the following steps: determining a voltage maximum value and a voltage minimum value in the voltage time sequence data, and calculating a voltage average value and a voltage time sequence change slope based on the voltage time sequence data to obtain the current voltage time sequence distribution characteristic; determining a current maximum value and a current minimum value in the current time sequence data, and calculating a current average value and a current fluctuation standard deviation based on the current time sequence data to obtain a current time sequence distribution characteristic; determining a temperature maximum value in the real-time temperature time sequence data, and calculating a temperature rising rate and a temperature fluctuation range based on the real-time temperature time sequence data to obtain a current temperature time sequence distribution characteristic; calculating the rising amplitude and the rising rate of the residual electric quantity based on the residual electric quantity time sequence data, and calculating the correlation coefficient of the residual electric quantity and the current based on the residual electric quantity time sequence data and the current time sequence data to obtain the current electric quantity time sequence distribution characteristic; And integrating the current voltage time sequence distribution characteristic, the current time sequence distribution characteristic, the current temperature time sequence distribution characteristic and the current electric quantity time sequence distribution characteristic to obtain the current charging time sequence distribution characteristic of the two-wheel vehicle to be detected.
7. The pattern recognition-based two-wheeled vehicle illicit charge detection method of claim 1, wherein the determining the illicit charge detection result of the two-wheeled vehicle to be detected based on the matching degree, the historical charge timing distribution characteristic and the current charge timing distribution characteristic comprises: Determining a target charger model and a target matching degree of a charger corresponding to the two-wheel vehicle to be detected in the preset charger model feature library according to the matching degree; calculating a target vector distance between the charging feature vector and a target charger model feature vector of the target charger model; And determining that the two-wheel vehicle to be detected is in illegal charging under the condition that the target matching degree is smaller than a matching degree threshold value corresponding to the target charger model and/or the target vector distance is larger than a preset vector distance.
8. A two-wheeled vehicle violation charge detection system based on pattern recognition, the system comprising: the charging mode determining module is used for acquiring battery end charging data, charging end charging data and positioning information of the two-wheel vehicle to be detected in real time, and determining the charging mode of the two-wheel vehicle to be detected according to the positioning information and basic charging data in the battery end charging data; The energy characteristic determining module is used for extracting a plurality of output frequency domain energy characteristics of the charger corresponding to the two-wheel vehicle to be detected based on the charging end charging data under the condition that the charging mode is a charged mode; The fingerprint matching degree calculation module is used for generating a charging feature vector of the charger based on the plurality of output frequency domain energy features and calculating the matching degree of the charging feature vector and each candidate charger model feature vector in a preset charger model feature library; The illegal charging detection module is used for acquiring historical charging time sequence distribution characteristics of the two-wheel vehicle to be detected, determining current charging time sequence distribution characteristics of the two-wheel vehicle to be detected according to battery state data in the battery end charging data, and determining illegal charging detection results of the two-wheel vehicle to be detected based on the matching degree, the historical charging time sequence distribution characteristics and the current charging time sequence distribution characteristics.
9. An electronic device comprising a processor, a memory and a program or instruction stored on the memory and running on the processor, which when executed by the processor, implements the steps of the pattern recognition based two-wheeled vehicle violation charge detection method of any of claims 1-7.
10. A readable storage medium, wherein a program or instructions is stored on the readable storage medium, which when executed by a processor, implements the steps of the pattern recognition based two-wheeled vehicle violation charge detection method of any of claims 1-7.

Description

Two-wheel vehicle illegal charging detection method and system based on pattern recognition Technical Field The application belongs to the technical field of data processing, and particularly relates to a two-wheel vehicle illegal charging detection method and system based on pattern recognition. Background With the rapid popularization of electric bicycles and electric motorcycles, electric two-wheeled vehicles become core vehicles for short-distance travel, but vehicle charging safety problems caused by illegal charging behaviors such as household charging, unmatched chargers, disordered wire connection and the like are increasingly prominent. Therefore, performing charging mode identification on the two-wheeled vehicle and detecting vehicle illegal charging behaviors based on the identification result has become a hot topic for improving urban governance and public safety. In the prior art, a camera is arranged at the entrance of a corridor and a community unit in advance in a detection mode of illegal charging of the two-wheeled vehicle, whether a vehicle mode is charged or not is determined through image recognition, and whether illegal charging behaviors such as vehicle home entry or flying line charging exist or not is further detected under the charged condition. However, the prior art relies on a large number of cameras to can't carry out accurate discernment to vehicle charging mode and violating charge behavior under the condition that the camera shelters from or the vehicle shelters from, there is the problem that detection result is not accurate enough to the detection of charging violating rules of two-wheeled vehicle. Disclosure of Invention The embodiment of the application aims to provide a two-wheel vehicle illegal charging detection method and system based on pattern recognition, which solve the problem that the detection result of the two-wheel vehicle illegal charging detection is not accurate enough in the prior art, identify the charging pattern of the two-wheel vehicle to be detected through the positioning information and basic charging data of the vehicle, extract a plurality of output frequency domain energy characteristics of a charger corresponding to the two-wheel vehicle to be detected according to charging end charging data, generate charging characteristic vectors of the charger, calculate the matching degree of the charging characteristic vectors and candidate charger model characteristic vectors in a preset charger model characteristic library, and determine the illegal charging detection result of the two-wheel vehicle to be detected based on the matching degree, historical charging time sequence distribution characteristics and current charging time sequence distribution characteristics, so that the aims of accurately identifying the charging pattern of the two-wheel vehicle and carrying out the two-wheel vehicle illegal charging detection by combining charger charging fingerprints and charging time sequence data can be achieved, and the comprehensiveness of the detection mode and the accuracy of the detection result are improved. In a first aspect, an embodiment of the present application provides a method for detecting a two-wheel vehicle illegal charging based on pattern recognition, where the method includes: acquiring battery end charging data, charging end charging data and positioning information of the two-wheel vehicle to be detected in real time, and determining a charging mode of the two-wheel vehicle to be detected according to the positioning information and basic charging data in the battery end charging data; Under the condition that the charging mode is a charged mode, extracting a plurality of output frequency domain energy characteristics of a charger corresponding to the two-wheel vehicle to be detected based on charging data of a charging end; Generating a charging feature vector of the charger based on the plurality of output frequency domain energy features, and calculating the matching degree of the charging feature vector and each candidate charger model feature vector in a preset charger model feature library; The method comprises the steps of obtaining historical charging time sequence distribution characteristics of a two-wheel vehicle to be detected, determining current charging time sequence distribution characteristics of the two-wheel vehicle to be detected according to battery state data in battery end charging data, and determining illegal charging detection results of the two-wheel vehicle to be detected based on matching degree, the historical charging time sequence distribution characteristics and the current charging time sequence distribution characteristics. Further, a plurality of output frequency domain energy characteristics of the charger corresponding to the two-wheel vehicle to be detected are extracted based on charging data of the charging end, and the method comprises the following steps: generating an output high-frequency