CN-121999551-A - Multi-feature fusion-based intelligent face recognition charging pile antitheft system, method and equipment

CN121999551ACN 121999551 ACN121999551 ACN 121999551ACN-121999551-A

Abstract

The invention belongs to the technical field of intelligent security and Internet of things, and relates to a face recognition intelligent charging pile anti-theft system, method and equipment based on multi-feature fusion. The face recognition module adopts a multi-feature fusion algorithm to finish face detection, feature extraction and similarity calculation, and related data are sent to the intelligent lock control module for final verification. And after receiving the data, the intelligent lock control module executes a three-level verification process, generates an unlocking or locking instruction according to the result, controls the opening and closing of the charging pile lock, and controls the lock to be closed after the charging is finished or an abnormal signal is received. The alarm module sends an alarm signal when the charging pile locking box is detected to be stressed and the authorized unlocking process is not completed. The power management module is responsible for powering and enforcing a hierarchical power management policy. According to the invention, the face recognition is performed through the multi-feature fusion algorithm, so that the authorized user is ensured to use the charging pile only in the correct time, and the charging safety is greatly improved.

Inventors

FENG RUIXUAN
LI PEILIN
Kang Linger

Assignees

电子科技大学

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. The intelligent face recognition charging pile antitheft system based on multi-feature fusion is characterized by comprising a face recognition module, an intelligent lock control module, an alarm module, a power management module and a microcontroller; The face recognition module is used for collecting the face image to be recognized, detecting and positioning the face to be recognized, extracting the characteristics and calculating the similarity, and sending the face detection result and the face similarity data obtained by calculation to the microcontroller; the intelligent lock control module is used for responding to an unlocking or locking instruction of the microcontroller and driving and controlling the opening and closing of the charging pile lock; The alarm module is used for monitoring the physical state of the charging pile lock box and sending an abnormal state signal to the microcontroller when the unauthorized opening behavior is detected; The power management module is used for supplying power to the system, monitoring the power consumption and the activity state of the system in real time, generating monitoring data, executing a power management instruction sent by the microcontroller, and reporting the monitoring data to the microcontroller; The microcontroller is respectively connected with the face recognition module, the intelligent lock control module, the alarm module and the power management module and is used for: According to the received face detection result and the face similarity data, performing three-stage verification, namely judging whether an effective face area meeting the preset gesture, shielding and quality requirements is detected based on the face detection result, judging whether the face similarity data is greater than or equal to a preset similarity threshold value if the effective face area is detected, performing living body detection on the effective face area if the face similarity data is greater than or equal to the preset similarity threshold value, generating an unlocking instruction and sending the unlocking instruction to the intelligent lock control module if the living body detection is passed; receiving an abnormal state signal sent by the alarm module, sending an alarm instruction to the alarm module based on the abnormal state signal, and/or sending a locking instruction to the intelligent lock control module; And receiving monitoring data reported by the power management module, and sending a power management instruction to the power management module based on the monitoring data or/and the result of whether the face is continuously not detected.
2. The system according to claim 1, wherein the face recognition module comprises an image acquisition unit, a face detection unit, an image preprocessing unit, a feature extraction unit and a similarity calculation unit; The image acquisition unit is used for acquiring images containing human faces; The face detection unit is connected with the image acquisition unit and is used for detecting and positioning a face area image from the acquired image and outputting a face detection result; The image preprocessing unit is connected with the face detection unit and is used for sequentially carrying out gray level conversion, histogram equalization and Gaussian filtering noise reduction on the detected face region image; The feature extraction unit is connected with the image preprocessing unit, and adopts an LBP operator to extract the LBP features of the preprocessed face image in a uniform mode, divides the face image into M multiplied by N uniform grids, and counts LBP feature histograms in each grid to be used as auxiliary grid statistical features; The similarity calculation unit is connected with the feature extraction unit, a registered face template is stored in the feature extraction unit, the similarity between the LBP feature of the uniform mode and the corresponding feature in the auxiliary grid statistical feature and the registered face template is calculated by adopting three similarity calculation methods of Euclidean distance, manhattan distance and correlation coefficient respectively, the three calculated similarities are weighted and fused to obtain final similarity S, and a final similarity value is sent to the microcontroller, wherein the calculation formula of the final similarity S is as follows: S=α*S euclidean +β*S manhattan +γ*S correlation ; Wherein α, β, γ are preset weight coefficients, and α+β+γ= 1;S euclidean represents a similarity value obtained by using the euclidean distance calculation method, S manhattan represents a similarity value obtained by using the manhattan distance calculation method, and S correlation represents a similarity value obtained by using the correlation coefficient calculation method.
3. The system of claim 2, wherein determining whether a specific condition of the valid face region meeting the preset pose, occlusion, and quality requirements is detected comprises: The gesture condition is that the absolute value of the yaw angle and the pitch angle of the face relative to the front face of the camera is not more than 30 degrees, and the rolling angle is not limited; the shielding condition is that the shielding areas of eyes, nose and mouth of the face are smaller than 20% of the area of the part; the quality requirement is that the face area image needs to meet the requirements that the resolution is not lower than 640x480 pixels and the signal to noise ratio is more than 20 dB; The region determined as a valid face is used for subsequent similarity comparison and in-vivo detection.
4. A system according to claim 2 or 3, wherein the validation rule for in-vivo detection of valid face regions when the face similarity data is greater than or equal to a preset similarity threshold comprises at least one of: Judging whether the face finishes the appointed biological characteristic action or not; Skin texture verification rules, namely judging whether the skin texture characteristic value of the human face is higher than a preset texture characteristic threshold value or not through texture characteristic analysis; and 3, three-dimensional structure verification rules, namely judging whether the facial depth information of the human face accords with preset real human face three-dimensional model parameters or not through three-dimensional or infrared imaging.
5. The system of claim 1, wherein the alarm module comprises a signal acquisition unit, an alarm control unit, and a buzzer; the signal acquisition unit is a film pressure sensor and is used for acquiring pressure signals received by the lock box of the charging pile in real time; The alarm control unit is connected with the signal acquisition unit and the microcontroller and is used for judging unauthorized opening behavior and sending an abnormal state signal to the microcontroller when the received pressure signal exceeds a preset pressure threshold value and an authorized unlocking signal is not received from the microcontroller; The buzzer is connected with the alarm control unit and used for giving an alarm according to an alarm instruction of the microcontroller.
6. The system of claim 1, further comprising a multi-modal feedback module including an indicator light and a display screen; the indicator lamp is used for displaying system states through different colors and flashing modes, including identification, identification success, identification failure and alarm states; the display screen is used for displaying the face recognition interface, the recognition result and the operation guidance in real time.
7. The system of claim 6, wherein the power management module comprises a hierarchical power management unit and a power monitoring unit; The hierarchical power management unit is used for controlling the power consumption state of the whole system by adopting a hierarchical management strategy according to the instruction of the microcontroller, and specifically comprises the following steps: After the time of continuously not detecting the human face exceeds a second preset time T2, wherein T2 is more than T1, controlling the display screen to enter a sleep mode; the power supply monitoring unit is used for monitoring the current and the voltage of the system in real time and reporting the monitored current and the monitored voltage to the microcontroller as monitoring data.
8. The system of claim 2, wherein the weighting coefficients α, β, γ used for the weighted fusion in the process of performing weighted fusion on the three calculated similarities to obtain the final similarity S are determined by cross-validation training optimization on the pre-collected face dataset.
9. The intelligent face recognition charging pile antitheft method based on multi-feature fusion is characterized by being applied to the system as claimed in claim 6 and comprising the following steps: the face recognition and feature fusion comprises the steps of acquiring a face image to be recognized by using a face recognition module, processing the face image to extract LBP features of a uniform mode and auxiliary grid statistical features, respectively calculating Euclidean distance, manhattan distance and correlation coefficient similarity between the LBP features of the uniform mode and the auxiliary grid statistical features and registered face templates, and carrying out weighted fusion on the LBP features and the auxiliary grid statistical features to obtain final similarity S; The third-level verification comprises the steps of executing third-level verification and decision by using a microcontroller, specifically, verifying whether an effective face area meeting the requirements of preset gesture, shielding and image parameters is detected by a priori, verifying whether the face similarity data exceeds a preset similarity threshold value, and finally performing living body detection verification; The lock control and alarm monitoring comprises the steps of responding to an unlocking instruction through an intelligent lock control module, unlocking the charging pile lock, monitoring the physical state of a charging pile lock box through an alarm module, and triggering an alarm if abnormal stress is detected when three-level verification is not completed; And the dynamic power management is that a power management module is used for monitoring the power consumption and the activity state of the system and generating monitoring data for reporting, and a microcontroller is used for controlling the system to enter different power consumption states including turning off the indicator light and the dormant display screen according to the result of whether the face is continuously not detected and the monitoring data.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the intelligent charging pile theft protection method based on multi-feature fusion of claim 9 when executing the computer program.

Description

Multi-feature fusion-based intelligent face recognition charging pile antitheft system, method and equipment Technical Field The invention relates to the technical field of intelligent security and Internet of things, in particular to a face recognition intelligent charging pile anti-theft system, method and equipment based on multi-feature fusion. Background With the rapid development of the electric automobile industry, the charging pile is used as a core supporting infrastructure, the market demand of the charging pile is in explosive growth, and the charging pile is widely applied to various scenes such as residential areas, commercial areas, public parking lots and the like. However, in the actual operation process, the safety management problem of the charging pile is increasingly remarkable, and the charging pile is particularly characterized in that in the charging process, the plug is maliciously pulled out to cause the conditions of charge interruption, the charging gun and accessory equipment are stolen, and the like, so that not only are the user property loss and the charging experience reduced, but also equipment faults are possibly caused by abnormal power failure, and potential safety hazards exist. At present, aiming at the anti-theft of the charging pile, solutions in the industry are mainly divided into two types. The mechanical lock is simple in structure and easy to be violently damaged or technically opened, meanwhile, the physical lock needs manual inspection and management, the state of the lock cannot be monitored in real time, and the anti-theft efficiency is extremely low in an unattended scene. Another category is intelligent schemes based on identification, such as verifying user rights by swiping a card, scanning a code, or biometric identification, in an attempt to control risk from the source of use. However, such schemes still face the following challenges in practical applications: 1. Under the complex outdoor environment conditions such as strong light, rain and snow, low temperature and the like, the response speed and the recognition accuracy of part of the recognition modules may be reduced; 2. most systems focus on the user identity verification link only, and cannot effectively link the identity recognition module with functions such as equipment state monitoring, abnormal alarm and the like, so that even if the identity recognition is passed, unauthorized interference or equipment theft in the charging process is difficult to prevent; 3. the partial scheme adopts a single identification mode, has limited environmental adaptability and anti-interference capability, and is difficult to meet the requirements of the charging pile on safety and reliability in all-weather operation environments. Disclosure of Invention The invention aims to overcome the defects of the prior art, and provides a face recognition intelligent charging pile anti-theft system, a face recognition intelligent charging pile anti-theft method and face recognition intelligent charging pile anti-theft equipment based on multi-feature fusion. In order to achieve the above purpose, the invention adopts the following technical scheme: Aiming at the defects of the prior art, the invention aims to provide the intelligent face recognition charging pile antitheft system based on multi-feature fusion, which effectively solves the antitheft problem faced by the existing charging pile by integrating an advanced face recognition technology and an intelligent antitheft mechanism, and improves charging safety and user experience. In order to achieve the above purpose, the invention is realized by the following technical scheme: A face recognition intelligent charging pile anti-theft system based on multi-feature fusion comprises a face recognition module, an intelligent lock control module, an alarm module, a power management module and a microcontroller; The face recognition module is used for collecting the face image to be recognized, detecting and positioning the face to be recognized, extracting the characteristics and calculating the similarity, and sending the face detection result and the face similarity data obtained by calculation to the microcontroller; the intelligent lock control module is used for responding to an unlocking or locking instruction of the microcontroller and driving and controlling the opening and closing of the charging pile lock; The alarm module is used for monitoring the physical state of the charging pile lock box and sending an abnormal state signal to the microcontroller when the unauthorized opening behavior is detected; The power management module is used for supplying power to the system, monitoring the power consumption and the activity state of the system in real time, generating monitoring data, executing a power management instruction sent by the microcontroller, and reporting the monitoring data to the microcontroller; The microcontroller is respectively connected with t