CN113823027A

CN113823027A - Intelligent non-inductive passing control system

Info

Publication number: CN113823027A
Application number: CN202111058285.0A
Authority: CN
Inventors: 张行; 姚信威; 陈树; 葛广志; 邢伟伟
Original assignee: Zhejiang Huixiang Information Technology Co ltd
Current assignee: Zhejiang Huixiang Information Technology Co ltd
Priority date: 2021-09-09
Filing date: 2021-09-09
Publication date: 2021-12-21

Abstract

The invention relates to an intelligent non-inductive traffic control system, which detects at least two characteristic sources of a user to be passed by a detection unit comprising at least two detection data acquisition mechanisms, processes multi-source detection data of the user to be passed by a pre-processing unit, interactively verifies the processed detection result by a control unit, outputs a passing instruction to the user to be passed by the verification, and realizes non-inductive traffic control. The intelligent community control method and the intelligent community control system solve the contradiction problem in the existing intelligent community setting, carry out non-inductive passing control on the user in a multi-source detection data combination mode, solve the problem that the existing intelligent community is not friendly to the application of the old people, and guarantee the safety of data on the premise of ensuring the application convenience and passing non-inductive.

Description

Intelligent non-inductive passing control system

Technical Field

The invention relates to the technical field of single input port or output port registers, in particular to an intelligent non-inductive traffic control system.

Background

With the rapid development of the internet of things technology and the national support and advocate for the intelligent community, the community service system is more complete and intelligent.

However, there are many problems in the application process of the smart community, including:

(1) for the communities with large occupation ratios of the middle-aged and the young, it is not difficult to implement intelligent communities and intelligent identification, and how to protect the information security of residents and users is difficult, especially when biological feature acquisition is involved, on one hand, the acceptance of the users needs to be ensured, and on the other hand, the confidentiality and the non-leakage of the information need to be ensured, so that practical application difficulty exists;

(2) for the community that the old person accounts for a relatively large proportion, except the existence of the above-mentioned problem, there is the invariance of using still, especially for the old person who can't understand thing networking science and technology, the operation degree of difficulty that face identification, cell-phone bluetooth control were opened the door, NFC swiped the card and are opened the door is huge undoubtedly, and is not high-efficient, intelligent.

How to effectively improve the convenience of the user in trip and the intellectualization of the access control equipment is worthy of exploration and urgent solution.

Disclosure of Invention

The invention solves the problems in the prior art and provides an optimized intelligent sensorless traffic control system.

The invention adopts the technical scheme that an intelligent non-inductive traffic control system comprises:

the detection unit comprises at least two detection data acquisition mechanisms and is used for detecting at least two feature sources of the user to be passed;

the pre-processing unit is matched with the detection unit and is used for processing multi-source detection data of the user to be passed;

and the control unit is matched with the front processing unit and used for carrying out interactive verification on the processed detection result and outputting a release instruction to the user to be passed who passes the verification so as to realize the non-inductive passing control.

Preferably, the detection unit comprises a communication module assembly, and the detection unit further comprises a camera, a microphone and a fingerprint collector.

Preferably, the cooperation be equipped with face identification module and/or iris recognition module in the leading processing unit of camera, the cooperation the leading processing unit of microphone is equipped with the voiceprint recognition module, and the cooperation fingerprint sampler's leading processing unit is equipped with the fingerprint recognition module, the cooperation the leading processing unit of communication module subassembly is equipped with equipment scanning module.

Preferably, the communication module assembly comprises a bluetooth transceiver and/or a radio frequency transceiver.

Preferably, the communication module component sends a communication connection request at a preset frequency, and the equipment scanning module completes the connection with the communication module component after receiving the request and successfully matching; and the equipment scanning module acquires the MAC address of the connected communication module assembly.

Preferably, the interactive authentication comprises the steps of:

step a.1: the communication module component is in communication connection with any communication component carried by a user;

step a.2: after the connection is successful, the communication component encrypts the characteristic information of the communication component by using a private key and sends the encrypted characteristic information and the identifier to the preprocessing unit;

step a.3: the preposed processing unit encrypts the encrypted characteristic information and the verification code by a public key and transmits the encrypted characteristic information and the encrypted verification code to the control unit;

step a.4: the control unit decrypts the data packet transmitted in the step 3 by using a public key to obtain encrypted characteristic information and an identifier, traverses a verifier corresponding to the communication component prestored in the control unit by using the identifier, and if a matchable item exists, performs the next step, otherwise, fails in verification and exits;

step a.5: taking a private key corresponding to the matched verifier, and decrypting the encrypted characteristic information; if the decrypted characteristic information is matched with the characteristic information corresponding to the identifier, extracting the other characteristics matched with the identifier at present, and carrying out the next step, otherwise, failing to verify and quitting;

step a.6: and the pre-processing unit processes other multi-source detection data of the user to be passed, transmits the characteristics to the control unit, if any characteristic corresponds to the characteristic matched with the identifier, the verification is successful, otherwise, the verification fails, and the operation is exited.

Preferably, in the step a.2, the characteristic information is a MAC address of the processed device.

Preferably, a barrier for blocking or releasing the user is arranged in cooperation with the detection unit, and a counting module is further arranged in cooperation with the barrier.

Preferably, the counting module is a thermal imaging device, and the counting of the thermal imaging device comprises the following steps:

step b.1: a user carries out thermal sensing imaging through a barrier gate and a thermal sensing imaging device;

step b.2: if the abnormal temperature person exists, directly alarming and stopping the whole body, otherwise, directly carrying out the next step;

step b.3: the detection unit sends the thermal imaging video frame of which the barrier gate is opened to fall to the preprocessing unit;

step b.4: the front-end processing unit processes the image of each frame of thermal imaging, performs anchor frame setting based on body temperature, and performs image deletion at a certain frequency;

step b.5: packaging the deleted images and sending the images to a control unit;

step b.6: the control unit processes the anchor frames in all the images, traces the trajectories of the persons and determines the number of the passers.

Preferably, the control unit pushes the determined number of the passing people to the user.

The invention relates to an optimized intelligent non-inductive traffic control system, which detects at least two characteristic sources of a user to be passed by a detection unit comprising at least two detection data acquisition mechanisms, processes multi-source detection data of the user to be passed by a pre-processing unit, interactively verifies the processed detection result by a control unit, outputs a passing instruction to the user to be passed by verification, and realizes non-inductive traffic control.

The intelligent community control method and the intelligent community control system solve the contradiction problem in the existing intelligent community setting, carry out non-inductive passing control on the user in a multi-source detection data combination mode, solve the problem that the existing intelligent community is not friendly to the application of the old people, and guarantee the safety of data on the premise of ensuring the application convenience and passing non-inductive.

Drawings

FIG. 1 is a schematic structural diagram of the present invention, wherein arrows indicate the direction of information transmission;

FIG. 2 is a flowchart illustrating the operation of interactive authentication in the present invention;

FIG. 3 is a flow chart of the operation of the counting module of the present invention;

FIG. 4 is a flow chart of the operation of the voice recognition module of the present invention.

Detailed Description

The present invention is described in further detail with reference to the following examples, but the scope of the present invention is not limited thereto.

The invention relates to an intelligent non-inductive traffic control system, which comprises:

The detection unit comprises a communication module assembly, and further comprises a camera, a microphone and a fingerprint collector.

The cooperation be equipped with face identification module and/or iris identification module in the leading processing unit of camera, the cooperation the leading processing unit of microphone is equipped with voiceprint identification module, and the cooperation fingerprint sampler's leading processing unit is equipped with fingerprint identification module, the cooperation the leading processing unit of communication module subassembly is equipped with equipment scanning module.

The communication module assembly comprises a Bluetooth transceiver and/or a radio frequency transceiver.

The communication module component sends a communication connection request at a preset frequency, and the equipment scanning module receives the request and completes the connection with the communication module component after successful matching; and the equipment scanning module acquires the MAC address of the connected communication module assembly.

In the invention, the detection result of the communication module assembly in the detection unit is used as main detection data, and the detection result of equipment including but not limited to a camera, a microphone and a fingerprint collector is used as auxiliary detection data.

In the invention, the Bluetooth transceiver and/or the radio frequency transceiver are/is equipment which is arranged at a user position and a detection end and has a transceiving function.

Firstly, recording the MAC address of hardware equipment of a user in a detection unit, scanning the address of nearby MAC equipment through Bluetooth transceiving equipment and/or radio frequency transceiving equipment when a camera, a microphone and a fingerprint collector obtain a trigger identification instruction, and further processing and comparing data collected by the camera, the microphone and the fingerprint collector after the MAC addresses are matched; the MAC address can be applied to realize coverage of various crowds through different hardware, and meanwhile, biological characteristic data are collected in a mode acceptable by different crowds to complete recheck of characteristics, so that the adaptability and the universality of the pass verification are ensured.

In the invention, the face recognition module, the iris recognition module and the fingerprint recognition module are common technologies in the prior art, and can be set by a person skilled in the art according to requirements.

In the present invention, for the voice Recognition module, as a part of voice Recognition, in this embodiment, a Voiceprint Recognition (VPR) is adopted to realize a function, which may also be called Speaker Recognition (Speaker Recognition), and is a process of recognizing a specific Speaker from different Speaker sets, and mainly studies the identification problem of the Speaker;

the voiceprint recognition and other biometric performance pairs are shown in table 1;

TABLE 1 comparison of voiceprint recognition with other biometric recognition performance

Feature(s)

Finger print

Palm print

Retina

Iris (iris)

Human face

Voiceprint

Ease of use

Height of

Is low in

Medium and high grade

Mountain, etc

Height of

Rate of accuracy

Height of

Medium and high grade

Height of

Medium and high grade

Height of

Cost of

Medium and high grade

Is very high

Mountain, etc

Is low in

Difficulty of sampling

Is low in

Height of

Is low in

Degree of user acceptance

Medium and high grade

Height of

Remote authentication

Must not

Can be used for

Mobile phone sampling

Is partially made of

Must not

Can be used for

It is thus clear that the ease of use of using the voiceprint to carry out authentication is strong, the sampling is simple, the acceptance is wide, and people's pronunciation has the uniqueness, be difficult to by the emulation, pronunciation do not expose and do not forget and steal easily like the password outside like facial yet, have certain security, it is very convenient to use pronunciation to judge the identity, basically as long as say one to two sentences can judge the identity, it is more practical to old person, the people who has not experienced education and disabled person, can be used for user's authentication with other safety inspection together.

In the invention, the voice recognition module mainly comprises three functional units, namely voice acquisition, voice processing, training and recognition; the voice acquisition is completed by voice acquisition equipment, the front-end processing and the feature processing are two components of voice processing, and in the actual processing, the front-end processing comprises pre-emphasis of tone of voice, voice detection (end point detection), noise suppression (voice enhancement) and the like; and the training and recognition of the voice are completed by a Deep Neural Network (DNN).

In the invention, the deep neural network can be regarded as a multilayer perceptron with a plurality of hidden layers, nodes between two adjacent layers are fully connected, and data is transmitted between each layer through parameters such as an activation function, a weight, a bias and the like to finally output a prediction result; namely, the voice acquisition equipment performs preprocessing and feature extraction after acquiring voice in the training stage, performs model training after registering voiceprints, establishes voiceprints corresponding to each person (individual) through comparison of a voiceprint library, performs the same preprocessing and feature extraction after acquiring voice in the actual comparison stage, calls the voiceprints corresponding to each person from the voiceprint library after identifying the voiceprints, and obtains comparison results.

In the invention, an implementation mode is given:

and (3) voiceprint input: directly inputting a voice signal as a tag array, detecting the voice signal by an endpoint, obtaining pure voice after voice enhancement, and storing the pure voice in a data set in a wave waveform form; setting frame length, framing the voice signal, wherein the frame number is the number of neurons in an input layer; determining the number of speakers as the target classification number k of DNN according to different names of the speakers in the voice library, namely, the output is divided into k types, which are defined as:

thus, the voiceprint signal forms a sample pair (x, y) between the speaker and the voiceprint thereof, wherein x is the amplitude of the voice and y is the speaker tag;

DNN model establishment: building a network parameter and sensor structure which needs to meet input and output, weight, bias and the like of a DNN, and comprises the following steps:

(1) the input is a voiceprint signal, and the output is the name of a speaker;

(2) the setting of network parameters is an important link for building a DNN model, and reasonable initialization parameters can effectively reduce training steps, save time and optimize a training result;

(3) the DNN model selects a sigmoid function as an activation function in a hidden layer, a softmax activation function is selected as a regression model in an output layer, an adam optimization algorithm is used as an optimizer to replace a traditional gradient descent algorithm, and a Dropout link is added to prevent an overfitting phenomenon;

training is performed after the setup is completed, as shown in fig. 4:

the weight parameter W and the bias parameter b in the DNN model need to be obtained by training samples, assuming that the total number of layers L, the maximum number of iterations is represented by Max (300 in this model), the number of input samples is m, Iter is the current number of iterations, n and x are the identifiers of the number of layers and the number of samples in the iteration process, respectively, and the training steps are as follows:

(1) inputting the total number of layers, the number of units of each layer, an activation function, a loss function, the maximum iteration number and the total number of samples;

(2) if the iteration number is not reached, the following loop is performed:

a. calculating each neuron output by forward propagation from a first hidden layer;

b. calculating the loss of an output layer;

c. respectively calculating the loss of each neuron of each layer from the last layer to the back propagation; reading the next sample, executing a until all samples are executed, and executing step d;

d. updating the weight and the offset of each layer from the first layer;

(3) when the iteration times are reached, outputting the weight and bias coefficient matrix of each layer, and if the iteration times are not reached, returning to the step (2);

identification: and collecting voice information of the personnel to be identified, uploading the voice information to a system, and carrying out personnel matching through a bone voiceprint identification system based on DNN.

The interactive authentication comprises the following steps:

in the step a.2, the characteristic information is the MAC address of the processed device.

In the invention, taking Bluetooth transceiver as an example, when the device scanning module tries to communicate with the Bluetooth transceiver carried by the user; after the connection is successful, the communication component encrypts the characteristic information of the communication component and the processed MAC address of the equipment by using a private key, and sends the encrypted MAC address and the identifier to the preprocessing unit, namely, the preprocessing unit cannot obtain the MAC address information of the user; the pre-processing unit encrypts the encrypted characteristic information and the verification code by a public key and then transmits the encrypted characteristic information and the encrypted verification code to the control unit, namely, secondary encryption; the control unit decrypts by using a public key, at the moment, the control unit obtains the encrypted characteristic information and the identifier, traverses the identifier corresponding to the communication component prestored in the control unit, if no matchable item exists, the equipment is probably a 'black-in' equipment, and when the matchable item exists, the encrypted characteristic information is decrypted by using a corresponding private key; based on the matching of the decrypted characteristic information and the characteristic information corresponding to the identifier, the front-end processing unit processes other multi-source detection data of the user to be passed and transmits the characteristics to the control unit.

The detection unit is provided with a barrier for blocking or releasing users, and the barrier is also provided with a counting module.

The counting module is a thermal imaging device, and the counting of the thermal imaging device comprises the following steps:

And the control unit pushes the determined number of the passing people to the user.

According to the invention, counting and heat sensation are integrated based on the current special social situation, and counting of personnel entering is completed while the body temperature is tested; and on the premise that the body temperature of the personnel is normal, tracking and counting the anchor frame of the person entering each wave, pushing the counting result to the user, and if the user thinks that the counting is wrong, feeding back and reversely managing the property.

Claims

1. The utility model provides an intelligent noninductive traffic control system which characterized in that: the system comprises:

2. The intelligent sensorless traffic control system according to claim 1, wherein: the detection unit comprises a communication module assembly, and further comprises a camera, a microphone and a fingerprint collector.

3. The intelligent sensorless traffic control system according to claim 2, wherein: the cooperation be equipped with face identification module and/or iris identification module in the leading processing unit of camera, the cooperation the leading processing unit of microphone is equipped with voiceprint identification module, and the cooperation fingerprint sampler's leading processing unit is equipped with fingerprint identification module, the cooperation the leading processing unit of communication module subassembly is equipped with equipment scanning module.

4. The intelligent sensorless traffic control system according to claim 2, wherein: the communication module assembly comprises a Bluetooth transceiver and/or a radio frequency transceiver.

5. The intelligent sensorless traffic control system according to claim 3, wherein: the communication module component sends a communication connection request at a preset frequency, and the equipment scanning module receives the request and completes the connection with the communication module component after successful matching; and the equipment scanning module acquires the MAC address of the connected communication module assembly.

6. The intelligent sensorless traffic control system according to claim 2, wherein: the interactive authentication comprises the following steps:

7. The intelligent sensorless traffic control system according to claim 6, wherein: in the step a.2, the characteristic information is the MAC address of the processed device.

8. The intelligent sensorless traffic control system according to claim 1, wherein: the detection unit is provided with a barrier for blocking or releasing users, and the barrier is also provided with a counting module.

9. The intelligent sensorless traffic control system according to claim 8, wherein: the counting module is a thermal imaging device, and the counting of the thermal imaging device comprises the following steps:

10. The intelligent sensorless traffic control system according to claim 9, wherein: and the control unit pushes the determined number of the passing people to the user.