A comprehensive rapid detection system for dangerous goods security gates with facial recognition

Abstract

This invention discloses a comprehensive rapid detection system for a hazardous materials security gate with facial recognition, comprising a security gate body, a facial recognition module, a hazardous materials detection module, a central data processing module, an alarm module, and a human-computer interaction module, which work together. This invention integrates three major technology modules—facial recognition, gas chromatography-ion mobility spectrometry (GC-IMS) gas detection, and active X-ray fluorescence (XRF) metal detection—into a single modular gate structure. This fundamentally solves the shortcomings of traditional metal detector gates with their single function, enabling a single security gate to simultaneously complete three core tasks: identity verification, flammable and explosive material screening, and toxic metal detection. This covers a wider range of hazardous materials and significantly improves the breadth and depth of security measures.

Description

Technical Field

[0001] This invention relates to the field of public safety, and in particular to a comprehensive rapid detection system for a security gate for dangerous goods with facial recognition. Background Technology

[0002] Traditional metal detectors (metal security gates) are the most widely used technology. Based on the principle of electromagnetic induction, they can effectively detect metals such as iron and copper, but they are completely ineffective against non-metallic hazardous materials (such as plastic knives, ceramic weapons, liquid explosives, and powdered toxic agents). They also cannot distinguish between everyday metal items (such as keys and belt buckles) and potential weapons, resulting in a high false alarm rate. They require manual inspection, making them inefficient. To detect explosives and drugs, trace explosive / drug detectors have emerged, typically using ion mobility spectrometry (IMS) or mass spectrometry. These require manual wiping and sampling, taking tens of seconds to test one person, making them unsuitable for large crowds and posing a high risk of missed detection for airtight packaging or highly clean items. Detection of toxic heavy metals usually relies on laboratory analysis or expensive, cumbersome portable X-ray fluorescence (XRF) equipment, which cannot achieve rapid screening of personnel at passageways. Although millimeter-wave imaging and terahertz technology can achieve non-contact imaging of the body surface and detect hidden items, they are extremely expensive and have weak ability to identify the chemical composition of items, making it impossible to determine whether they are specific flammable, explosive or toxic metals.

[0003] A more prominent problem is the "information silo" state of existing technologies. The aforementioned devices typically operate independently, with detection results (alarm or no alarm) disconnected from the identity information of the inspected person. For example, if someone is detected carrying suspicious powder, their identity still requires secondary manual verification, making it impossible to achieve automatic correlation and rapid tracing of "person-object-event." Although some high-end venues have attempted to place facial recognition turnstiles alongside security gates, the two are not deeply integrated in terms of data and judgment logic; they are merely physically adjacent and cannot be correlated.

[0004] Therefore, this invention proposes a comprehensive rapid detection system for a security gate for dangerous goods with facial recognition. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a comprehensive and rapid detection system for a dangerous goods security gate with facial recognition.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A comprehensive rapid detection system for dangerous goods security gates with facial recognition includes: The security gate includes at least two cabinets arranged side by side and a top plate on top of the cabinets, forming an inspection channel between the cabinets. The cabinets are equipped with baffles to guide airflow. The top and bottom of the security gate are equipped with airflow systems that connect to the inspection channel to form a directional airflow from bottom to top within the inspection channel. A face recognition module is installed at the entrance of the inspection channel to capture the face image of the person passing through, extract the face feature vector, compare it with the pre-stored database, and output identity information or visitor identifier. The hazardous materials detection module includes a gas detection submodule and a metal detection submodule. The gas detection submodule includes a gas chromatography-ion mobility spectrometry sensor array arranged inside the inspection channel for sampling and analyzing volatile organic compounds in the directional gas flow and outputting a gas concentration vector. The metal detection submodule includes an active X-ray fluorescence sensor symmetrically arranged on both sides of the inspection channel for emitting X-rays to personnel passing through and receiving characteristic fluorescence to analyze toxic metal elements and outputting a metal intensity vector. The central data processing module is communicatively connected to the face recognition module and the hazardous materials detection module. It is used to receive identity information, gas concentration vector and metal strength vector, execute data fusion and decision-making algorithms, and calculate the comprehensive hazard index. An alarm module, which is communicatively connected to the central data processing module and the human-machine interaction module, is used to generate corresponding audible and visual alarm signals and display detection details based on the judgment result. The human-computer interaction module is used for human-computer interaction between the administrator and the system.

[0007] Preferably, the airflow system includes louvers at the air inlet at the bottom of the cabinet, a mesh cover at the air outlet at the top of the cabinet, and multiple exhaust fans at the top and bottom respectively. The exhaust fans at the bottom are used to draw in external air to form a directional airflow, and the exhaust fans at the top are used to extract the gas flowing through the inspection channel and guide it to the air intake of the gas detection submodule.

[0008] Preferably, the step of the face recognition module extracting and comparing facial feature vectors includes: A1: Use a deep convolutional neural network model to extract a feature vector v from the captured face image, and calculate the feature vector v and the feature vector v in the pre-stored database. i cosine similarity S(v, v) i ); A2: If the maximum similarity exceeds the first threshold T face If the identification is successful, the corresponding identity will be associated with it; otherwise, it will be marked as a visitor. The identity information includes the successfully identified identity or visitor identifier and its corresponding feature vector.

[0009] Preferably, the data fusion and decision-making algorithm is as follows: B1: Regarding the gas concentration vector and metal strength vector Normalization is performed to obtain the normalized vector. and The normalized component , and This is the preset upper limit of the safety threshold; B2: Determine the risk level based on identity information. If the identity is on the blacklist, then =1, if it is a whitelist, then =0, if it is a visitor, then The preset baseline risk value; B3: Calculate the overall risk index ,in As weight, max() is a function that takes the maximum value in a vector; B4: Compare R with the second threshold T alarm If R≥T alarm Then an alarm judgment result will be generated.

[0010] Preferably, the gas chromatography-ion mobility spectrometry (GC-IMS) sensor in the gas detection submodule is arranged in an array of multiple layers on the cabinet inside the test channel, with its inlet facing the path of the directional airflow, to capture trace amounts of gas desorbed from different heights of the human body. The GC-IMS sensor has a built-in characteristic spectral library of volatile gases of common flammable and explosive substances, and qualitative and semi-quantitative detection is achieved through comparative analysis.

[0011] Preferably, the active X-ray fluorescence sensor in the metal detection submodule includes a low-power miniature X-ray tube and a silicon drift detector. The energy of the X-ray beam emitted by the X-ray tube is adjustable to optimize the excitation of characteristic X-ray fluorescence generated by the target toxic metal element. The silicon drift detector measures the intensity I of the characteristic X-ray peak through energy spectrum analysis. i And according to the formula The relative abundance levels of the target toxic metal element were calculated, where w i The target element's quality fraction. σ is the incident X-ray flux. i denoted as the fluorescence excitation cross section, and k as the instrument constant.

[0012] Preferably, the alarm module includes a multi-color LED indicator located above the exit side of the security gate. The central data processing module controls the indicator based on the judgment result: a solid green light is displayed when the system is ready; when R≥T... alarm The red indicator light flashes and a high-decibel buzzer is triggered when... or Exceeds the corresponding warning threshold, but R < T alarm The yellow indicator light will flash when the light is on; otherwise, the green light will remain on.

[0013] Preferably, the human-computer interaction module (11) is a touch screen located near the security gate, used to display in real time the face image of the person passing through, the detection status, the real-time curves or values ​​of the gas concentration vector and the metal strength vector, the calculated value of the comprehensive danger index R, and the final judgment result and alarm type.

[0014] Preferably, the touch screen also provides a management function entry point for configuring weighting coefficients. First threshold T face Second threshold T alarm The upper limit of the security threshold and And check historical security check records.

[0015] A security inspection method for a comprehensive rapid detection system for dangerous goods using a security gate with facial recognition includes the following steps: S1: When a person approaches the inspection channel, the face recognition module is triggered to complete the face image acquisition, feature extraction and identity comparison, and generate a security inspection event bound with identity information; S2: When personnel enter the inspection channel, the airflow system is activated, creating a directional airflow around the personnel within the channel; S3: The gas detection submodule continuously collects and analyzes gas samples in the directional airflow to generate a gas concentration vector; at the same time, the metal detection submodule emits X-rays and detects fluorescence when people pass by to generate a metal intensity vector. S4: The central data processing module receives identity information, gas concentration vector, and metal strength vector, executes data fusion and decision-making algorithms, calculates the comprehensive risk index R, and generates a judgment result; S5: Based on the judgment result, drive the alarm module to emit the corresponding sound and light signals, and display a detailed detection report on the human-machine interaction module.

[0016] The beneficial effects of this invention are as follows: 1. This invention integrates three major technology modules—facial recognition, gas chromatography-ion mobility spectrometry (GC-IMS) gas detection, and active X-ray fluorescence (XRF) metal detection—into a single modular door structure. This fundamentally solves the shortcomings of traditional metal detector doors that have only one function, enabling a single security gate to simultaneously complete three core tasks: identity verification, screening for flammable and explosive materials, and detection of toxic metals. This covers a wider range of hazardous materials and significantly improves the breadth and depth of security.

[0017] 2. A carefully designed airflow system forms a stable "air curtain" within the passageway, which can actively and imperceptibly collect trace amounts of substances detached from a person's body and deliver them to the GC-IMS sensor. This replaces the traditional time-consuming manual wiping sampling. At the same time, low-dose XRF scanning can be completed instantly as people walk through. All detection processes are completed automatically within 2-3 seconds, without the need for personnel to stop and cooperate. The central data processing module automatically receives, integrates, and judges all data, driving audible and visual alarms. The entire process requires no manual intervention, achieving "non-contact, non-stop, and fully automatic" rapid security checks. It is particularly suitable for high-traffic scenarios during peak hours, effectively alleviating congestion.

[0018] 3. This invention does not simply add up the alarm results of each sensor using an "OR" logic. Instead, it introduces a risk coefficient based on identity information and normalizes the gas concentration and metal strength data. Then, it performs weighted fusion through configurable weights to calculate a quantitative comprehensive risk index R. This model allows the system to flexibly adapt to different scenario focuses (such as chemical zones focusing on gases and battery factories focusing on metals). It also comprehensively considers the risk attributes of "people" themselves, making it more scientific than a single threshold alarm. It can reduce false alarms while improving the detection rate of real complex threats.

[0019] 4. By binding the security check event ID and personnel identity through the facial recognition module, all subsequent gas and metal detection data and alarm results are associated with this unique ID. This allows any alarm to be immediately linked to a specific person and their facial image, enabling second-level traceability. At the same time, normal passage records are also retained, forming a complete security check log. This completely breaks the "information silo" state of traditional security checks where detection results and personnel identities are separated. It provides a solid data foundation for in-process early warning, post-event review, big data analysis, and security strategy optimization, realizing the informatization and intelligent management of the security check process. Attached Figure Description

[0020] Figure 1 This is a hardware structure diagram of a comprehensive rapid detection system for a dangerous goods security gate with facial recognition proposed in this invention; Figure 2This is a flowchart of a comprehensive rapid detection system for a dangerous goods security gate with facial recognition proposed in this invention.

[0021] In the picture: 1. Cabinet; 2. Top panel; 3. Face recognition module; 4. Alarm module; 5. Wire mesh cover; 6. Exhaust fan; 7. Louvers; 8. Gas chromatography-ion mobility spectrometry sensor; 9. Low-power miniature X-ray tube; 10. Silicon drift detector; 11. Human-computer interaction module. Detailed Implementation

[0022] The technical solution of the present invention will be further described in detail below with reference to specific embodiments.

[0023] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "setting" should be interpreted broadly. For example, they can refer to a fixed connection or setting, a detachable connection or setting, or an integral connection or setting. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0024] Example 1: A comprehensive rapid detection system for dangerous goods security gates with facial recognition includes: The security gate includes at least two cabinets 1 arranged side by side and a top plate 2 on the top of the cabinets 1. An inspection channel is formed between the cabinets 1. The cabinets 1 are provided with a baffle plate for guiding airflow. The top and bottom of the security gate are provided with an airflow system that connects to the inspection channel to form a directional airflow from bottom to top in the inspection channel. The face recognition module 3 is located at the entrance of the inspection channel and is used to capture the face image of the person passing through, extract the face feature vector, compare it with the pre-stored database, and output identity information or visitor identifier. The hazardous materials detection module includes a gas detection submodule and a metal detection submodule. The gas detection submodule includes a gas chromatography-ion mobility spectrometry sensor 8 arrayed inside the inspection channel, used to sample and analyze volatile organic compounds in the directional gas flow and output a gas concentration vector. The metal detection submodule includes an active X-ray fluorescence sensor symmetrically arranged on both sides of the inspection channel, used to emit X-rays to personnel passing through and receive characteristic fluorescence, analyze toxic metal elements, and output a metal intensity vector. The central data processing module is communicatively connected to the face recognition module and the hazardous materials detection module. It is used to receive identity information, gas concentration vector and metal strength vector, execute data fusion and decision-making algorithms, and calculate the comprehensive hazard index. The alarm module 4 and the human-machine interaction module 11 are communicatively connected to the central data processing module, and are used to generate corresponding audible and visual alarm signals and display detection details based on the judgment result. Human-computer interaction module 11, which is used for human-computer interaction between the administrator and the system.

[0025] Preferably, the airflow system includes a louver 7 at the bottom air inlet of the cabinet 1, a mesh cover 5 at the top air outlet of the cabinet 1, and multiple exhaust fans 6 at the top and bottom respectively. The exhaust fan 6 at the bottom is used to draw in external air to form a directional airflow, and the exhaust fan 6 at the top is used to extract the gas flowing through the inspection channel and guide it to the air intake of the gas detection submodule.

[0026] Preferably, the step of the face recognition module 3 extracting and comparing facial feature vectors includes: A1: Use a deep convolutional neural network model to extract a feature vector v from the captured face image, and calculate the feature vector v and the feature vector v in the pre-stored database. i cosine similarity S(v, v) i ); A2: If the maximum similarity exceeds the first threshold T face If the identification is successful, the corresponding identity will be associated with it; otherwise, it will be marked as a visitor. The identity information includes the successfully identified identity or visitor identifier and its corresponding feature vector.

[0027] Preferably, the data fusion and decision-making algorithm is as follows: B1: Regarding the gas concentration vector and metal strength vector Normalization is performed to obtain the normalized vector. and The normalized component , and This is the preset upper limit of the safety threshold; B2: Determine the risk level based on identity information. If the identity is on the blacklist, then =1, if it is a whitelist, then =0, if it is a visitor, then The preset baseline risk value; B3: Calculate the overall risk index ,in As weight, max() is a function that takes the maximum value in a vector; B4: Compare R with the second threshold T alarm If R≥T alarmThen an alarm judgment result will be generated.

[0028] Preferably, the gas chromatography-ion mobility spectrometry sensor 8 in the gas detection submodule is arranged in an array of multiple layers on the cabinet inside the test channel, with its inlet facing the path of the directional airflow, to capture trace amounts of gas desorbed from different heights of the human body. The gas chromatography-ion mobility spectrometry sensor 8 has a built-in characteristic spectral library of volatile gases of common flammable and explosive substances, and qualitative and semi-quantitative detection is achieved through comparative analysis.

[0029] Preferably, the active X-ray fluorescence sensor in the metal detection submodule includes a low-power miniature X-ray tube 9 and a silicon drift detector 10. The X-ray beam energy emitted by the X-ray tube 9 is adjustable to optimize the excitation of characteristic X-ray fluorescence generated by the target toxic metal element. The silicon drift detector 10 measures the intensity I of the characteristic X-ray peak through energy spectrum analysis. i And according to the formula The relative abundance levels of the target toxic metal element were calculated, where w i The target element's quality fraction. σ is the incident X-ray flux. i denoted as the fluorescence excitation cross section, and k as the instrument constant.

[0030] Preferably, the alarm module 4 includes a multi-color LED indicator located above the exit side of the security gate. The central data processing module controls the indicator based on the judgment result: a solid green light is displayed when the system is ready; when R≥T... alarm The red indicator light flashes and a high-decibel buzzer is triggered when... or Exceeds the corresponding warning threshold, but R < T alarm The yellow indicator light will flash when the light is on; otherwise, the green light will remain on.

[0031] Preferably, the human-computer interaction module 11 is a touch screen located near the security gate, used to display in real time the face image of the person passing through, the detection status, the real-time curves or values ​​of the gas concentration vector and the metal strength vector, the calculated value of the comprehensive hazard index R, and the final judgment result and alarm type.

[0032] Preferably, the touch screen also provides a management function entry point for configuring weighting coefficients. First threshold T face Second threshold T alarm The upper limit of the security threshold and And check historical security check records.

[0033] Example 2: A security inspection method for a comprehensive rapid detection system for dangerous goods using a security gate with facial recognition includes the following steps: S1: When a person approaches the inspection channel, the face recognition module is triggered to complete the face image acquisition, feature extraction and identity comparison, and generate a security inspection event bound with identity information; S2: When personnel enter the inspection channel, the airflow system is activated, creating a directional airflow around the personnel within the channel; S3: The gas detection submodule continuously collects and analyzes gas samples in the directional airflow to generate a gas concentration vector; at the same time, the metal detection submodule emits X-rays and detects fluorescence when people pass by to generate a metal intensity vector. S4: The central data processing module receives identity information, gas concentration vector, and metal strength vector, executes data fusion and decision-making algorithms, calculates the comprehensive risk index R, and generates a judgment result; S5: Based on the judgment result, drive the alarm module to emit the corresponding sound and light signals, and display a detailed detection report on the human-machine interaction module.

[0034] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A comprehensive rapid detection system for a dangerous goods security gate with facial recognition, characterized in that, include: The security gate body includes at least two cabinets (1) arranged side by side and a top plate (2) set on the top of the cabinets (1). An inspection channel is formed between the cabinets (1). A baffle plate for guiding airflow is provided inside the cabinets (1). An airflow system connecting the top and bottom of the security gate body is provided to form a directional airflow from bottom to top in the inspection channel. The face recognition module (3) is located at the entrance of the inspection channel and is used to capture the face image of the person passing through, extract the face feature vector, compare it with the pre-stored database, and output identity information or visitor identifier. The hazardous materials detection module includes a gas detection submodule and a metal detection submodule. The gas detection submodule includes a gas chromatography-ion mobility spectrometry sensor (8) arrayed inside the inspection channel, used to sample and analyze volatile organic compounds in the directional gas flow and output a gas concentration vector. The metal detection submodule includes an active X-ray fluorescence sensor symmetrically arranged on both sides of the inspection channel, used to emit X-rays to the personnel passing through and receive characteristic fluorescence, analyze toxic metal elements, and output a metal intensity vector. The central data processing module is communicatively connected to the face recognition module and the hazardous materials detection module. It is used to receive identity information, gas concentration vector and metal strength vector, execute data fusion and decision-making algorithms, and calculate the comprehensive hazard index. The alarm module (4) and the human-machine interaction module (11) are communicatively connected to the central data processing module, and are used to generate corresponding sound and light alarm signals and display detection details according to the judgment result; Human-computer interaction module (11), which is used for human-computer interaction between the administrator and the system.

2. The integrated rapid detection system for a dangerous goods security gate with facial recognition as described in claim 1, characterized in that, The airflow system includes a louver (7) at the bottom air inlet of the cabinet (1), a mesh cover (5) at the top air outlet of the cabinet (1), and multiple exhaust fans (6) at the top and bottom respectively. The exhaust fan (6) at the bottom is used to draw in external air to form a directional airflow, and the exhaust fan (6) at the top is used to extract the gas flowing through the inspection channel and guide it to the air intake of the gas detection submodule.

3. The integrated rapid detection system for a dangerous goods security gate with facial recognition as described in claim 1, characterized in that, The steps of the face recognition module (3) to extract and compare facial feature vectors include: A1 : extracting a feature vector v from a captured face image using a deep convolutional neural network model, computing a cosine similarity S(v, v i ) of this feature vector v with a feature vector v i stored in a pre-stored database; A2: If the maximum similarity exceeds a first threshold T face then the recognition is determined to be successful and the corresponding identity is associated, otherwise the person is marked as a visitor; The identity information includes the successfully identified identity or visitor identifier and its corresponding feature vector.

4. The integrated rapid detection system for a dangerous goods security gate with facial recognition as described in claim 1, characterized in that, The data fusion and decision-making algorithm is specifically as follows: B1: Regarding the gas concentration vector and metal strength vector Normalization is performed to obtain the normalized vector. and The normalized component , and This is the preset upper limit of the safety threshold; B2: Determine the risk level based on identity information. If the identity is on the blacklist, then =1, if it is a whitelist, then =0, if it is a visitor, then The preset baseline risk value; B3: Calculate the overall risk index ,in As weight, max() is a function that takes the maximum value in a vector; B4: compare R with a second threshold value T alarm , if R ≥ T alarm , then generate an alarm decision result.

5. The integrated rapid detection system for a dangerous goods security gate with facial recognition as described in claim 1, characterized in that, The gas chromatography-ion mobility spectrometry sensor (8) in the gas detection submodule is arranged in an array of multiple layers on the cabinet inside the test channel. Its inlet faces the path of the directional airflow and is used to capture trace amounts of gas desorbed from different heights of the human body. The gas chromatography-ion mobility spectrometry sensor (8) has a built-in characteristic spectrum library of volatile gases of common flammable and explosive substances, and qualitative and semi-quantitative detection is achieved through comparative analysis.

6. The integrated rapid detection system for a dangerous goods security gate with facial recognition as described in claim 1, characterized in that, The active X-ray fluorescence sensor in the metal detection submodule includes a low-power miniature X-ray tube (9) and a silicon drift detector (10). The X-ray beam energy emitted by the X-ray tube (9) is adjustable to optimize the excitation of characteristic X-ray fluorescence generated by the target toxic metal element. The silicon drift detector (10) measures the intensity I of the characteristic X-ray peak through energy spectrum analysis. i And according to the formula The relative abundance levels of the target toxic metal element were calculated, where w i The target element's quality fraction. σ is the incident X-ray flux. i denoted as the fluorescence excitation cross section, and k as the instrument constant.

7. The integrated rapid detection system for a dangerous goods security gate with facial recognition as described in claim 1, characterized in that, The alarm module (4) includes a multi-color LED indicator located above the exit side of the security gate. The central data processing module controls the indicator according to the judgment result: when the system is ready, it displays a solid green light; when R≥T alarm The red indicator light flashes and a high-decibel buzzer is triggered when... or Exceeds the corresponding warning threshold, but R < T alarm The yellow indicator light will flash when the light is on; otherwise, the green light will remain on.

8. The integrated rapid detection system for a dangerous goods security gate with facial recognition as described in claim 1, characterized in that, The human-computer interaction module (11) is a touch screen located near the security gate, used to display in real time the face image of the person passing through, the detection status, the real-time curves or values ​​of the gas concentration vector and the metal strength vector, the calculated value of the comprehensive hazard index R, and the final judgment result and alarm type.

9. A comprehensive rapid detection system for a dangerous goods security gate with facial recognition as described in claim 8, characterized in that, The touchscreen display also provides a management function entry point for configuring weighting coefficients. First threshold T face Second threshold T alarm The upper limit of the security threshold and And check historical security check records.

10. A security inspection method based on the integrated rapid detection system of the dangerous goods security gate with facial recognition as described in any one of claims 1-9, characterized in that, Includes the following steps: S1: When a person approaches the inspection channel, the face recognition module is triggered to complete the face image acquisition, feature extraction and identity comparison, and generate a security inspection event bound with identity information; S2: When personnel enter the inspection channel, the airflow system is activated, creating a directional airflow around the personnel within the channel; S3: The gas detection submodule continuously collects and analyzes gas samples in the directional airflow to generate a gas concentration vector; at the same time, the metal detection submodule emits X-rays and detects fluorescence when people pass by to generate a metal intensity vector. S4: The central data processing module receives identity information, gas concentration vector, and metal strength vector, executes data fusion and decision-making algorithms, calculates the comprehensive risk index R, and generates a judgment result; S5: Based on the judgment result, drive the alarm module to emit the corresponding sound and light signals, and display a detailed detection report on the human-machine interaction module.

Images