A bullet cabinet with alcohol detection function

Abstract

The application provides a bullet cabinet with alcohol detection function, and the alcohol detection system comprises: a data acquisition module which acquires detection video of a cabinet opening personnel through a camera, acquires detection voice of the cabinet opening personnel through a microphone, acquires historical cabinet opening data, and acquires current time data; a personnel state determination module which processes the detection video, the detection voice, the historical cabinet opening data, and the current time data to obtain a face alcohol state score, an eye state abnormality score, a body action amplitude score, a voice instability score, and a time risk score; a risk index determination module which determines a final risk index according to the face alcohol state score, the eye state abnormality score, the body action amplitude score, the voice instability score, and the time risk score; and a cabinet lock control module which controls the opening and closing state of an intelligent cabinet lock according to the final risk index. The alcohol detection precision of the cabinet opening personnel is improved, and the risk of using a gun by the cabinet opening personnel is improved.

Description

Technical Field

[0001] This invention relates to the field of alcohol monitoring technology, and specifically to a gun cabinet with alcohol detection function. Background Technology

[0002] Gun cabinets are containers used to store firearms and ammunition. Firearms and ammunition have significant lethality and require careful maintenance and safe storage to prevent wear and tear or theft.

[0003] Most existing gun and ammunition cabinets require manual access by opening the door, and their security features are relatively basic. Gun and ammunition cabinets originated from safes and were primarily designed for the proper management of firearms and ammunition. As a tool for firearms and ammunition management, they are an important component of the unified management system used by judicial departments. While their security and reliability have led to their widespread use, the ease with which some internal personnel and police officers can open them, especially those under the influence of alcohol, poses a significant risk and thus restricts their use. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a gun and ammunition cabinet with alcohol detection function, aiming to improve the accuracy of alcohol detection for personnel opening the cabinet and reduce the risk of personnel using firearms.

[0005] In a first aspect, embodiments of this application provide a gun cabinet with an alcohol detection function. The gun cabinet includes a camera, a speaker, an alcohol detection system, and a smart lock. The alcohol detection system includes a voice command output module, a data acquisition module, a personnel status determination module, a risk index determination module, and a lock control module. The alcohol detection system includes: The data acquisition module acquires detection video of the person opening the cabinet through a camera, detection voice of the person opening the cabinet through a microphone, historical cabinet opening data, and current time data. The personnel status determination module processes the detection video, the detection voice, the historical cabinet opening data, and the current time data to obtain facial alcohol status score, eye abnormality score, body movement amplitude score, voice instability score, and time risk score. The risk index determination module determines the final risk index based on facial alcohol status score, abnormal eye condition score, body movement amplitude score, voice instability score, and time risk score. The cabinet lock control module controls the opening and closing status of the smart cabinet lock based on the final risk index.

[0006] Optionally, the personnel status determination module includes the following steps: The detection video is processed to obtain facial alcohol status score, eye abnormality score, and body movement amplitude score; The detected speech is processed to obtain a speech instability score; The historical data on opening cabinets and the current time data are processed to obtain a time risk score.

[0007] Optionally, the step of processing the detected video to obtain a facial alcohol status score, an eye condition abnormality score, and a body movement amplitude score includes: The detected video is input into the face alcohol status recognition model for processing to obtain a face alcohol status score. The face alcohol status recognition model is an improved ResNet-18 artificial network model that has been trained. Based on the detected video, an abnormal eye condition score is obtained; The human pose estimation model is used to identify joint points in the detected video to obtain human joint point data. Based on the data of the human joint points, a score for the range of motion of the body is obtained.

[0008] Optionally, the step of obtaining an eye condition abnormality score based on the detected video includes: Based on the detected video, extract key eye point data for each frame; Blink frequency, eyelid opening ratio, and fixation stability are calculated based on key eye point data. The blink frequency, the eyelid opening ratio, and the gaze stability are normalized and weighted to obtain an abnormal eye condition score.

[0009] Optionally, the step of normalizing and weighting the blink frequency, the eyelid opening ratio, and the fixation stability to obtain an ocular state abnormality score includes: ; in, Score abnormal eye condition; The normalized blink frequency; This represents the normalized eyelid opening ratio; This represents the gaze stability after normalization.

[0010] Optionally, the body movement amplitude score includes an arm swing amplitude score, a body swing amplitude score, and a head swing amplitude score. The step of obtaining the body movement amplitude score based on the human joint data includes: The human joint data is smoothed and normalized to obtain left wrist trajectory data, right wrist joint data, trunk center point trajectory data, and nose tip trajectory data. Based on the left wrist trajectory data and the right wrist joint data, an arm swing amplitude score is obtained; Based on the trajectory data of the torso center point, a score for the body swing amplitude is obtained; Based on the trajectory data of the nose tip, a score for the head sway amplitude is obtained.

[0011] Optionally, the step of processing the historical opening data and the current time data to obtain the time risk score includes: Based on the historical cabinet opening data, a historical behavior score is determined; Determine the current time score based on the current time data; A time risk score is obtained based on the historical behavior score and the current time score.

[0012] Optionally, the risk index determination module includes the following steps: A physiological aggregate score is obtained based on the facial alcohol status score (Fscore) and the eye condition abnormality score (Escore). The behavior aggregate score is obtained based on the body movement amplitude score Ascore, Tscore, Hscore and speech instability score Vscore; The final risk index is obtained based on the physiological aggregate score, the behavioral aggregate score, and the time risk score.

[0013] Optionally, the step of obtaining the final risk index based on the physiological aggregate score, the behavioral aggregate score, and the time risk score includes: ; in; This is the final risk index; Sensitivity coefficient; For physiological aggregation score; Aggregate scores for behaviors; The first polymerization coefficient; The second polymerization coefficient; This is a value indicating a medium risk level. Score the time risk.

[0014] Optionally, the gun cabinet with alcohol detection function also includes an infrared sensor and a speaker, and the alcohol detection system also includes an infrared sensing module and a voice command output module; The infrared sensing module acquires infrared sensing signals through an infrared sensor and detects whether the infrared sensing signals are greater than a preset infrared threshold. If the infrared sensing signal is greater than the preset infrared threshold, the voice command output module sends a voice prompt command to the speaker so that the speaker can broadcast the voice prompt command so that the person opening the cabinet can perform the corresponding action according to the voice prompt command. If the infrared sensing signal is less than or equal to the preset infrared threshold, the voice command output module will not send a voice prompt command to the speaker.

[0015] In this embodiment, the data acquisition module acquires detection video of the person opening the cabinet via a camera, detection voice via a microphone, historical cabinet opening data, and current time data. The personnel status determination module processes the detection video, detection voice, historical cabinet opening data, and current time data to obtain facial alcohol status score, eye abnormality score, body movement amplitude score, voice instability score, and time risk score. The risk index determination module determines the final risk index based on the facial alcohol status score, eye abnormality score, body movement amplitude score, voice instability score, and time risk score. The cabinet lock control module controls the opening and closing state of the smart cabinet lock based on the final risk index. This improves the accuracy of alcohol detection for cabinet openers and reduces the risk of cabinet openers using firearms. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0017] Figure 1 This is a schematic diagram of the structure of a gun and ammunition cabinet with alcohol detection function provided in one embodiment of this application; Detailed Implementation

[0018] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0019] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0020] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0021] Figure 1 This illustration shows a structural diagram of a gun cabinet with alcohol detection function provided in an embodiment of this application. The gun cabinet with alcohol detection function includes a camera, a microphone, an alcohol detection system, and a smart lock. The alcohol detection system includes a data acquisition module 21, a personnel status determination module 22, a risk index determination module 23, and a lock control module 24. Figure 1 As shown, the alcohol detection system may include: The data acquisition module 21 acquires the detection video of the person opening the cabinet through the camera, the detection voice of the person opening the cabinet through the microphone, historical cabinet opening data, and current time data. When the person opening the cabinet stands in front of the gun and ammunition cabinet, the data acquisition module 21 acquires the detection video of the person opening the cabinet through the camera, the detection voice of the person opening the cabinet through the microphone, historical cabinet opening data, and current time data.

[0022] In one implementation, the gun cabinet with alcohol detection function also includes an infrared sensor and a microphone, and the alcohol detection system also includes an infrared sensing module 25 and a voice command output module 26; wherein, the alcohol detection system is connected to the camera, speaker, infrared sensor, microphone and smart cabinet lock respectively.

[0023] Before the alcohol detection system acquires the detection video of the person opening the cabinet through the camera, the detection voice of the person opening the cabinet through the microphone, the historical opening data, and the current time data, the infrared sensing module 25 of the alcohol detection system 21 acquires the infrared sensing signal through the infrared sensor and detects whether the infrared sensing signal is greater than the preset infrared threshold. If the infrared sensing signal is greater than the preset infrared threshold, the voice command output module 26 sends a voice prompt command to the speaker so that the speaker can broadcast the voice prompt command so that the person opening the cabinet can perform the corresponding action according to the voice prompt command. If the infrared sensing signal is less than or equal to the preset infrared threshold, the voice command output module 26 will not send a voice prompt command to the speaker.

[0024] The personnel status determination module 22 processes the detection video, the detection voice, the historical cabinet opening data, and the current time data to obtain a facial alcohol status score, an abnormal eye status score, a body movement amplitude score, a voice instability score, and a time risk score. As one implementation method, the personnel status determination module 22 includes the following steps: A1. Process the detected video to obtain facial alcohol status score, eye abnormality score, and body movement amplitude score; After acquiring the detection video, the personnel status determination module 22 processes the detection video to obtain facial alcohol status score, eye abnormality score, and body movement amplitude score.

[0025] As one implementation method, the step of processing the detected video to obtain a facial alcohol status score, an eye condition abnormality score, and a body movement amplitude score includes: B1. The detected video is input into the facial alcohol status recognition model for processing to obtain a facial alcohol status score. The facial alcohol status recognition model is a trained and improved ResNet-18 artificial network model. The basic framework of the facial alcohol status recognition model is 3D ResNet-18 for extracting spatiotemporal features. Lightweight improvements to the facial alcohol status recognition model include introducing Ghost Net to replace some convolutions, reducing redundant computation. Specifically, Ghost Net replaces some convolutions by replacing standard convolutions with Ghost convolutions in all Bottleneck expansion and projection convolutions. Attention enhancement of the facial alcohol status recognition model: A Coordinate Attention (CA) module is embedded to improve sensitivity to key facial regions (such as cheeks and the area around the eyes). The Coordinate Attention (CA) module is inserted after the residual blocks in Stage 3 and Stage 4 and before downsampling; the CA attention mechanism can learn to adjust the weights of the input feature map channels. Temporal modeling of the facial alcohol status recognition model: A Bi-LSTM is connected after the 3D-CNN to capture dynamic patterns of micro-expressions and blood flow changes. The facial alcohol status recognition model achieves high-precision, low-latency facial alcohol status determination through Ghost convolution, Coordinate Attention, and 3D-CNN + Bi-LSTM. Ghost convolution is lightweight and suitable for edge deployment; Coordinate Attention accurately focuses on alcohol-sensitive facial areas. The 3D-CNN and Bi-LSTM networks jointly model spatial texture and temporal dynamics (such as facial blood flow and micro-expression tremors).

[0026] B2. Based on the detected video, obtain an abnormal eye condition score. ; As one implementation method, the step of obtaining an eye condition abnormality score based on the detected video includes: C1, Based on the detected video, extract key eye point data for each frame; MediaPipe Eyes was used to extract key eye points (including upper and lower eyelids, pupil center, etc.) from each frame of the detected video.

[0027] C2 calculates blink frequency, eyelid opening ratio, and fixation stability based on key eye point data; Then, based on the data of key eye points, the blink frequency, eyelid opening ratio, and fixation stability are calculated. Among them, the blink frequency is the number of blinks per unit time (normal range 15-20 times / minute); the eyelid opening ratio (EAR) is the standard deviation of the EAR of consecutive frames; and the fixation stability is the variance of the displacement of the pupil center in the screen coordinate system.

[0028] C3. The blink frequency, eyelid opening ratio, and fixation stability are normalized and weighted and fused to obtain an ocular condition abnormality score. .

[0029] As one implementation method, the step of normalizing and weighting the blink frequency, the eyelid opening ratio, and the fixation stability to obtain an ocular state abnormality score includes: ; in, Score abnormal eye condition; The normalized blink frequency; This represents the normalized eyelid opening ratio; This represents the gaze stability after normalization.

[0030] Where σ(x) is the number of blinks, eyelid opening ratio and fixation stability normalized to [0,1] according to min-max.

[0031] B3, use a human pose estimation model to identify joint points in the detected video to obtain human joint point data; That is, the personnel status determination module 22 of the alcohol detection system performs key point recognition on the detection video through a human posture estimation model to obtain human key point data; wherein, the human posture estimation model is the MediaPipe Pose model, and the human key point data includes data of 33 human key points in each frame.

[0032] B4. Based on the aforementioned human joint data, a score for the range of motion of the body is obtained.

[0033] As one implementation, the body movement range score includes an arm swing range score. Body swing amplitude score and head sway amplitude score The step of obtaining a body movement range score based on the human joint data includes: D1. Perform trajectory smoothing and normalization processing on the human joint point data to obtain the left wrist trajectory data. Right wrist joint data Trajectory data of the center point of the torso And the trajectory data of the tip of the nose; Savitzky-Golay filters can be applied to the key point sequence to denoise it, converting the coordinates into a local coordinate system relative to the center of the torso, thus eliminating the influence of overall displacement. D2, based on the left wrist trajectory data and the right wrist joint data, obtain the arm swing amplitude score. ; Among them, the arm swing amplitude score The Sigmoid function is defined as follows: ; ;in, The left wrist is represented by its two-dimensional coordinate trajectory in the video time series. Standard deviation is used to measure the degree of positional fluctuation of the keypoint within a video time period; The two-dimensional coordinate trajectory of the right wrist in the video time series; These are the core scale parameters. Calibration can be performed based on the deployment scenario (such as camera height and focal length). Standard deviation is used to measure the degree of positional fluctuation of the keypoint within a video time period; The larger the value, the more violent and unstable the shaking of that part, and the more likely it is to be affected by alcohol.

[0034] D3, based on the trajectory data of the torso center point, obtain a score for the body swing amplitude. ; Among them, the body swing amplitude score The Sigmoid function is defined as follows: ; ;in, The coordinate trajectory of the center point of the torso is used to measure the overall sway of the body. Standard deviation is used to measure the degree of positional fluctuation of the keypoint within a video time period; These are the core scale parameters. Calibration can be performed based on the deployment scenario (such as camera height and focal length). Standard deviation is used to measure the degree of positional fluctuation of the keypoint within a video time period; The larger the value, the more violent and unstable the shaking of that part, and the more likely it is to be affected by alcohol. This is used to avoid deviations caused by using only the hips (which are easily affected by leg micro-movements) or only the shoulders (which are easily affected by arm swing coupling).

[0035] D4. Based on the nose tip trajectory data, obtain the head sway amplitude score. .

[0036] Among them, the head sway amplitude score The Sigmoid function is defined as follows: ; ;in, The coordinate trajectory of the key point at the tip of the nose represents the position of the head. Standard deviation is used to measure the degree of positional fluctuation of the keypoint within a video time period; These are the core scale parameters. Calibration can be performed based on the deployment scenario (such as camera height and focal length). Standard deviation is used to measure the degree of positional fluctuation of the keypoint within a video time period; The larger the value, the more violent and unstable the shaking of that part, and the more likely it is to be affected by alcohol. This is used to avoid deviations caused by using only the hips (which are easily affected by leg micro-movements) or only the shoulders (which are easily affected by arm swing coupling).

[0037] A2, process the detected speech to obtain a speech instability score. ; After acquiring the detected speech, the alcohol detection system transcribes the speech using ASR (Acoustic Resonance Speech) and extracts acoustic features such as fundamental frequency standard deviation, speech rate fluctuation, and word error rate. Then, by normalizing and fusing these acoustic features, a speech instability score is obtained. If there is no audio input, then set And it will be automatically ignored in subsequent weighting.

[0038] Among them, the fundamental frequency standard deviation is used to measure pitch jitter; word error rate (which indirectly reflects pronunciation clarity and is used to obtain a speech instability score); A3. The historical data on opening cabinets and the current time data are processed to obtain a time risk score.

[0039] After acquiring historical opening data and current time data, the alcohol detection system processes the historical opening data and current time data to obtain a time risk score.

[0040] As one implementation method, the step of processing the historical cabinet opening data and the current time data to obtain a time risk score includes: E1, Based on the historical cabinet opening data, determine the historical behavior score. ; If the person opening the cabinet has at least one record of opening the cabinet while intoxicated: No record: ; Identity unknown: .

[0041] E2, determine the current time score based on the current time data. ; When the current time is a high-risk period High-risk periods can be defined as (22:00-06:00 or 2 hours before and after statutory holidays).

[0042] When the current time is another time period ; E3, based on the historical behavior score and the current time score, obtain the time risk score. .

[0043] Among them, time risk score ;in, For behavior coefficients; This is the time coefficient. It can be 0.6; It can be 0.4. Among them, the time risk score... By taking into account both individual risk propensity and environmental risk, the final result can be amplified or suppressed globally.

[0044] The risk index determination module 23 scores the alcohol status of the face. Eye condition abnormality score Body movement range score, speech instability score and time risk score Determine the final risk index; The risk index determination module 23 includes the following steps: F1, based on the facial alcohol status score (Fscore) and the eye condition abnormality score (Escore), yields a physiological aggregate score. ; Physiological aggregate score ;in, Score facial alcohol status; Score abnormal eye condition; The first state coefficient; This is the second state coefficient.

[0045] F2, based on scores for body movement range and speech instability. This yields a behavioral aggregate score; The body movement range score includes the arm swing range score. Body swing amplitude score and head sway amplitude score ; Behavioral aggregate scoring ; Rate the range of arm swing; Rate the amplitude of body sway; Score the amplitude of head movement; The coefficient is the first row; The coefficient is the second row; The coefficient is the third row; The fourth row is the coefficient; Aggregate total weight for behavior .

[0046] F3, based on the physiological aggregate score, the behavioral aggregate score, and the time risk score, yields the final risk index.

[0047] As one implementation, the step of obtaining the final risk index based on the physiological aggregate score, the behavioral aggregate score, and the time risk score includes: ; in; This is the final risk index; For sensitivity coefficient, sensitivity coefficient It can be equal to 8, making The output value rapidly increases from 0.1 to 0.9 within the range of [0.4, 0.6], enhancing the discrimination sensitivity; For physiological aggregation score; Aggregate scores for behaviors; The first polymerization coefficient; The second polymerization coefficient; The value represents a medium risk level. It can be 0.5; Score the time risk.

[0048] even though For medium risk (0.6), if M=1.0 (high-risk individuals + late at night), D can reach 0.6; if M=0.3 (low-risk scenario), then D=0.18, which can avoid false alarms.

[0049] This application organically combines deep learning model innovation, multimodal behavior analysis, spatiotemporal context awareness, and interpretable fusion logic to form a high-precision, low-false-alarm, and deployable visual alcohol screening system.

[0050] The cabinet lock control module 24 controls the opening and closing status of the smart cabinet lock according to the final risk index. The infrared sensor and speaker are electrically connected to the alcohol detection system.

[0051] A smart lock can be installed inside the gun cabinet. The cabinet lock control module 24 will not open the cabinet if it detects that the final risk index of the person opening the cabinet is greater than the preset risk threshold. The cabinet lock control module 24 opens the cabinet when it detects that the final risk index of the person opening the cabinet is less than or equal to the preset risk threshold.

[0052] In summary, this application utilizes a data acquisition module to acquire detection video of the person opening the cabinet via a camera, detection voice via a microphone, historical cabinet opening data, and current time data. The personnel status determination module processes the detection video, detection voice, historical cabinet opening data, and current time data to obtain facial alcohol status scores, eye abnormality scores, body movement amplitude scores, voice instability scores, and time risk scores. The risk index determination module determines a final risk index based on these scores. The cabinet lock control module controls the opening and closing status of the smart cabinet lock based on the final risk index. This improves the accuracy of alcohol detection for cabinet openers and reduces the risk of cabinet openers using firearms.

[0053] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0054] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] In the embodiments provided in this application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0056] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

Claims

1. A gun and ammunition cabinet with alcohol detection function, characterized in that, The gun cabinet with alcohol detection function includes a camera, a microphone, an alcohol detection system, and a smart cabinet lock. The alcohol detection system includes a voice command output module, a data acquisition module, a personnel status determination module, a risk index determination module, and a cabinet lock control module. The alcohol detection system includes: The data acquisition module acquires detection video of the person opening the cabinet through a camera, detection voice of the person opening the cabinet through a microphone, historical cabinet opening data, and current time data. The personnel status determination module processes the detection video, the detection voice, the historical cabinet opening data, and the current time data to obtain facial alcohol status score, eye abnormality score, body movement amplitude score, voice instability score, and time risk score. The risk index determination module determines the final risk index based on facial alcohol status score, abnormal eye condition score, body movement amplitude score, voice instability score, and time risk score. The cabinet lock control module controls the opening and closing status of the smart cabinet lock based on the final risk index.

2. The gun and ammunition cabinet with alcohol detection function according to claim 1, characterized in that, The personnel status determination module includes the following steps: The detection video is processed to obtain facial alcohol status score, eye abnormality score, and body movement amplitude score; The detected speech is processed to obtain a speech instability score; The historical data on opening cabinets and the current time data are processed to obtain a time risk score.

3. The gun and ammunition cabinet with alcohol detection function according to claim 2, characterized in that, The steps of processing the detected video to obtain facial alcohol status scores, eye condition abnormality scores, and body movement amplitude scores include: The detected video is input into the face alcohol status recognition model for processing to obtain a face alcohol status score. The face alcohol status recognition model is an improved ResNet-18 artificial network model that has been trained. Based on the detected video, an abnormal eye condition score is obtained; The human pose estimation model is used to identify joint points in the detected video to obtain human joint point data. Based on the data of the human joint points, a score for the range of motion of the body is obtained.

4. The gun and ammunition cabinet with alcohol detection function according to claim 3, characterized in that, The step of obtaining an eye condition abnormality score based on the detected video includes: Based on the detected video, extract key eye point data for each frame; Blink frequency, eyelid opening ratio, and fixation stability are calculated based on key eye point data. The blink frequency, the eyelid opening ratio, and the gaze stability are normalized and weighted to obtain an abnormal eye condition score.

5. The gun and ammunition cabinet with alcohol detection function according to claim 4, characterized in that, The step of normalizing and weighting the blink frequency, the eyelid opening ratio, and the fixation stability to obtain an ocular condition abnormality score includes: ； in, Score abnormal eye condition; The normalized blink frequency; This represents the normalized eyelid opening ratio; This represents the gaze stability after normalization.

6. The gun and ammunition cabinet with alcohol detection function according to claim 3, characterized in that, The body movement amplitude score includes an arm swing amplitude score, a body swing amplitude score, and a head swing amplitude score. The step of obtaining the body movement amplitude score based on the human joint point data includes: The human joint data is smoothed and normalized to obtain left wrist trajectory data, right wrist joint data, trunk center point trajectory data, and nose tip trajectory data. Based on the left wrist trajectory data and the right wrist joint data, an arm swing amplitude score is obtained; Based on the trajectory data of the torso center point, a score for the body swing amplitude is obtained; Based on the trajectory data of the nose tip, a score for the head sway amplitude is obtained.

7. The gun and ammunition cabinet with alcohol detection function according to claim 2, characterized in that, The step of processing the historical opening data and the current time data to obtain the time risk score includes: Based on the historical cabinet opening data, a historical behavior score is determined; Determine the current time score based on the current time data; A time risk score is obtained based on the historical behavior score and the current time score.

8. The gun and ammunition cabinet with alcohol detection function according to claim 1, characterized in that, The risk index determination module includes the following steps: A physiological aggregate score is obtained based on the facial alcohol status score (Fscore) and the eye condition abnormality score (Escore). The behavior aggregate score is obtained based on the body movement amplitude score Ascore, Tscore, Hscore and speech instability score Vscore; The final risk index is obtained based on the physiological aggregate score, the behavioral aggregate score, and the time risk score.

9. The gun and ammunition cabinet with alcohol detection function according to claim 8, characterized in that, The step of obtaining the final risk index based on the physiological aggregate score, the behavioral aggregate score, and the time risk score includes: ； in; This is the final risk index; Sensitivity coefficient; For physiological aggregation score; Aggregate scores for behaviors; The first polymerization coefficient; The second polymerization coefficient; This is a value indicating a medium risk level. Score the time risk.

10. The gun and ammunition cabinet with alcohol detection function according to claim 1, characterized in that, The gun cabinet with alcohol detection function also includes an infrared sensor and a speaker, and the alcohol detection system also includes an infrared sensing module and a voice command output module. The infrared sensing module acquires infrared sensing signals through an infrared sensor and detects whether the infrared sensing signals are greater than a preset infrared threshold. If the infrared sensing signal is greater than the preset infrared threshold, the voice command output module sends a voice prompt command to the speaker so that the speaker can broadcast the voice prompt command so that the person opening the cabinet can perform the corresponding action according to the voice prompt command. If the infrared sensing signal is less than or equal to the preset infrared threshold, the voice command output module will not send a voice prompt command to the speaker.

Images