Building construction access control method and system based on voice recognition

Abstract

The invention discloses a building construction access control method and system based on voice recognition, and relates to the technical field of voice recognition. The method comprises the steps that under the condition that a target user enters an identity verification area, a verification sequence structure is obtained, and the verification sequence structure is used for representing a priority sequence among multiple identity verification modes; under the condition that the first syn-position of the verification sequence structure is a voice recognition mode, obtaining a first to-be-verified audio input by the target user on site, a pre-stored reference audio of the target user and an environment audio of the identity verification area; based on the environment audio, performing denoising processing on the first to-be-verified audio to obtain a second to-be-verified audio; and comparing the second to-be-verified audio with the reference audio to obtain a voice verification result of the voice recognition mode, so that access of the target user is controlled based on the voice verification result. According to the invention, the efficiency and accuracy of access control can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of access control, and in particular to a construction access control method and system based on speech recognition. Background Art

[0002] Construction sites are characterized by high turnover, diverse roles, and complex construction environments, with high-risk operations such as working at height and operating large machinery. Unauthorized access not only disrupts the construction process but can also easily lead to safety incidents, endangering both personnel and project property. Therefore, establishing a reliable and efficient access control system to strictly manage personnel entry and exit is a critical component of construction management.

[0003] Currently, facial recognition and access card recognition are commonly used in construction to control access. Facial recognition verifies identity based on facial features, with the camera capturing an image and comparing it with a pre-stored facial template to determine legitimacy. Access card recognition uses a card as a carrier, which stores personnel information. When a person holds the card to a reader, the information read is matched with the system database to determine access rights. Some construction sites support both recognition methods to enhance access control convenience.

[0004] However, construction sites require workers to wear protective equipment such as hard hats and masks, which can seriously interfere with facial recognition, obscuring key facial features and making it difficult for cameras to clearly capture information, leading to recognition errors or even failure to recognize. Access cards are also inconvenient to use. Workers holding tools with both hands interrupt their work, reducing efficiency. Cards are also easily lost, damaged, or forgotten. Furthermore, while supporting multiple recognition methods, there is a lack of a robust judgment mechanism, which affects the accuracy of access control. This results in poor efficiency and accuracy in access control. Summary of the Invention

[0005] The embodiments of the present invention provide a method and system for construction access control based on speech recognition, which can improve the efficiency and accuracy of access control.

[0006] A first aspect of an embodiment of the present invention provides a method for controlling access to a construction site based on speech recognition, comprising: When the target user enters the identity authentication area, a verification sequence structure is obtained, where the verification sequence structure is used to represent the priority order among multiple identity authentication methods; When the first priority of the verification sequence structure is the voice recognition method, obtaining the first audio to be verified input by the target user on site, the pre-stored reference audio of the target user, and the ambient audio of the identity verification area; Based on the ambient audio, denoising is performed on the first audio to be verified to obtain a second audio to be verified; The second audio to be verified is compared with the reference audio to obtain a voice verification result in a voice recognition manner, so as to control the entry and exit of the target user based on the voice verification result.

[0007] Furthermore, the present invention also proposes that, when the target user enters the identity authentication area, before obtaining the verification sequence structure, the method further includes: Obtain historical verification information and attendance record information for various identity authentication methods within a historical time period; Based on each historical information to be verified and the corresponding attendance record information, respectively determine the verification validity of each identity authentication method in each attendance record information; Based on the verification validity of each identity authentication method in each attendance record information and the verification order of each identity authentication method in each attendance record information, respectively determine the verification priority of each identity authentication method in each verification order; Based on the verification priority of each identity authentication method in each verification order, a verification sequence structure is constructed.

[0008] Furthermore, the present invention also proposes to determine the verification validity of each identity authentication method in each attendance record information based on each historical information to be verified and the corresponding attendance record information, including: Determining the verification possibility index of each historical information to be verified based on the difference between each historical information to be verified and the corresponding reference verification information; Determining the verification reliability of each identity authentication method based on the verification possibility indicators included in each identity authentication method; Based on each attendance record information and the verification reliability of each identity verification method, the verification validity of each identity verification method in each attendance record information is determined respectively.

[0009] Furthermore, the present invention also proposes that the historical information to be verified is the historical audio to be verified, and the reference verification information is the reference audio; Based on the differences between each piece of historical information to be verified and the corresponding reference verification information, the verification possibility index of each piece of historical information to be verified is determined, including: An audio segment in the historical audio to be verified that is consistent with the reference audio waveform is determined as a valid audio segment; Perform short-time Fourier transform on the effective audio segment to obtain the instantaneous phase difference and amplitude spectrum of the effective audio segment in each time frame; Connect the harmonic peak points in adjacent amplitude spectra in sequence to form the spectrum envelope; The maximum envelope slope difference is selected from the envelope slope differences, where the envelope slope difference is the slope difference between adjacent spectrum envelopes; The variance of the instantaneous phase difference, the maximum envelope slope difference and the mean of the spectrum centroid are used to determine the verification possibility index of the historical audio to be verified. The mean of the spectrum centroid is the mean of the spectrum centroids corresponding to each time frame.

[0010] Furthermore, the present invention also proposes to determine the verification reliability of each identity authentication method based on each verification possibility index included in each identity authentication method, including: Obtaining the accuracy of first historical information to be verified, where the first historical information to be verified is each piece of historical information to be verified corresponding to a target identity authentication method, where the target identity authentication method is any identity authentication method, and the information accuracy is the degree of match between the first historical information to be verified and the reference authentication information; Performing mean processing on the verification possibility index and information accuracy of each first historical information to be verified, respectively, to obtain the mean of the verification possibility index and the mean of the information accuracy; Determining the local verification reliability of each first historical information to be verified by using the verification possibility index, information accuracy, verification possibility index average, and information accuracy average of each first historical information to be verified; Based on the local verification reliability of each piece of first historical information to be verified, the target verification reliability of the target identity authentication method is determined.

[0011] Furthermore, the present invention also proposes that, based on each attendance record information and the verification reliability of each identity authentication method, the verification validity of each identity authentication method in each attendance record information is determined separately, including: Obtaining a first attendance record vector of the first attendance record information and a first key feature vector in the second historical information to be verified, where the second historical information to be verified is historical information to be verified of a target identity authentication method corresponding to the first attendance record information, and the target identity authentication method is any identity authentication method; Obtaining a second attendance record vector of the second attendance record information and a second key feature vector in the third historical information to be verified, where the second attendance record information is the attendance record information previous to the first attendance record information, and the third historical information to be verified is the historical information to be verified of the target identity authentication method corresponding to the second attendance record information; The verification validity of the target identity authentication method in the first attendance record information is determined by using the first attendance record vector, the first key feature vector, the second attendance record vector, the second key feature vector and the verification reliability of the target identity authentication method.

[0012] Furthermore, the present invention also proposes to determine the verification priority of each identity authentication method in each verification sequence based on the verification validity of each identity authentication method in each attendance record information and the verification sequence of each identity authentication method in each attendance record information, including: Filtering out third attendance record information from each attendance record information, where the third attendance record information is used to represent the attendance record information of the target identity authentication method in the target authentication sequence; The verification priority of the target identity authentication method in the target verification sequence is determined by using the verification validity of the target identity authentication method in each third attendance record information and the number of the third attendance record information.

[0013] Furthermore, the present invention also proposes comparing the second audio to be verified with the reference audio to obtain a voice verification result in a voice recognition manner, so as to control the entry and exit of the target user based on the voice verification result, including: Comparing the second audio to be verified with the reference audio to obtain a voice verification result in a voice recognition manner; If the voice verification result indicates that the identity verification is passed, a door opening instruction signal is sent to the controller, so that the controller controls the opening of the gate based on the door opening instruction signal to allow the target user to enter and exit; In the case where the voice verification result indicates that the identity authentication fails, auxiliary identity authentication is performed based on the verification sequence structure, so that the entry and exit of the target user is controlled based on the auxiliary identity authentication result.

[0014] Furthermore, the present invention also proposes to perform auxiliary identity verification based on the verification sequence structure, so as to control the entry and exit of the target user based on the auxiliary identity verification result, including: Perform auxiliary identity authentication based on the identity authentication method in the next order of the authentication sequence structure to obtain an auxiliary identity authentication result; If the auxiliary identity authentication result indicates that the identity authentication fails, the loop returns to execute the auxiliary identity authentication based on the next identity authentication method in the verification sequence structure to obtain the auxiliary identity authentication result until a preset stop condition is reached. The preset stop condition is that the auxiliary identity authentication result indicates that the identity authentication passes or each identity authentication method in the verification sequence structure is traversed; When the auxiliary identity authentication result indicates that the identity authentication is passed, a manual verification process is triggered, and the entry and exit of the target user is controlled based on the manual verification result.

[0015] A second aspect of an embodiment of the present invention provides a construction entry and exit control system based on speech recognition, comprising: A structure acquisition module is used to acquire a verification sequence structure when a target user enters the identity authentication area. The verification sequence structure is used to represent the priority order among multiple identity authentication methods. An audio acquisition module is used to acquire a first audio to be verified input by a target user on site, a pre-stored reference audio of the target user, and an ambient audio of the identity verification area when the first priority of the verification sequence structure is a voice recognition method; An audio processing module, configured to perform denoising on the first audio to be verified based on the ambient audio to obtain a second audio to be verified; The result determination module is used to compare the second audio to be verified with the reference audio to obtain a voice verification result in a voice recognition manner, so as to control the entry and exit of the target user based on the voice verification result.

[0016] The present invention has the following beneficial effects: In the method for controlling access to and from construction sites based on voice recognition provided by an embodiment of the present invention, by setting a verification sequence structure, when voice recognition is the first priority, the first audio to be verified input by the target user on site, the pre-stored reference audio, and the ambient audio are obtained, and the first audio to be verified is denoised based on the ambient audio to obtain the second audio to be verified, which is then compared with the reference audio to obtain a voice verification result. Voice recognition is not affected by the obstruction of protective equipment, and workers do not need to put down their tools to perform additional operations, thus avoiding the situation where work is interrupted by taking a card, and improving access efficiency. At the same time, the denoising process can effectively reduce environmental noise interference, making the second audio to be verified clearer, and the voice verification result obtained by comparison with the reference audio is more accurate. In combination with the verification sequence structure, multiple recognition methods are used scientifically and rationally, which makes up for the shortcomings of traditional methods in the use of multiple recognition methods and comprehensively improves the accuracy of personnel access control. Therefore, the present invention can effectively solve the problems of traditional recognition methods in construction scenarios and improve the efficiency and accuracy of access control. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flowchart of a first method for controlling access to and from a construction site based on speech recognition according to an embodiment of the present invention is provided; Figure 2 A flow chart of a second method for controlling access to and from a construction site based on speech recognition provided by one embodiment of the present invention; Figure 3 A schematic diagram of the process of S202 provided in one embodiment of the present invention; Figure 4 A schematic diagram of the process of S104 provided in one embodiment of the present invention; Figure 5A schematic structural diagram of a construction access control system based on speech recognition provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0019] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of a construction access control method and system based on voice recognition proposed by the present invention. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0020] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0021] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of the present invention comply with the relevant provisions of laws and regulations.

[0022] It should be noted that in the embodiments of the present invention, certain software, components, models and other existing solutions in the industry may be mentioned. They should be regarded as exemplary and their purpose is only to illustrate the feasibility of implementing the technical solution of the present invention, but it does not mean that the applicant has or will necessarily use the solution.

[0023] In traditional construction access control systems, identity verification methods that rely on a single biometric or physical carrier have significant limitations. Facial recognition technology is easily affected by occlusion from safety equipment, resulting in reduced feature extraction accuracy. Infrared fill light modules struggle to accurately reconstruct three-dimensional facial models under strong ambient light interference. Access card verification relies on physical contact interaction, which can easily cause radio frequency signal collisions in multi-person scenarios and cannot verify the biometric correlation between the cardholder and their identity information in real time. Although multimodal verification systems use a parallel verification mechanism, they lack a dynamic priority scheduling algorithm. Conflicts in the timing of data collection between different verification methods can trigger competition for system resources, resulting in a nonlinear increase in the overall verification response time.

[0024] When faced with the above problems, the present invention first considers how to break through the traditional verification method's dependence on physical carriers and facial features. The noise interference in the complex acoustic environment of the construction site directly affects the accuracy of speech recognition, and a targeted noise reduction mechanism needs to be designed for this. The dynamic scheduling problem of the multimodal verification system requires the establishment of a scientific verification sequence decision model to avoid resource competition causing system congestion. The inefficiency of traditional verification methods in concurrent scenarios suggests the need to introduce a priority strategy, and the contactless nature of voice interaction can alleviate the operational burden of people holding objects. The present invention further analyzed and found that different time periods and environmental factors have a dynamic impact on the effectiveness of the verification method, and it is necessary to construct a verification sequence structure that can be adaptively adjusted, where the generation of the verification sequence needs to be combined with the reliability characteristics in the historical verification data.

[0025] In this regard, Figure 1 As shown, the present invention proposes a construction access control method based on voice recognition, which can be applied to a server, including the following steps S101 to S104: S101, when a target user enters an identity authentication area, obtaining an authentication sequence structure, where the authentication sequence structure is used to represent a priority order among multiple identity authentication methods; S102, when the first priority of the verification sequence structure is a voice recognition method, obtaining a first audio to be verified input by the target user on site, a pre-stored reference audio of the target user, and an ambient audio of the identity verification area; S103, performing denoising processing on the first audio to be verified based on the ambient audio to obtain a second audio to be verified; S104: Compare the second audio to be verified with the reference audio to obtain a voice verification result in a voice recognition manner, so as to control the entry and exit of the target user based on the voice verification result.

[0026] In this embodiment, the verification sequence structure refers to the logical arrangement structure of the priority order among multiple identity authentication methods, which can be implemented by dynamic adjustment strategies or preset rules. Its function is to dynamically optimize the execution order of the identity authentication process according to the requirements of different scenarios and improve verification efficiency.

[0027] The authentication area refers to the pre-demarcated spatial range used to perform authentication operations. It can be implemented using infrared sensing or camera positioning technology. Its function is to ensure that the triggering timing of the authentication process matches the user's actual location through physical or virtual boundary restrictions.

[0028] Ambient audio refers to the background sound signal present in the authentication area, which can be collected using a directional microphone array. Its function is to provide noise samples for subsequent audio denoising and eliminate the negative impact of environmental interference on speech recognition.

[0029] Denoising refers to the noise reduction and enhancement of the target speech signal based on the characteristics of the ambient noise. It can be implemented using spectral subtraction or adaptive filtering algorithms. Its function is to improve the accuracy of subsequent voiceprint comparison by eliminating the spectral overlap between the ambient noise and the speech signal.

[0030] Reference audio refers to the pre-registered user voiceprint feature data, which can be implemented using Mel-frequency cepstral coefficient modeling. Its function is to provide a comparison benchmark for the real-time collected voice to ensure the uniqueness of the biological characteristics for identity verification.

[0031] The core invention of this invention lies in building an adaptive voice recognition framework in complex construction scenarios by combining dynamic priority verification order with multimodal identity authentication technology, thereby improving the accuracy and execution efficiency of personnel access control while ensuring safety.

[0032] As an example, consider setting up an identity verification area at the entrance of a construction site, equipped with a high-sensitivity microphone array and ambient noise collection equipment. When a target user enters the authentication area, the server first retrieves a verification sequence structure. This structure represents the priority order of multiple authentication methods and is used to dynamically adjust the authentication strategy.

[0033] When voice recognition is the first priority in the verification sequence, the server acquires three types of audio data: the target user's on-site audio input to be verified, pre-stored reference audio of the target user, and ambient audio from the authentication area. Specifically, the target user speaks a preset passphrase into the microphone, which the server records as the first audio to be verified.

[0034] Next, the server uses the ambient audio to denoise the first audio to be verified, generating a second audio to be verified. This step improves speech recognition accuracy by eliminating ambient noise. Specifically, the server can use an adaptive filtering algorithm to denoise the first audio to be verified based on the collected ambient noise, generating a second audio to be verified.

[0035] The processed second audio to be verified is then compared with the reference audio to obtain a voice verification result. Based on this result, the target user's access rights are controlled. Specifically, a dynamic time warping algorithm can be used to compare the second audio to be verified with the reference audio, calculate a similarity score, and then obtain a voice verification result based on the similarity score. Then, if the voice verification result indicates that the identity verification is successful, an opening signal is sent to the gate to allow the target user to enter. If the voice verification result indicates that the identity verification is unsuccessful, the next verification method, such as facial recognition or access card verification, will be switched according to the verification sequence structure.

[0036] The introduction of a sequential verification structure enables the server to flexibly select the most appropriate verification method based on different scenarios. The collection and utilization of ambient audio effectively addresses the interference of complex acoustic environments on construction sites with speech recognition. De-noising further improves the quality of speech samples. By comparing the processed second audio to be verified with the reference audio, the service reliably determines the user's identity, enabling precise access control.

[0037] Through this embodiment, a verification sequence structure is first set up. When the first priority is voice recognition, the first audio to be verified input by the target user on site, the pre-stored reference audio and the environmental audio are obtained, and the first audio to be verified is denoised based on the environmental audio to obtain the second audio to be verified, and then compared with the reference audio to obtain the voice verification result. Voice recognition is not affected by the obstruction of protective equipment, and workers do not need to put down their tools to perform additional operations, avoiding the situation of interrupting work by taking a card, and improving entry and exit efficiency. At the same time, the denoising process can effectively reduce the interference of environmental noise, making the second audio to be verified clearer, and the voice verification result obtained by comparison with the reference audio is more accurate. In addition, combined with the verification sequence structure, a variety of recognition methods are used scientifically and rationally, which makes up for the shortcomings of traditional methods in the use of multiple recognition methods and comprehensively improves the accuracy of personnel entry and exit control. Therefore, the present invention can effectively solve the problems of traditional recognition methods in construction scenarios and can improve the efficiency and accuracy of entry and exit control.

[0038] In some of the aforementioned solutions, the verification order structure uses fixed rules to prioritize identity verification methods, which lacks adaptability to dynamic changes in real-world scenarios. This fixed ordering fails to reflect the effectiveness of verification methods over time, preventing dynamic adjustment of the verification order based on actual verification results, impacting the accuracy and efficiency of access control.

[0039] In this regard, Figure 2 As shown, the present invention further proposes that before S101, the construction access control method based on speech recognition may also include the following S201 to S204: S201, obtaining historical verification information and attendance record information of various identity authentication methods within a historical time period; S202, based on each historical information to be verified and the corresponding attendance record information, respectively determining the verification validity of each identity authentication method in each attendance record information; S203, based on the verification validity of each identity authentication method in each attendance record information and the verification order of each identity authentication method in each attendance record information, respectively determine the verification priority of each identity authentication method in each verification order; S204: Construct a verification sequence structure based on the verification priority of each identity authentication method at each verification order.

[0040] In this embodiment, the historical information to be verified includes actual verification data of identity authentication methods within different time periods, and the attendance record information reflects the actual entry and exit records corresponding to the verification operation. The verification validity is calculated by the correlation between the historical information to be verified and the attendance record, and is used to evaluate the reliability of the verification method within a specific time period. The verification priority refers to the priority position of the verification method in the historical verification process. The verification preference is comprehensively calculated by combining the verification validity and the historical verification priority. The verification sequence structure generates a dynamic ranking by integrating the preference of each verification method at different priorities.

[0041] Specifically, the historical time period is divided into multiple consecutive intervals, and verification request data and actual attendance results corresponding to each identity verification method are collected within each interval. For example, within a certain period, 100 facial recognition verification requests were collected, of which 85 were successful and valid attendance records were recorded. Similarly, 80 access card verification requests were collected, of which 76 were successful and valid attendance records were recorded. Verification effectiveness is determined by calculating the ratio of the number of successful verifications to the degree of matching of attendance records. The average of the verification effectiveness of a particular verification method across multiple attendance records is used as the overall effectiveness indicator for that method.

[0042] The verification priority is determined by the statistical analysis of the startup sequence of each method during the historical verification process. For example, voice recognition verification is started 50 times in a certain period of time, and the subsequent verification steps are skipped after 40 successful verifications. The verification priority is obtained by weighted calculation of the verification effectiveness and the frequency of use of the priority. For example, the historical usage frequency of voice recognition in the second priority is 30%, and its verification effectiveness is 0.92, then the verification priority is 0.92×0.3=0.276. The verification priority of each verification method at different priorities is compared horizontally, and the verification method with the highest verification priority is assigned to the corresponding verification priority, forming a dynamically adjusted verification order structure.

[0043] As an example, first obtain the historical pending verification information and attendance record information for various authentication methods within a historical time period. For example, obtain the historical voice data, facial images, and fingerprint images for the three authentication methods of voice recognition, facial recognition, and fingerprint recognition, as well as the corresponding attendance clock-in records for the past month.

[0044] Based on each piece of historical information to be verified and the corresponding attendance record information, the verification validity of each authentication method in each attendance record is determined. Specifically, by comparing the degree of match between the historical information to be verified and the pre-stored reference information, combined with the accuracy of the attendance record, a verification validity score for each authentication method in each attendance record is calculated.

[0045] Based on the verification validity of each identity authentication method in each attendance record and its verification order in each attendance record, the verification preference of each identity authentication method at each verification order is determined. For example, the verification validity score distribution of each identity authentication method in different verification orders is calculated to calculate its verification preference at each verification order.

[0046] Finally, based on the verification priority of each authentication method in each verification order, a verification order structure is constructed. Furthermore, the authentication methods can be sorted according to the level of verification priority to generate a priority order list as the final verification order structure.

[0047] Through this embodiment, the present invention can dynamically optimize the identity verification sequence based on historical verification data, improving verification efficiency and accuracy. This allows for the selection of the most appropriate combination of verification methods to suit different work environments and personnel characteristics. Furthermore, this solution reduces unnecessary verification attempts, shortens personnel entry and exit times, and improves construction site management efficiency.

[0048] In some of the above-mentioned schemes of the present invention, the difference between the historical information to be verified and the reference verification information is not fully quantified, resulting in a lack of objective basis for determining the verification possibility index, affecting the accuracy of subsequent verification reliability calculations, and further affecting the determination of verification effectiveness.

[0049] In this regard, Figure 3 As shown, the present invention further proposes that S202 may specifically include the following S301 to S303: S301, determining a verification possibility index for each piece of historical information to be verified based on the difference between each piece of historical information to be verified and the corresponding reference verification information; S302, determining the verification reliability of each identity authentication method based on each verification possibility indicator included in each identity authentication method; S303: Based on each attendance record information and the verification reliability of each identity verification method, respectively determine the verification validity of each identity verification method in each attendance record information.

[0050] In this embodiment, the historical information to be verified is various information used to verify the employee's identity in past attendance scenarios. For example, it may be the employee's password, facial image data, fingerprint data, etc. This information is the original data received by the server for identity verification.

[0051] Attendance records are information related to employee attendance, typically including employee identification (such as employee number and name), attendance time (entry clock-in time, exit clock-out time, etc.). They are data records that reflect employees' actual attendance behavior.

[0052] Reference verification information is standard information pre-stored in the system and used to compare with historical verification information to determine the legitimacy of an employee's identity. For example, for password verification, the reference verification information is the employee's pre-set correct password; for facial recognition verification, the reference verification information is the pre-collected and stored employee facial feature data.

[0053] The Verification Likelihood Index (VCI) is a quantitative indicator used to measure the degree of match between historical information to be verified and reference verification information. It reflects the likelihood that the historical information to be verified will pass identity verification. For example, in face recognition, the VCI can be determined by calculating the similarity between the historical image to be verified and the reference verification image. The higher the similarity, the greater the VCI.

[0054] Verification reliability is a metric used to measure a specific authentication method's ability to accurately identify employees within the attendance system. It comprehensively considers the verification probability indicators of all historical pending verification information for that authentication method, reflecting the stability and accuracy of that authentication method in practical applications. For example, if the verification probability indicators for most historical pending verification information for a particular authentication method are high, then the authentication method has high verification reliability.

[0055] Verification effectiveness is a metric that measures the effectiveness of a particular authentication method within a specific attendance record scenario. It combines the characteristics of the attendance record itself with the verification reliability of the authentication method, reflecting the applicability and accuracy of the authentication method within that specific attendance record. For example, if the employee's historical verification likelihood indicator for a particular authentication method is generally high within a particular attendance record, and the authentication method also has high verification reliability, then the authentication method has high verification effectiveness within that attendance record.

[0056] As an example, based on the differences between the historical pending verification information and the corresponding reference verification information, a verification likelihood index is determined for each piece of historical pending verification information. For example, in password verification, the historical pending verification information (i.e., the employee's previously entered password) is compared character by character with the corresponding reference verification information (i.e., the pre-set correct password). If there is a complete match, the verification likelihood index is set to the highest value. If there are partial character mismatches, the verification likelihood index can be calculated based on the number and position of the mismatched characters. For example, for each character mismatch, a certain score is subtracted from the highest value, ultimately resulting in a specific value as the verification likelihood index. In fingerprint verification, fingerprint feature points are extracted from the historical pending verification fingerprint data (i.e., the data collected when the employee punched in) and the reference verification fingerprint data (i.e., the pre-collected and stored employee fingerprint feature data). The verification likelihood index is determined by comparing the degree of match between the two sets of fingerprint feature points. For example, the ratio of the number of matching feature points to the total number of feature points can be calculated and then multiplied by 100 to obtain the verification likelihood index.

[0057] Then, based on the various verification possibility indicators included in each identity authentication method, the verification reliability of each identity authentication method is determined separately. Specifically, the mean value, variance and other statistical quantities of these verification possibility indicators can be calculated. The mean value reflects the average level of the degree of matching between the historical information to be verified and the reference verification information under this identity authentication method, and the variance reflects the fluctuation of the matching degree. Based on the statistical quantities such as the mean value and variance, the verification reliability of the identity authentication method is determined in combination with preset rules or models. For example, a threshold value can be set. If the mean value is higher than the threshold value and the variance is small, the verification reliability of the identity authentication method is considered to be high, and a higher reliability score can be assigned; if the mean value is low or the variance is large, the verification reliability is considered to be low, and a lower score is assigned.

[0058] Finally, based on the attendance record information and the verification reliability of each identity verification method, the verification validity of each identity verification method in each attendance record information is determined. Specifically, based on the verification reliability of each identity verification method and the characteristics of the attendance record information, a preset algorithm or model is used to determine the verification validity of each identity verification method in the attendance record information. For example, if the attendance time in the attendance record information is during peak hours, and the verification accuracy of a certain identity verification method is generally low during peak hours (determined through historical data analysis), then even if the verification reliability of that identity verification method is high, the verification validity of that identity verification method in the attendance record information will be correspondingly reduced.

[0059] This embodiment analyzes the effectiveness of each identity verification method based on historical data, improving verification accuracy. By considering multiple factors, such as the differences between historical verification information and reference information, and the relevance of attendance records, the reliability of each verification method is comprehensively assessed. This data-driven verification effectiveness analysis method can adapt to different construction environments and personnel characteristics, dynamically optimize verification strategies, and improve the accuracy and efficiency of access control.

[0060] In some of the aforementioned solutions, the verification likelihood indicator determination method suffers from insufficient accuracy in complex construction environments. Due to environmental noise, the difference between the historical audio to be verified and the reference audio is difficult to accurately quantify, leading to bias in subsequent verification reliability assessments.

[0061] In this regard, the present invention further proposes that the historical information to be verified is the historical audio to be verified, and the reference verification information is the reference audio; S301 may specifically include: An audio segment in the historical audio to be verified that is consistent with the reference audio waveform is determined as a valid audio segment; Perform short-time Fourier transform on the effective audio segment to obtain the instantaneous phase difference and amplitude spectrum of the effective audio segment in each time frame; Connect the harmonic peak points in adjacent amplitude spectra in sequence to form the spectrum envelope; The maximum envelope slope difference is selected from the envelope slope differences, where the envelope slope difference is the slope difference between adjacent spectrum envelopes; The variance of the instantaneous phase difference, the maximum envelope slope difference and the mean of the spectrum centroid are used to determine the verification possibility index of the historical audio to be verified. The mean of the spectrum centroid is the mean of the spectrum centroids corresponding to each time frame.

[0062] In this embodiment, the valid audio segment is located by a waveform matching algorithm to ensure that subsequent analysis focuses on areas with similar characteristics to the reference audio. The short-time Fourier transform performs time-frequency decomposition of the audio signal with a fixed time window. The time window length is set to 20 milliseconds, and the overlap rate of adjacent windows is 50%. The spectrum envelope is connected to the harmonic peak points by linear interpolation, and the harmonic peak points are extracted from the amplitude spectrum by a peak detection algorithm. The maximum envelope slope difference is obtained by calculating the difference in the absolute values ​​of the slopes of adjacent envelope segments, and the largest difference is selected as the feature parameter. The spectrum centroid mean is obtained by calculating the arithmetic mean of the spectrum centroid of each time frame. The spectrum centroid is determined by the weighted sum of each frequency component and its amplitude.

[0063] Specifically, valid audio segments are extracted using a dynamic time warping algorithm. This algorithm aligns the time axis of the historical audio to be verified with the reference audio and marks continuous regions where waveform similarity exceeds a preset threshold. During the short-time Fourier transform (SFT) process, a Hamming window is used to reduce spectral leakage. The instantaneous phase difference of each time frame is obtained by comparing the phase information of adjacent frames. The construction of a spectral envelope further distinguishes noise interference from true speech characteristics. Environmental noise typically exhibits broadband spectral characteristics, while the harmonic structure of the speech signal exhibits regular changes along the envelope. The maximum envelope slope difference quantifies the degree of sudden changes in the envelope shape and effectively reflects the transient components in the speech signal. The variance of the instantaneous phase difference describes the phase stability of the speech signal, while the mean spectral centroid characterizes the distribution center of the signal energy in the frequency domain. These three parameters are weighted and fused to form a verification probability index. This verification probability index integrates time, frequency, and phase characteristics, improving the accuracy of speech verification in complex environments.

[0064] As an example, the verification possibility index of the historical audio to be verified can be specifically determined by the following formula 1: Formula 1 In formula 1, It is used to characterize the verification possibility index of the kth historical information to be verified for the i-th identity authentication method. Used to characterize the variance of the instantaneous phase difference in the k-th historical information to be verified, Used to represent the maximum envelope slope difference in the k-th historical information to be verified, Used to represent the mean of the spectrum centroid in the k-th historical information to be verified.

[0065] Among them, if the waveform of the valid audio segment of a historical audio to be verified is relatively consistent with that of the reference audio, and the instantaneous phase difference of the valid audio segment fluctuates greatly, the slope of the frequency domain envelope fluctuates less (that is, the energy distribution of the audio segment is more regular), and the spectrum centroid is smaller, then the valid audio segment is more likely to be used for correct identity authentication.

[0066] This embodiment enables analysis of audio features from multiple dimensions, extracting richer and more stable audio feature parameters. This improves the accuracy and robustness of voice verification and reduces interference from factors such as environmental noise. Furthermore, this solution can adapt to the voice verification needs of different speakers and environments, enhancing the effectiveness of voice recognition in construction access control.

[0067] Some of the aforementioned solutions of the present invention propose determining the verification validity of an identity authentication method based on historical pending verification information and attendance records, thereby constructing a verification sequence structure. However, this process evaluates verification reliability based solely on the mean of the verification likelihood index, which fails to reflect the individual differences in historical pending verification information, resulting in inaccurate verification reliability assessments of the target identity authentication method.

[0068] In this regard, the present invention further proposes that S302 may specifically include: Obtaining the accuracy of first historical information to be verified, where the first historical information to be verified is each piece of historical information to be verified corresponding to a target identity authentication method, where the target identity authentication method is any identity authentication method, and the information accuracy is the degree of match between the first historical information to be verified and the reference authentication information; Performing mean processing on the verification possibility index and information accuracy of each first historical information to be verified, respectively, to obtain the mean of the verification possibility index and the mean of the information accuracy; Determining the local verification reliability of each first historical information to be verified by using the verification possibility index, information accuracy, verification possibility index average, and information accuracy average of each first historical information to be verified; Based on the local verification reliability of each piece of first historical information to be verified, the target verification reliability of the target identity authentication method is determined.

[0069] In this embodiment, information accuracy is calculated based on the degree of match between the first historical information to be verified and the reference verification information, for example, quantified by voiceprint similarity or text matching. Local verification reliability is determined by the deviation between information accuracy and a verification probability indicator, for example, by multiplying the deviation by information accuracy and then normalizing the result. Target verification reliability is obtained by taking the maximum or average of the local verification reliabilities.

[0070] Specifically, to determine the verification reliability of a target identity authentication method, the database first extracts multiple historical verification information and corresponding reference verification information corresponding to that method. For each historical verification information, the degree of match with the reference verification information is calculated, generating an information accuracy value between 0 and 1. Simultaneously, the verification likelihood index of each historical verification information is averaged to obtain a baseline value. Subsequently, the verification likelihood index of each historical verification information is compared with this baseline value, the degree of difference is calculated, and the difference is multiplied by the information accuracy to obtain a local verification reliability, reflecting individual reliability. Finally, the maximum value of all local verification reliabilities is used as the target verification reliability to avoid interference from outliers. For example, if the information accuracy of a historical verification is 0.9, the verification likelihood index is 0.85, and the mean is 0.8, then the degree of difference is 0.05, and the local verification reliability is 0.9 × 0.05 = 0.045, which after normalization is 0.045 / 0.1 = 0.45. This method comprehensively evaluates individual differences and overall trends, improving the accuracy of the target verification reliability calculation.

[0071] As an example, the target verification reliability of the target identity authentication method can be specifically determined by the following formula 2: Formula 2 In formula 2, Used to characterize the verification reliability of the i-th authentication method, It is used to characterize the verification possibility index of the kth historical information to be verified for the i-th identity authentication method. It is used to characterize the accuracy of the k-th historical information to be verified for the i-th authentication method. The mean value of the verification possibility index used to characterize the i-th authentication method, Used to characterize the mean information accuracy of the i-th authentication method. It is used to represent the number of historical information to be verified in the i-th authentication method, and exp is used to represent the exponential function operation. Represents an infinitesimal quantity greater than 0, used to prevent the denominator from being 0.

[0072] Among them, when The smaller the value of is, the higher the reliability of the authentication method i is. The larger the value, the smaller the difference between the verification possibility index and the mean value of information accuracy, and the higher the verification reliability.

[0073] Through this embodiment, based on the individual differences between different historical information to be verified, the verification reliability of the identity authentication method is accurately evaluated, which can improve the evaluation accuracy of the verification reliability of the target identity authentication method, thereby improving the rationality of the constructed verification sequence structure.

[0074] In some of the aforementioned solutions, the verification effectiveness assessment is inaccurate. Because the verification effectiveness calculation is based solely on the degree of match between the current attendance record and the historical information to be verified, there is no consideration of the temporal correlation between attendance records. This makes it impossible to dynamically capture the actual effectiveness of the verification method over different time periods, which in turn affects the optimization accuracy of the verification sequence structure.

[0075] In this regard, the present invention further proposes that S303 may specifically include: Obtaining a first attendance record vector of the first attendance record information and a first key feature vector in the second historical information to be verified, where the second historical information to be verified is historical information to be verified of a target identity authentication method corresponding to the first attendance record information, and the target identity authentication method is any identity authentication method; Obtaining a second attendance record vector of the second attendance record information and a second key feature vector in the third historical information to be verified, where the second attendance record information is the attendance record information previous to the first attendance record information, and the third historical information to be verified is the historical information to be verified of the target identity authentication method corresponding to the second attendance record information; The verification validity of the target identity authentication method in the first attendance record information is determined by using the first attendance record vector, the first key feature vector, the second attendance record vector, the second key feature vector and the verification reliability of the target identity authentication method.

[0076] In this embodiment, the server retrieves two consecutive attendance records from the database, marks the current attendance record as the first attendance record, and the previous record as the second attendance record. The attendance record vector contains at least two dimensional elements: a timestamp and a verification result status, and is normalized to form a standardized vector.

[0077] The first key feature vector and the second key feature vector may include key feature information extracted during the identity verification process. For example, for voice recognition, keywords may be extracted from the historical audio to be verified as key feature vectors; for face recognition, feature points may be extracted from the face image as key feature vectors.

[0078] As an example, the verification validity of each identity authentication method in each attendance record information can be determined by the following formula 3: Formula 3 In formula 3, It is used to represent the verification validity of the i-th identity authentication method in the j-th attendance record information. Used to represent the attendance record vector corresponding to the j-th attendance record information, Used to represent the attendance record vector corresponding to the j-1th attendance record information, The key feature vector used to characterize the historical information to be verified for the i-th identity authentication method corresponding to the j-th attendance record information, The key feature vector used to characterize the historical information to be verified for the i-th authentication method corresponding to the j-1-th attendance record information, It is used to characterize the verification reliability of the i-th authentication method, norm is used to characterize the standardization processing, and || is used to characterize the modulus length of the vector. Represents an infinitesimal quantity greater than 0, used to prevent the denominator from being 0.

[0079] in, It reflects the degree of difference between adjacent attendance record information. If the adjacent attendance record vectors differ greatly, it means that the attendance status may have changed significantly. The modulus difference between these two key feature vectors measures the degree of variation between the key features of adjacent historical information to be verified. When the difference between adjacent attendance record vectors is small, while the difference between adjacent key feature vectors of historical information to be verified is large, or when the authentication method is less reliable, the numerator is small and the denominator is large, resulting in a smaller overall ratio and lower verification validity after standardization.

[0080] This embodiment comprehensively considers attendance record information, historical information to be verified, and the verification reliability of the identity verification method to accurately calculate the verification validity of the identity verification method in the attendance record. This improves the accuracy of the verification validity calculation, provides a reliable basis for the subsequent construction of the verification sequence structure, and further enhances the accuracy and reliability of the entire identity verification system.

[0081] In some of the above-mentioned schemes of the present invention, it is difficult to accurately quantify the verification priority of each identity authentication method in each verification sequence based on the verification validity of each identity authentication method in each attendance record information, resulting in poor accuracy of the verification priority.

[0082] In this regard, the present invention further proposes that S203 may specifically include: Filtering out third attendance record information from each attendance record information, where the third attendance record information is used to represent the attendance record information of the target identity authentication method in the target authentication sequence; The verification priority of the target identity authentication method in the target verification sequence is determined by using the verification validity of the target identity authentication method in each third attendance record information and the number of the third attendance record information.

[0083] In this embodiment, when filtering the third attendance records, the verification priority is correlated with the time dimension of the attendance records to extract the attendance records for the target identity authentication method at the target verification priority. When calculating the verification preference, a weighted calculation model can be constructed by combining the mean verification validity and the number of third attendance records.

[0084] Specifically, all attendance records with the target identity authentication method in the target verification priority are first selected from historical attendance records as the third attendance records. Next, the verification validity of the target identity authentication method in these third attendance records is averaged, and the number of third attendance records is used as a weighting factor to calculate the verification preference for this verification priority.

[0085] As an example, the authentication priority of each authentication method in each authentication order can be determined by the following formula 4: Formula 4 In formula 4, It is used to represent the verification preference of the i-th authentication method in the s-th verification order. It is used to represent the verification validity of the i-th identity authentication method in the j-th attendance record information. It is used to represent the number of third attendance records of the sth verification order for the i-th authentication method. s is used to represent the sth verification order. Used to characterize the maximum and minimum value normalization.

[0086] This embodiment enables dynamic optimization of the identity verification sequence based on historical attendance data, improving the efficiency and accuracy of the verification process. By analyzing the actual performance of each verification method at different levels, the system can adaptively adjust the verification strategy to better suit the actual application scenario. This data-driven approach avoids the potential bias caused by subjective judgment, making the identity verification process more scientific and reasonable. Furthermore, because the optimization of the verification sequence is based on a large amount of historical data, it can adapt to dynamic changes in the construction site environment and personnel composition, maintaining the long-term effectiveness of the verification system.

[0087] In some of the above-mentioned schemes of the present invention, the identity of the target user is authenticated through a verification sequence structure. When there is an error in the voice recognition method in the first priority, there may be a risk of misjudgment if the entry and exit of the target user is controlled only by relying on the voice verification result, resulting in authorized personnel not entering the construction area and delaying the construction process.

[0088] In this regard, Figure 4 As shown, the present invention further proposes that S104 may specifically include the following S401 to S403: S401, comparing the second audio to be verified with the reference audio to obtain a voice verification result in a voice recognition manner; S402, when the voice verification result indicates that the identity verification is passed, sending a door opening instruction signal to the controller, so that the controller controls the opening of the gate based on the door opening instruction signal to allow the target user to enter and exit; S403: When the voice verification result indicates that the identity verification fails, an auxiliary identity verification is performed based on the verification sequence structure, so as to control the entry and exit of the target user based on the auxiliary identity verification result.

[0089] In this embodiment, the secondary identity verification process is executed based on the next authentication method in the verification sequence, such as fingerprint or iris recognition. If the secondary identity verification result still fails, the next authentication method in the sequence is executed in a loop until a preset stop condition is met. The preset stop condition includes passing the verification or traversing all authentication methods in the verification sequence. If the secondary identity verification passes, manual verification is triggered, such as on-site confirmation of the target user's identity by security personnel. The manual verification result serves as the basis for final access control.

[0090] Specifically, after a target user enters the identity verification area, voice recognition is first performed. If voice verification fails, the next-highest priority authentication method in the verification sequence is automatically invoked as a secondary verification method. For example, if fingerprint recognition is the next-highest priority, the target user's fingerprint information is collected and compared with pre-stored data. If the secondary verification passes, the automatic process is paused and manual verification is triggered. For example, a surveillance camera transmits real-time footage to a security personnel terminal, who manually confirms the target user's identity and then manually opens the gate. If the secondary verification fails, the subsequent authentication methods are invoked. If all authentication methods fail, the target user is directly prohibited from entering or exiting, and an alarm is issued. By combining automatic verification processes with manual verification, the probability of false positives is effectively reduced, ensuring that only authorized personnel are allowed into high-risk areas. For example, if noise interference at a construction site causes voice recognition to fail, fingerprint recognition is used to verify identity. If the fingerprint matches, a secondary verification check is performed to verify the compliance of the safety equipment worn by the user before allowing entry.

[0091] For example, when a user enters the identity verification area, the server first captures their voice input and compares it with a pre-stored reference audio. If voice verification fails, the secondary verification process is initiated: the server automatically invokes the next authentication method in the pre-set verification order. For example, if voice recognition is currently in priority, iris recognition is next in line. The server then activates the iris sensor to capture the user's eye features and compares them with the iris template in the database. If iris verification still fails, fingerprint verification is invoked to capture the user's fingerprint for comparison. If all verification methods still fail, the server triggers an alarm and logs the exception. If a secondary verification method passes, manual verification is initiated, with a manager remotely retrieving multimodal data from the verification process (such as voice waveform, iris image, and fingerprint feature point distribution) for secondary confirmation. After manual verification confirms the identity is legitimate, the server generates a temporary pass and controls the gate opening.

[0092] This embodiment solves the technical problem of misjudgments caused by the failure of a single verification method in multimodal identity authentication scenarios. By dynamically switching verification levels and combining them with a manual review mechanism, it effectively avoids false identity rejections due to environmental interference or equipment failure, while also preventing unauthorized individuals from circumventing system defenses by forging biometrics. While maintaining the efficiency of automated verification, the coordinated mechanism of hierarchical verification and manual intervention significantly improves the reliability of personnel access control in complex construction environments.

[0093] In some of the above-mentioned schemes of the present invention, when the voice verification result indicates that the identity authentication fails, auxiliary identity authentication is directly performed based on the verification sequence structure. However, if the auxiliary verification result is passed and the person is released directly, there may be security risks of identity fraud or verification method deception, resulting in insufficient security of access control.

[0094] In this regard, the present invention further proposes that S403 may specifically include: Perform auxiliary identity authentication based on the identity authentication method in the next order of the authentication sequence structure to obtain an auxiliary identity authentication result; If the auxiliary identity authentication result indicates that the identity authentication fails, the loop returns to execute the auxiliary identity authentication based on the next identity authentication method in the verification sequence structure to obtain the auxiliary identity authentication result until a preset stop condition is reached. The preset stop condition is that the auxiliary identity authentication result indicates that the identity authentication passes or each identity authentication method in the verification sequence structure is traversed; When the auxiliary identity authentication result indicates that the identity authentication is passed, a manual verification process is triggered, and the entry and exit of the target user is controlled based on the manual verification result.

[0095] In this embodiment, the preset stopping condition is defined as the auxiliary authentication result indicating a successful authentication or traversing all authentication methods in the verification sequence structure. If the auxiliary authentication passes, manual verification processing involves security personnel verifying the user's identity information on-site, such as checking work credentials or inquiring about construction project details. The cyclic execution of the auxiliary authentication follows the priority order pre-set in the verification sequence structure, for example, fingerprint recognition or iris recognition is initiated in sequence after voice recognition fails. The manual verification result and the auxiliary authentication result form a dual confirmation mechanism. For example, if fingerprint verification passes but manual verification finds that the ID information does not match, access will be denied.

[0096] Specifically, if voice verification fails, the server automatically switches to the next verification method in the verification sequence. For example, if facial recognition is the second-priority method, the server activates the camera to collect facial data and compares it with a pre-stored template. If facial recognition fails again, it switches to access card recognition, the third-priority method. If any secondary verification method succeeds, the server triggers a manual verification process, such as sending a verification request containing the user's photo and work number to the duty office. After confirming the authenticity of the identity through on-site observation or document verification, security personnel manually enter the verification results. The server only sends the door opening command to the gate controller if the manual verification result matches the secondary verification result. If all verification methods fail, the server automatically triggers an alarm and records the abnormal access event. For example, if a user fails verification three times in a row, the server locks the gate and uploads the abnormality log to the security management platform.

[0097] Through this embodiment, the problem of inefficiency caused by redundant verification processes in multimodal verification scenarios is solved. Accurate scheduling of verification resources is achieved by dynamically switching verification levels. At the same time, through the connection mechanism between manual verification and automatic verification, the risk of misjudgment caused by environmental interference or equipment abnormalities is effectively avoided, thereby ensuring the decision-making accuracy and execution reliability of the construction site access control system.

[0098] Based on the speech recognition-based construction access control method provided by the present invention, the present invention also provides a specific embodiment of the speech recognition-based construction access control system.

[0099] like Figure 5 , a schematic diagram of a construction access control system based on speech recognition is provided. The construction access control system based on speech recognition 500 may include a structure acquisition module 510 , an audio acquisition module 520 , an audio processing module 530 and a result determination module 540 .

[0100] The structure acquisition module 510 is used to acquire the authentication sequence structure when the target user enters the authentication area. The authentication sequence structure is used to represent the priority order among multiple authentication methods. The audio acquisition module 520 is used to acquire the first audio to be verified input by the target user on site, the pre-stored reference audio of the target user, and the ambient audio of the identity verification area when the first priority of the verification sequence structure is the voice recognition method; The audio processing module 530 is configured to perform denoising on the first audio to be verified based on the ambient audio to obtain a second audio to be verified; The result determination module 540 is used to compare the second audio to be verified with the reference audio to obtain a voice verification result in a voice recognition manner, so as to control the entry and exit of the target user based on the voice verification result.

[0101] In the construction access control system based on voice recognition provided by an embodiment of the present invention, by setting a verification sequence structure, when the first priority is voice recognition, the first audio to be verified input by the target user on site, the pre-stored reference audio and the ambient audio are obtained, and the first audio to be verified is denoised based on the ambient audio to obtain the second audio to be verified, and then compared with the reference audio to obtain the voice verification result. Voice recognition is not affected by the obstruction of protective equipment, and workers do not need to put down their tools to perform additional operations, avoiding the situation of interrupting work by taking a card, and improving access efficiency. At the same time, the denoising process can effectively reduce the interference of environmental noise, making the second audio to be verified clearer, and the voice verification result obtained by comparison with the reference audio is more accurate. In combination with the verification sequence structure, a variety of recognition methods are used scientifically and rationally, which makes up for the shortcomings of traditional methods in the use of multiple recognition methods and comprehensively improves the accuracy of personnel access control. Therefore, the present invention can effectively solve the problems of traditional recognition methods in construction scenarios and can improve the efficiency and accuracy of access control.

[0102] It should be understood that the present invention is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted. In the above embodiments, several specific steps are described and illustrated as examples. However, the method of the present invention is not limited to the specific steps described and illustrated. Those skilled in the art may make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present invention.

[0103] It should also be noted that the exemplary embodiments described herein describe methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the steps described above. In other words, the steps may be performed in the order described in the embodiments, or in a different order, or several steps may be performed simultaneously.

[0104] The above description is only a specific embodiment of the present invention. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited to this. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention.

Claims

1. A construction entry and exit control method based on speech recognition, characterized in that: The method comprises: When the target user enters the identity authentication area, obtaining an authentication sequence structure, wherein the authentication sequence structure is used to represent a priority order among multiple identity authentication methods; In the case where the first order of the verification sequence structure is a voice recognition method, obtaining a first audio to be verified input on-site by the target user, a pre-stored reference audio of the target user, and an ambient audio of the identity verification area; Based on the environmental audio, denoising the first audio to be verified to obtain a second audio to be verified; The second audio to be verified is compared with the reference audio to obtain a voice verification result of the voice recognition method, so as to control the entry and exit of the target user based on the voice verification result.

2. The construction entry and exit control method based on speech recognition according to claim 1 is characterized in that: When the target user enters the identity verification area, before obtaining the verification sequence structure, the method further includes: Obtain historical verification information and attendance record information for various identity authentication methods within a historical time period; Based on each of the historical information to be verified and the corresponding attendance record information, respectively determining the verification validity of each of the identity authentication methods in each of the attendance record information; Based on the verification validity of each identity authentication method in each attendance record information and the verification order of each identity authentication method in each attendance record information, respectively determining the verification priority of each identity authentication method in each verification order; Based on the verification priority of each of the identity authentication methods in each of the verification orders, a verification sequence structure is constructed.

3. The construction entry and exit control method based on speech recognition according to claim 2 is characterized in that: The determining, based on each of the historical information to be verified and the corresponding attendance record information, respectively, of the verification validity of each of the identity authentication methods in each of the attendance record information includes: Determining a verification possibility index for each piece of historical information to be verified based on a difference between each piece of historical information to be verified and the corresponding reference verification information; Determining the verification reliability of each identity authentication method based on the verification possibility indicators included in each identity authentication method; Based on the attendance record information and the verification reliability of each identity authentication method, the verification validity of each identity authentication method in each attendance record information is determined respectively.

4. The construction entry and exit control method based on speech recognition according to claim 3 is characterized in that: The historical information to be verified is the historical audio to be verified, and the reference verification information is the reference audio; Determining the verification possibility index of each piece of historical information to be verified based on the difference between each piece of historical information to be verified and the corresponding reference verification information includes: Determining an audio segment in the historical audio to be verified that is consistent with the reference audio waveform as a valid audio segment; Performing short-time Fourier transform processing on the valid audio segment to obtain the instantaneous phase difference and amplitude spectrum of the valid audio segment in each time frame; sequentially connecting adjacent harmonic peak points in the amplitude spectrum to form a spectrum envelope; Screening a maximum envelope slope difference from each envelope slope difference, wherein the envelope slope difference is a slope difference between adjacent spectrum envelopes; The verification possibility index of the historical audio to be verified is determined by using the variance of the instantaneous phase difference, the maximum envelope slope difference and the spectrum centroid mean, where the spectrum centroid mean is the mean of the spectrum centroids corresponding to each time frame.

5. The construction entry and exit control method based on speech recognition according to claim 3 is characterized in that: The determining of the verification reliability of each identity authentication method based on each verification possibility indicator included in each identity authentication method includes: Obtaining information accuracy of first historical information to be verified, where the first historical information to be verified is each historical information to be verified corresponding to a target identity authentication method, the target identity authentication method being any of the identity authentication methods, and the information accuracy being a degree of match between the first historical information to be verified and the reference authentication information; Performing mean processing on the verification possibility index and the information accuracy of each of the first historical information to be verified, respectively, to obtain a verification possibility index mean and an information accuracy mean; Determining the local verification reliability of each of the first historical information to be verified by using the verification possibility index, the information accuracy, the average of the verification possibility index, and the average of the information accuracy of each of the first historical information to be verified; Based on the local verification reliability of each of the first historical information to be verified, the target verification reliability of the target identity authentication method is determined.

6. The construction entry and exit control method based on speech recognition according to claim 3 is characterized in that: The method of determining the verification validity of each identity authentication method in each attendance record information based on each attendance record information and the verification reliability of each identity authentication method includes: Obtaining a first attendance record vector of the first attendance record information and a first key feature vector in the second historical information to be verified, where the second historical information to be verified is the historical information to be verified of the target identity authentication method corresponding to the first attendance record information, and the target identity authentication method is any one of the identity authentication methods; Obtaining a second attendance record vector of second attendance record information and a second key feature vector in third historical information to be verified, wherein the second attendance record information is the attendance record information previous to the first attendance record information, and the third historical information to be verified is the historical information to be verified of the target identity authentication method corresponding to the second attendance record information; The verification validity of the target identity authentication method in the first attendance record information is determined by using the first attendance record vector, the first key feature vector, the second attendance record vector, the second key feature vector and the verification reliability of the target identity authentication method.

7. The construction entry and exit control method based on speech recognition according to claim 2 is characterized in that: The verification priority of each identity authentication method in each verification sequence is determined based on the verification validity of each identity authentication method in each attendance record information and the verification sequence of each identity authentication method in each attendance record information, including: Filtering out third attendance record information from each of the attendance record information, the third attendance record information being used to represent the attendance record information of the target identity authentication method in the target verification order; The verification priority of the target identity authentication method in the target verification sequence is determined by utilizing the verification validity of the target identity authentication method in each of the third attendance record information and the number of the third attendance record information.

8. The construction entry and exit control method based on speech recognition according to any one of claims 1 to 7, characterized in that: The step of comparing the second audio to be verified with the reference audio to obtain a voice verification result of the voice recognition method, so as to control entry and exit of the target user based on the voice verification result, includes: Comparing the second audio to be verified with the reference audio to obtain a voice verification result of the voice recognition method; If the voice verification result indicates that the identity verification is passed, sending a door opening instruction signal to the controller, so that the controller controls the opening of the gate based on the door opening instruction signal to allow the target user to enter and exit; In the case where the voice verification result indicates that the identity authentication fails, auxiliary identity authentication is performed based on the verification sequence structure, so that the entry and exit of the target user is controlled based on the auxiliary identity authentication result.

9. The construction entry and exit control method based on speech recognition according to claim 8 is characterized in that: The performing auxiliary identity authentication based on the authentication sequence structure, so as to control the entry and exit of the target user based on the auxiliary identity authentication result, includes: Perform auxiliary identity authentication based on the identity authentication method in the next order of the authentication sequence structure to obtain the auxiliary identity authentication result; If the auxiliary identity authentication result indicates that the identity authentication fails, returning to the loop to perform the auxiliary identity authentication based on the identity authentication method at the next order in the verification sequence structure, obtaining the auxiliary identity authentication result, until a preset stop condition is reached, wherein the preset stop condition is that the auxiliary identity authentication result indicates that the identity authentication passes or each identity authentication method in the verification sequence structure is traversed; In the case where the auxiliary identity authentication result indicates that the identity authentication is passed, a manual verification process is triggered, and the entry and exit of the target user is controlled based on the manual verification result.

10. A construction entry and exit control system based on voice recognition, characterized in that: The system comprises: A structure acquisition module is used to acquire a verification sequence structure when a target user enters the identity authentication area, wherein the verification sequence structure is used to represent a priority order among multiple identity authentication methods; an audio acquisition module, configured to acquire, when the first priority of the verification sequence structure is a voice recognition mode, a first audio to be verified inputted on-site by the target user, a pre-stored reference audio of the target user, and an ambient audio of the identity verification area; an audio processing module, configured to perform denoising on the first audio to be verified based on the ambient audio to obtain a second audio to be verified; A result determination module is used to compare the second audio to be verified with the reference audio to obtain a voice verification result of the voice recognition method, so as to control the entry and exit of the target user based on the voice verification result.

Images