Voice quality detection system

Abstract

The invention relates to the technical field of voice data processing, in particular to a voice quality detection system, which is characterized in that a monitoring data acquisition module is used for acquiring monitoring video information in a user house; the sound data analysis module is used for analyzing the environment sound characteristics and the sound type of the monitoring video information; the interaction data acquisition module is used for acquiring interaction information of intelligent equipment in a user housing; the interaction data analysis module is used for storing the monitoring video information to obtain historical interaction information; the video data analysis module is used for performing frame extraction operation on the monitoring video information to obtain a monitoring frame image and analyzing image feature parameters; the historical data analysis module is used for analyzing the interactive voice features; the voice quality analysis module is used for analyzing the quality evaluation parameters; and the voice quality evaluation module is used for analyzing and outputting the voice quality. According to the invention, accurate analysis of voice quality is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of voice data processing, and in particular to a voice quality detection system. Background Art

[0002] The detection of voice signal quality in surveillance videos mainly relies on technical means such as computer vision, image processing and data analysis. With the widespread application of voice communication and video communication, it becomes very important to ensure efficient monitoring and optimization of real-time voice quality.

[0003] Chinese Patent Publication No.: CN102737480A discloses an abnormal voice monitoring system and method based on intelligent video, including a monitoring area field device, an intelligent monitoring and analysis device connected to the monitoring area field device, and a monitoring center computer respectively connected to the monitoring area field device and the intelligent monitoring and analysis device; the monitoring area field device includes a video acquisition module, a voice receiving module triggered by the intelligent monitoring and analysis device, and an explosion-proof warning module triggered by the monitoring center computer; the intelligent monitoring and analysis device includes a central processor respectively connected to the video acquisition module and the voice receiving module, whose output end is connected to the monitoring center computer, a data storage module connected to the central processor, a video processing module connected to the central processor, and a sound intensity detection module connected to the central processor. The invention realizes the abnormal recognition and detection of voice signals in security monitoring, but does not realize the comprehensive detection and analysis of voice quality by combining video data, sound data and interactive data of surrounding intelligent devices. There are problems of low analysis efficiency of voice data in the video and inaccurate analysis of voice quality. Summary of the invention

[0004] The object of the present invention is to provide a voice quality detection system to solve at least one of the problems existing in the prior art.

[0005] To achieve the above object, the present invention adopts the following technical solutions: A voice quality detection system, comprising: A monitoring data collection module is used to collect monitoring video information in the user's house; A sound data analysis module is used to analyze the environmental sound characteristics according to the monitoring video information, and is also used to analyze the sound type of the monitoring video information according to the environmental sound characteristics; An interactive data collection module is used to collect interactive information of smart devices in the user's house according to the sound type of the monitoring video information; An interactive data analysis module is used to store the monitoring video information according to the interactive information and the sound type of the monitoring video information to obtain historical interactive information; The video data analysis module is used to extract frames from the monitoring video information to obtain monitoring frame images, and is also used to analyze image feature parameters based on the monitoring frame images; A historical data analysis module is used to analyze interactive voice features based on historical interactive information; A voice quality analysis module is used to analyze quality assessment parameters based on surveillance video information, environmental sound characteristics, sound type of surveillance video information, image feature parameters and interactive voice characteristics; The speech quality assessment module is used to analyze and output the speech quality according to the quality assessment parameters.

[0006] Furthermore, the sound data analysis module is provided with a sound feature analysis unit, which is used to extract the monitoring sound data within K seconds before the current analysis time as environmental analysis data, and analyze the environmental sound features according to the audio time domain diagram of the environmental analysis data to obtain the environmental sound feature E(t), where t represents the time number and K represents the time extraction parameter.

[0007] Furthermore, the sound data analysis module is also provided with a sound type analysis unit, which is used to compare the environmental sound feature E(t) with the environmental comparison threshold α, and analyze the sound type of the monitoring video information according to the comparison result to determine the sound type of the monitoring video information, and the sound type of the monitoring video information includes one category and two categories.

[0008] Furthermore, the interaction data analysis module stores the monitoring video information whose interaction information is voice interaction and whose sound type is Class II, and uses the stored monitoring video information as historical interaction information.

[0009] Furthermore, the video data analysis module is provided with a video frame extraction processing unit, which is used to extract m frames of images from the monitoring video data per second as monitoring frame images, wherein m represents a frame extraction parameter; The video data analysis module is also provided with an image feature analysis unit, which is used to analyze the gray value L of each pixel in the monitoring frame image. t (x, y) analyzes the image feature parameters to obtain the image feature parameters P(t).

[0010] Furthermore, the historical data analysis module analyzes the historical normalization parameter according to the historical interaction information to obtain the historical normalization parameter B(t), and analyzes the interactive speech feature according to the historical normalization parameter B(t) to determine the interactive speech feature G.

[0011] Furthermore, the voice quality analysis module is provided with a voice quality analysis unit, which is used to analyze the quality assessment parameters according to the monitoring sound data and the environmental sound characteristics E(t) when the sound type of the monitoring video information is Class II, so as to obtain the quality assessment parameter H(t).

[0012] Furthermore, the speech quality analysis module is also provided with an interactive feature analysis unit, which is used to compare the interactive speech feature G with the interactive comparison threshold g. When the interactive speech feature meets the threshold, the analysis process of the quality assessment parameter is adjusted, and the adjusted quality assessment parameter is H1(t).

[0013] Furthermore, the speech quality analysis module is also provided with an image feature analysis unit, which is used to count the number of difference pixels in the monitoring frame image as the difference number, and compare the difference number NL(t) with the difference threshold β. When the difference number does not meet the threshold, the adjustment process of the quality assessment parameter is optimized, and the optimized quality assessment parameter is H2(t).

[0014] Furthermore, the speech quality assessment module compares the quality assessment parameter H(t) with the quality assessment thresholds h1 and h2, and analyzes the speech quality according to the comparison result to determine the speech quality, where the speech quality includes: poor, good and excellent.

[0015] The beneficial effects of the present invention are as follows: through the collection of user monitoring video information and interactive information by the monitoring data collection module and the interactive data collection module, and the comprehensive analysis of the collected data by other modules, voice quality detection of sound data in the monitoring video is realized, and comprehensive analysis of video data, sound data and interactive data is realized, thereby improving the system's analysis efficiency of voice data in the video and improving the accuracy of voice quality analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0017] Figure 1 Schematic diagram of the structure of the voice quality detection system of this embodiment.

[0018] Figure 2 Schematic diagram of the structure of the sound data analysis module of this embodiment.

[0019] Figure 3 Schematic diagram of the structure of the video data analysis module of this embodiment.

[0020] Figure 4 Schematic diagram of the structure of the voice quality analysis module in this embodiment. DETAILED DESCRIPTION

[0021] In order to more clearly illustrate the present invention, the present invention is further described below in conjunction with preferred embodiments and accompanying drawings. Similar components in the accompanying drawings are represented by the same reference numerals. It should be understood by those skilled in the art that the content specifically described below is illustrative rather than restrictive, and should not be used to limit the scope of protection of the present invention.

[0022] It should be noted that, although the terms first, second, third, etc. may be used to describe in the embodiments of the present application, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.

[0023] See also Figure 1 As shown, the voice quality detection system of this embodiment includes: The monitoring data acquisition module is used to collect monitoring video information in the user's house. The video monitoring information includes monitoring video data and monitoring sound data. The monitoring video data is the video data in the monitoring video, and the monitoring sound data is the sound data in the monitoring video. The monitoring video information is obtained by importing the background data of the smart camera data management system used by the user.

[0024] Specifically, this embodiment is applied to the quality analysis of voice signal acquisition in indoor monitoring, ensuring the data quality of voice data in the monitoring data and meeting the signal quality requirements of users when reviewing and querying monitoring.

[0025] Please continue reading Figure 1 As shown, the voice quality detection system also includes: A sound data analysis module, used to analyze the environmental sound characteristics according to the monitoring video information, and also used to analyze the sound type of the monitoring video information according to the environmental sound characteristics, the sound data analysis module is connected to the monitoring data acquisition module; See also Figure 2 As shown, the sound data analysis module includes: The waveform analysis unit is used to convert the monitoring sound data into an audio time domain diagram. The horizontal axis of the audio time domain diagram represents time in seconds, and the vertical axis represents amplitude in V. The waveform analysis unit analyzes the monitoring sound data to convert the sound data into an audio time domain diagram, and the amplitude in the sound data is matched one by one with the time when it is generated. The sound data is quantified to ensure the accuracy of the system's analysis of the sound data, thereby improving the system's analysis efficiency of the voice data in the video and improving the accuracy of the voice quality analysis.

[0026] Please continue reading Figure 2 As shown, the sound data analysis module also includes: The sound feature analysis unit is used to analyze the environmental sound features according to the audio time domain diagram, and use the environmental sound features to represent the difference features of the environmental noise data that may exist in the video. The sound feature analysis unit is connected to the limp analysis unit.

[0027] Specifically, the sound feature analysis unit in this embodiment extracts the monitoring sound data within K seconds before the current analysis time as the environmental analysis data, and analyzes the environmental sound features according to the audio time domain diagram of the environmental analysis data, sets the environmental sound features as E(t), and sets , , where D(t) represents the amplitude of the monitoring sound data, t represents the time number, k represents the environmental analysis parameter, k∈N and k≤K, K represents the time extraction parameter, 60≤K≤120, and A(t) represents the amplitude mean feature; the monitoring sound data is analyzed by the sound feature analysis unit to extract the environmental analysis data, realize the extraction of data within a period of time, set the sample data range for analysis, thereby analyzing the environmental sound features, and realizing the analysis of the difference characteristics of the sound amplitude within a period of time, thereby improving the system's analysis efficiency of the voice data in the video and improving the accuracy of the voice quality analysis; it can be understood that the present embodiment does not make specific restrictions on the value of the time extraction parameter, and the technicians in this field can set it freely, as long as it meets the analysis of the environmental sound features. The optimal value of the time extraction parameter is: K=60.

[0028] Please continue reading Figure 2 As shown, the sound data analysis module also includes: The sound type analysis unit is used to analyze the sound type of the monitoring video information according to the environmental sound characteristics, and divide the sound data in the monitoring video information into two categories according to the size of its short-term amplitude. The sound type analysis unit is connected to the sound feature analysis unit.

[0029] Specifically, the sound type analysis unit described in this embodiment compares the environmental sound characteristics with the environmental comparison threshold, and analyzes the sound type of the monitoring video information according to the comparison result. If D(t) / [A(t)-E(t)]≤α, the sound type analysis unit determines that the sound type of the monitoring video information is one category; if D(t) / [A(t)-E(t)]>α, the sound type analysis unit determines that the sound type of the monitoring video information is two categories; wherein α represents the environmental comparison threshold, 0.8≤α<1; the sound type analysis unit analyzes the environmental sound characteristics to determine the sound type of the monitoring video information, and realizes the analysis of the sound amplitude data in the video to determine whether it is environmental noise, thereby improving the system's analysis efficiency of the voice data in the video and improving the accuracy of the voice quality analysis; it can be understood that in this embodiment, the value of the environmental comparison threshold is not specifically limited, and the technicians in this field can freely set it, as long as it satisfies the analysis of the sound type of the monitoring video information. The optimal value of the environmental comparison threshold is: α=0.9.

[0030] Please continue reading Figure 1 As shown, the voice quality detection system also includes: The interactive data collection module is used to collect interactive information of smart devices in the user's house according to the sound type of the monitoring video information. The interactive data collection module is connected to the sound data analysis module. The smart devices in this embodiment include but are not limited to smart speakers, smart TVs and other devices that can perform voice interaction. The interactive information includes voice interaction and non-interaction. The interactive information is data obtained by judging whether the smart device responds to the user's interaction request. The interactive information is obtained by importing the background data of the data management system running on the smart device used by the user.

[0031] Please continue reading Figure 1 As shown, the voice quality detection system also includes: The interactive data analysis module is used to store the monitoring video information according to the interactive information and the sound type of the monitoring video information to obtain the historical interactive information. The interactive data analysis module is connected to the interactive data acquisition module.

[0032] Specifically, the interaction data analysis module in this embodiment stores the monitoring video information in which the interaction information is voice interaction and the sound type of the monitoring video information is Class II, and uses the stored monitoring video information as historical interaction information; the interaction data analysis module analyzes the sound types of the interaction information and the monitoring video information to extract and store the fragments of the monitoring video information in which the user performs voice interaction with the smart device, thereby increasing the diversity of sample analysis, thereby improving the system's analysis efficiency of the voice data in the video and improving the accuracy of the voice quality analysis.

[0033] Please continue reading Figure 1 As shown, the voice quality detection system also includes: The video data analysis module is used to perform frame extraction operation on the monitoring video information to obtain the monitoring frame image, and is also used to analyze the image feature parameters according to the monitoring frame image. The video data analysis module is connected to the monitoring data acquisition module.

[0034] See also Figure 3 As shown, the video data analysis module includes: The video frame extraction processing unit is used to perform frame extraction operation on the monitoring video data to obtain the monitoring frame image.

[0035] Specifically, the video frame extraction processing unit in this embodiment extracts m frames of images from the monitoring video data per second as monitoring frame images, wherein m represents a frame extraction parameter, 1≤m≤3; the video frame extraction processing unit performs a frame extraction operation on the monitoring video data, so as to use each frame in the extracted video as a monitoring frame image, thereby realizing the extraction of analysis data every second in the monitoring video, thereby improving the system's analysis efficiency of the voice data in the video and improving the accuracy of the voice quality analysis; in this embodiment, the frame extraction parameter is set to 1, and python+opencv is used to implement video frame extraction, so as to realize the extraction of 1 frame from each second of the monitoring video data as a monitoring frame image. It can be understood that the frame extraction operation of the monitoring video data is not specifically limited in this embodiment, and technical personnel in this field can freely set it, such as using video processing tools such as ffmpeg to perform frame extraction operations on the video.

[0036] Please continue reading Figure 3 As shown, the video data analysis module also includes: The image feature analysis unit is used to analyze the image feature parameters according to the monitoring frame image, and use the image feature parameters to represent the difference characteristics of the grayscale value of each pixel in the image. The image feature analysis unit is connected to the video frame extraction processing unit.

[0037] Specifically, when analyzing the monitoring frame image, the image feature analysis unit described in this embodiment takes the pixel point in the lower left corner of the monitoring frame image as the coordinate origin, and takes the two sides adjacent to the coordinate origin as the x-axis and y-axis respectively. The x-axis increases from left to right, and the y-axis increases from bottom to top. The unit length is 1 pixel. A plane rectangular coordinate system is established, and the coordinate points are used to represent the position of each pixel in the monitoring frame image.

[0038] Specifically, the image feature analysis unit in this embodiment analyzes the image feature parameter according to the gray value of each pixel in the monitoring frame image, sets the image feature parameter to P(t), and sets , where L t(x, y) represents the grayscale value of each pixel in the monitoring frame image, and NXY represents the number of pixels in the monitoring frame image.

[0039] The image feature analysis unit analyzes the monitoring frame image to analyze the image feature parameters, realizes the comprehensive feature analysis of the gray value of each pixel in the monitoring frame image, ensures the accuracy of the system's analysis of video data, thereby improving the system's analysis efficiency of voice data in the video and improving the accuracy of voice quality analysis.

[0040] Please continue reading Figure 1 As shown, the voice quality detection system also includes: The historical data analysis module is used to analyze the interactive voice features according to the historical interactive information. The historical data analysis module is connected with the video data analysis module and the interactive data analysis module.

[0041] Specifically, the historical data analysis module in this embodiment analyzes the historical normalization parameter according to the historical interaction information, sets the historical normalization parameter to B(t), and sets B(t)=D(t) max -D(t) / [D(t) max -D(t) min ], t∈U, and analyze the interactive speech features according to the historical normalization parameters, set the interactive speech features to G, and set , where D(t) max Indicates the maximum amplitude of the monitored sound data in the historical interaction information, D(t) min represents the minimum value of the amplitude of the monitoring sound data in the historical interaction information, U represents the set of time numbers in the historical interaction information, and NU represents the number of elements in the set of time numbers in the historical interaction information.

[0042] By analyzing historical interaction information through the historical data analysis module, voice feature analysis in the monitoring video data when the user performs voice interaction can be realized, and difference analysis between voice signals can be realized, thereby improving the system's analysis efficiency of voice data in the video and improving the accuracy of voice quality analysis.

[0043] Please continue reading Figure 1 As shown, the voice quality detection system also includes: The voice quality analysis module is used to analyze the quality assessment parameters according to the monitoring video information, environmental sound characteristics, the sound type of the monitoring video information, the image feature parameters and the interactive voice characteristics. The voice quality analysis module is connected to the historical data analysis module.

[0044] See also Figure 4 As shown, the voice quality analysis module includes: The voice quality analysis unit is used to analyze the quality evaluation parameters according to the monitoring video information and the environmental sound characteristics when the sound type of the monitoring video information is the second type.

[0045] Specifically, when the sound type of the monitoring video information is Class II, the voice quality analysis unit in this embodiment analyzes the quality assessment parameters according to the monitoring sound data and the environmental sound characteristics, sets the quality assessment parameters to H(t), and sets H(t)=[D(t)-E(t)] / A(t).

[0046] The voice quality analysis unit analyzes the monitoring video information and environmental sound characteristics to analyze the quality evaluation parameters, and uses the quality evaluation parameters to represent the short-time energy characteristics of sound data with large sound information amplitude, thereby improving the system's analysis efficiency of voice data in the video and improving the accuracy of voice quality analysis.

[0047] Please continue reading Figure 4 As shown, the voice quality analysis module also includes: The interactive feature analysis unit is used to adjust the analysis process of the quality assessment parameter according to the interactive voice feature so that the adjusted quality assessment parameter is related to the sound feature of the user when performing voice interaction. The interactive feature analysis unit is connected to the voice quality analysis unit.

[0048] Specifically, the interactive feature analysis unit in this embodiment compares the interactive speech feature with the interactive comparison threshold, and adjusts the analysis process of the quality assessment parameter according to the comparison result. If G≥g, the interactive feature analysis unit determines that the interactive speech feature meets the threshold, and adjusts the analysis process of the quality assessment parameter. The adjusted quality assessment parameter is H1(t), and H1(t)=H(t)×e 1-G ; If G<g, the interactive feature analysis unit determines that the interactive speech feature does not meet the threshold and does not adjust the analysis process of the quality assessment parameters; wherein g represents the interactive comparison threshold, 0.7≤g≤0.8; the interactive feature analysis unit analyzes the interactive speech feature to adjust the analysis process of the quality assessment parameters, and adjusts the quality assessment parameters when the difference in the interactive speech features is small, so as to achieve a comprehensive analysis of the interactive speech data and the current sound data, thereby improving the system's analysis efficiency of the speech data in the video and improving the accuracy of the speech quality analysis; it can be understood that in this embodiment, the value of the interactive comparison threshold is not specifically limited, and technical personnel in this field can set it freely, as long as the adjustment of the quality assessment parameter analysis process is satisfied. The optimal value of the interactive comparison threshold is: g=0.75.

[0049] Please continue reading Figure 4 As shown, the voice quality analysis module also includes: The image feature analysis unit is used to optimize the adjustment process of the quality assessment parameter according to the image feature parameter so that the optimized quality assessment parameter is related to the characteristics of the image data in the monitoring video. The image feature analysis unit is connected to the interactive feature analysis unit.

[0050] Specifically, the image feature analysis unit in this embodiment analyzes the image similarity according to the image feature parameters, sets the image similarity to C(t), sets C(t)=P(t) / P(t-1); and when |C(t)-1|<c, analyzes the difference pixels according to the monitoring frame image, if |L t (x,y)-L t-1 (x, y)|<P(t), the image feature analysis unit determines that the current analysis pixel is a similar pixel. If |L t (x,y)-L t-1 (x,y)|≥P(t), the image feature analysis unit determines that the current analysis pixel is a difference pixel; wherein c represents the similarity comparison threshold, 0.1≤c≤0.2. It can be understood that the similarity comparison threshold is not specifically limited in this embodiment, and those skilled in the art can freely set it as long as it satisfies the analysis of the difference pixel. The optimal value of the similarity comparison threshold is: c=0.15.

[0051] Specifically, the image feature analysis unit in this embodiment counts the number of difference pixels in the monitoring frame image as the difference number, and compares the difference number with the difference threshold, and optimizes the adjustment process of the quality assessment parameter according to the comparison result. If NL(t) / NXY≤β, the image feature analysis unit determines that the difference number meets the threshold, and does not optimize the adjustment process of the quality assessment parameter; if NL(t) / NXY>β, the image feature analysis unit determines that the difference number does not meet the threshold, and optimizes the adjustment process of the quality assessment parameter. The optimized quality assessment parameter is H2(t), and H2(t)=H1(t)×e is set. NL(t) / [NXY×2] , where NL(t) represents the number of differences, β represents the difference threshold, and 0.15≤β≤0.3. The image feature analysis unit analyzes the image feature parameters to determine the similarity between adjacent images, thereby analyzing the number of differences, and analyzing the changes in the image in the surveillance video to determine whether there are image changes and sounds in the video at the time when the sound signal exists, thereby improving the system's analysis efficiency of the voice data in the video and improving the accuracy of the voice quality analysis; it can be understood that the value of the difference threshold is not specifically limited in this embodiment, and those skilled in the art can freely set it as long as the optimization of the quality assessment parameter adjustment process is satisfied. The optimal value of the difference threshold is: β=0.2.

[0052] Please continue reading Figure 1 As shown, the voice quality detection system also includes: The voice quality assessment module is used to analyze and output the voice quality according to the quality assessment parameters, and the voice quality assessment module is connected to the voice quality analysis module.

[0053] Specifically, the voice quality assessment module described in this embodiment compares the quality assessment parameter with the quality assessment threshold, and analyzes the voice quality according to the comparison result. If H(t)

[0054] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the embodiments here. All obvious changes or modifications derived from the technical solution of the present invention are still within the protection scope of the present invention.​

Claims

1. A voice quality detection system, characterized in that: include: A monitoring data collection module is used to collect monitoring video information in the user's house; A sound data analysis module is used to analyze the environmental sound characteristics according to the monitoring video information, and is also used to analyze the sound type of the monitoring video information according to the environmental sound characteristics; An interactive data collection module is used to collect interactive information of smart devices in the user's house according to the sound type of the monitoring video information; An interactive data analysis module is used to store the monitoring video information according to the interactive information and the sound type of the monitoring video information to obtain historical interactive information; The video data analysis module is used to extract frames from the monitoring video information to obtain monitoring frame images, and is also used to analyze image feature parameters based on the monitoring frame images; A historical data analysis module is used to analyze interactive voice features based on historical interactive information; A voice quality analysis module is used to analyze quality assessment parameters based on surveillance video information, environmental sound characteristics, sound type of surveillance video information, image feature parameters and interactive voice characteristics; The speech quality assessment module is used to analyze and output the speech quality according to the quality assessment parameters.

2. The voice quality detection system according to claim 1, characterized in that: The sound data analysis module is provided with a sound feature analysis unit, which is used to extract the monitoring sound data within K seconds before the current analysis time as the environmental analysis data, and analyze the environmental sound features according to the audio time domain diagram of the environmental analysis data to obtain the environmental sound feature E(t), where t represents the time number and K represents the time extraction parameter.

3. The voice quality detection system according to claim 2, characterized in that: The sound data analysis module is also provided with a sound type analysis unit, which is used to compare the environmental sound feature E(t) with the environmental comparison threshold α, and analyze the sound type of the monitoring video information according to the comparison result to determine the sound type of the monitoring video information, and the sound type of the monitoring video information includes one category and two categories.

4. The voice quality detection system according to claim 3, characterized in that: The interaction data analysis module stores the monitoring video information whose interaction information is voice interaction and whose sound type is Class II, and uses the stored monitoring video information as historical interaction information.

5. The voice quality detection system according to claim 4, characterized in that: The video data analysis module is provided with a video frame extraction processing unit, which is used to extract m frames of images per second from the monitoring video data as monitoring frame images, wherein m represents a frame extraction parameter; The video data analysis module is also provided with an image feature analysis unit, which is used to analyze the gray value L of each pixel in the monitoring frame image. t (x, y) analyzes the image feature parameters to obtain the image feature parameters P(t).

6. The voice quality detection system according to claim 5, characterized in that: The historical data analysis module analyzes the historical normalization parameter according to the historical interaction information to obtain the historical normalization parameter B(t), and analyzes the interactive speech feature according to the historical normalization parameter B(t) to determine the interactive speech feature G.

7. The voice quality detection system according to claim 6, characterized in that: The voice quality analysis module is provided with a voice quality analysis unit, which is used to analyze the quality evaluation parameter according to the monitoring sound data and the environmental sound characteristics E(t) when the sound type of the monitoring video information is the second type, so as to obtain the quality evaluation parameter H(t).

8. The voice quality detection system according to claim 7, characterized in that: The speech quality analysis module is also provided with an interactive feature analysis unit, which is used to compare the interactive speech feature G with the interactive comparison threshold g. When the interactive speech feature meets the threshold, the analysis process of the quality assessment parameter is adjusted, and the adjusted quality assessment parameter is H1(t).

9. The voice quality detection system according to claim 8, characterized in that: The speech quality analysis module is also provided with an image feature analysis unit, which is used to count the number of difference pixels in the monitoring frame image as the difference number, and compare the difference number NL(t) with the difference threshold β. When the difference number does not meet the threshold, the adjustment process of the quality assessment parameter is optimized, and the optimized quality assessment parameter is H2(t).

10. The voice quality detection system according to claim 9, characterized in that: The speech quality assessment module compares the quality assessment parameter H(t) with the quality assessment thresholds h1 and h2, and analyzes the speech quality according to the comparison result to determine the speech quality, which includes: poor, good and excellent.

Images