A noise reduction method and device based on state monitoring of directional pickup reactor
Patent Information
- Application Number
- CN202411327930.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-24
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-09-24
AI Technical Summary
The audio files collected by traditional reactor voiceprint recognition systems have a high signal-to-noise ratio under environmental noise interference, which cannot guarantee the reliability of equipment failure judgment and affect the user experience.
Using a noise reduction method based on directional pickup, the plane image of the reactor is obtained through the top camera, the outline pattern is identified, the target position of the microphone array is determined, the beam guide algorithm and inverse Fourier transform are used to process the audio signal, combined with the deep learning model , accurately monitor the reactor status to reduce noise interference.
It improves the accuracy and reliability of the soundprint monitoring of reactors, reduces the impact of environmental noise on detection, and enhances the reliability of fault determination.
Smart Images

Figure CN118843027B_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the technical field of signal processing, and in particular relates to a noise reduction method and device based on state monitoring of a directional pickup reactor. Background Art
[0002] In recent years, the voiceprint recognition monitoring of reactors has been gradually applied. There are two main solutions for fault recognition of voiceprints: reverse recognition, that is, judging the fault by identifying the characteristics of the sound when the monitoring equipment fails; and forward recognition, that is, judging the fault by the proportion of the sound of the equipment deviating from the normal operation. Regardless of the solution, the interference of environmental noise will affect the monitoring results, so how to filter the interference of environmental noise during monitoring has become an urgent problem to be solved.
[0003] The audio files collected by the traditional reactor voiceprint recognition system have a high signal-to-noise ratio when there is environmental noise interference, which cannot guarantee the reliability of determining whether the equipment is faulty, thus affecting the user experience of the reactor. Summary of the invention
[0004] In order to solve the technical defects mentioned above, such as the audio files collected by the traditional reactor voiceprint recognition system have high signal-to-noise ratio when there is environmental noise interference, and the reliability of the device failure judgment cannot be guaranteed, which further affects the user experience of the reactor, a noise reduction method and device based on directional pickup reactor state monitoring is proposed, and its technical solution is as follows:
[0005] In a first aspect, an embodiment of the present application provides a noise reduction method based on state monitoring of a directional pickup reactor, comprising:
[0006] A planar image of the reactor is acquired based on a top camera, and a contour graphic of the reactor is identified in the planar image of the reactor; wherein the top camera is located directly above the reactor;
[0007] According to the contour graph of the reactor, a minimum regular graph including the contour graph of the reactor is marked, and a center point of the minimum regular graph and any two non-adjacent first vertices are determined;
[0008] Determine whether the center point and the line connecting the two first vertices include any two non-adjacent second vertices in the contour graph of the reactor;
[0009] When it is determined that the connecting line includes any two non-adjacent second vertices in the contour graph of the reactor, the two first vertices are used as target positions of the microphone array, and the initial audio signal collected by the microphone array is obtained based on the target positions;
[0010] The initial audio signal is processed based on a beam steering algorithm to obtain a three-dimensional sound source model, and sound source data corresponding to a central symmetrical position of the microphone array is determined in the three-dimensional sound source model; wherein the sound source data corresponding to the central position of the microphone array includes at least two frequency values and a sound pressure value corresponding to each frequency value;
[0011] The sound source data corresponding to the center position of the microphone array is processed by inverse Fourier transform to obtain the target audio signal.
[0012] In an optional solution of the first aspect, after determining whether the connection line between the center point and the two first vertices includes any two non-adjacent second vertices in the contour graph of the reactor, the method further includes:
[0013] When it is determined that the connecting line does not include any two non-adjacent second vertices in the contour graph of the reactor, marking the first intersection point between the connecting line and the contour graph of the reactor, and the third vertex closest to each first intersection point in the contour graph of the reactor;
[0014] Mark the midpoint between each first intersection point and the nearest third vertex in the contour graph of the reactor, and determine the second intersection point between the extended line of the connection and the minimum regular graph based on the extended line of the connection of all the midpoints;
[0015] The second intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0016] In another optional solution of the first aspect, after marking the first intersection points between the connecting line and the contour graph of the reactor and the third vertex closest to each first intersection point in the contour graph of the reactor, the method further includes:
[0017] When it is detected that the distance between any first intersection point and the nearest third vertex is within a preset distance interval, the first intersection point is used as the fourth vertex;
[0018] When it is detected that the distance between any first intersection point and the nearest third vertex exceeds the preset distance interval, the midpoint between the first intersection point and the nearest third vertex is marked in the contour graph of the reactor, and the third intersection point between the extended line of the connection line and the minimum regular graph is determined based on the extended line of the connection line between the midpoint and the fourth vertex;
[0019] The third intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0020] In another optional solution of the first aspect, the initial audio signal is processed based on a beam steering algorithm to obtain a three-dimensional sound source model, including:
[0021] Based on a preset target array element signal expression and an initial audio signal, an average power expression is constructed; wherein the preset target array element signal expression includes a preset target vector, a preset target signal, and a preset interference signal;
[0022] Calculate the target weight vector according to the average power expression and the preset proportional constant;
[0023] The initial audio signal and the target weight vector are processed based on the beam steering algorithm to obtain a three-dimensional sound source model.
[0024] In another optional solution of the first aspect, after performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain the target audio signal, the method further includes:
[0025] Inputting the initial audio signal into a deep learning model to obtain a first prediction result; wherein the deep learning model is trained based on sample audio signals of multiple known fault types;
[0026] Inputting the target audio signal into the deep learning model to obtain a second prediction result, and when the first prediction result is a device failure and the second prediction result is a device failure, determining that the reactor has a device failure; or
[0027] When the first prediction result is no fault, and the second prediction result is a device fault, it is determined that the reactor has a device fault.
[0028] In another optional solution of the first aspect, after inputting the target audio signal into the deep learning model to obtain the second prediction result, the method further includes:
[0029] When the first prediction result is that the device is faulty, and the second prediction result is that there is no fault, it is determined that the reactor has environmental interference.
[0030] In a second aspect, an embodiment of the present application provides a noise reduction device based on state monitoring of a directional pickup reactor, comprising:
[0031] An image recognition module, used for acquiring a planar image of the reactor based on a top camera, and identifying a contour of the reactor in the planar image of the reactor; wherein the top camera is located directly above the reactor;
[0032] An image processing module, used for marking a minimum regular graph including the contour graph of the reactor according to the contour graph of the reactor, and determining a center point of the minimum regular graph and any two non-adjacent first vertices;
[0033] An image judgment module is used to judge whether the center point and the line connecting the two first vertices contain any two non-adjacent second vertices in the contour graph of the reactor;
[0034] A signal acquisition module, for, when it is determined that the connecting line includes any two non-adjacent second vertices in the contour graph of the reactor, using the two first vertices as target positions of the microphone array, and acquiring the initial audio signal collected by the microphone array based on the target positions;
[0035] A signal analysis module, used to process the initial audio signal based on a beam steering algorithm to obtain a three-dimensional sound source model, and determine the sound source data corresponding to the central symmetric position of the microphone array in the three-dimensional sound source model; wherein the sound source data corresponding to the central position of the microphone array includes at least two frequency values and a sound pressure value corresponding to each frequency value;
[0036] The signal processing module is used to perform inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain the target audio signal.
[0037] In an optional solution of the second aspect, the device further includes:
[0038] After determining whether the center point and the connecting line between the two first vertices include any two non-adjacent second vertices in the contour graph of the reactor, when it is determined that the connecting line does not include any two non-adjacent second vertices in the contour graph of the reactor, marking the first intersection point between the connecting line and the contour graph of the reactor, and the third vertex closest to each first intersection point in the contour graph of the reactor;
[0039] Mark the midpoint between each first intersection point and the nearest third vertex in the contour graph of the reactor, and determine the second intersection point between the extended line of the connection and the minimum regular graph based on the extended line of the connection of all the midpoints;
[0040] The second intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0041] In yet another optional solution of the second aspect, the device further comprises:
[0042] After marking the first intersection point between the connecting line and the contour graph of the reactor and the third vertex closest to each first intersection point in the contour graph of the reactor, when it is detected that the distance between any first intersection point and the third vertex closest to the first intersection point is within a preset distance interval, the first intersection point is used as the fourth vertex;
[0043] When it is detected that the distance between any first intersection point and the nearest third vertex exceeds the preset distance interval, the midpoint between the first intersection point and the nearest third vertex is marked in the contour graph of the reactor, and the third intersection point between the extended line of the connection line and the minimum regular graph is determined based on the extended line of the connection line between the midpoint and the fourth vertex;
[0044] The third intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0045] In yet another optional solution of the second aspect, the signal analysis module includes:
[0046] Based on a preset target array element signal expression and an initial audio signal, an average power expression is constructed; wherein the preset target array element signal expression includes a preset target vector, a preset target signal, and a preset interference signal;
[0047] Calculate the target weight vector according to the average power expression and the preset proportional constant;
[0048] The initial audio signal and the target weight vector are processed based on the beam steering algorithm to obtain a three-dimensional sound source model.
[0049] In yet another optional solution of the second aspect, the device further comprises:
[0050] After performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain the target audio signal, the initial audio signal is input into the deep learning model to obtain a first prediction result; wherein the deep learning model is trained based on sample audio signals of multiple known fault types;
[0051] Inputting the target audio signal into the deep learning model to obtain a second prediction result, and when the first prediction result is a device failure and the second prediction result is a device failure, determining that the reactor has a device failure; or
[0052] When the first prediction result is no fault, and the second prediction result is a device fault, it is determined that the reactor has a device fault.
[0053] In yet another optional solution of the second aspect, the device further comprises:
[0054] After the target audio signal is input into the deep learning model and the second prediction result is obtained, when the first prediction result is a device failure and the second prediction result is no failure, it is determined that there is environmental interference in the reactor.
[0055] In a third aspect, an embodiment of the present application further provides a noise reduction device based on state monitoring of a directional pickup reactor, comprising a processor and a memory;
[0056] The processor is connected to the memory;
[0057] A memory for storing executable program codes;
[0058] The processor runs the program corresponding to the executable program code by reading the executable program code stored in the memory, so as to implement the noise reduction method based on directional pickup reactor state monitoring provided by the first aspect of the embodiment of the present application or any implementation method of the first aspect.
[0059] In a fourth aspect, an embodiment of the present application provides a computer storage medium, which stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, the noise reduction method based on directional pickup reactor state monitoring provided by the first aspect of the embodiment of the present application or any one of the implementation methods of the first aspect can be implemented.
[0060] In an embodiment of the present application, in the process of signal denoising for voiceprint monitoring of a reactor, a planar image of the reactor can be obtained based on a top camera, and a contour graphic of the reactor can be identified in the planar image of the reactor; based on the contour graphic of the reactor, a minimum regular graphic including the contour graphic of the reactor is marked, and the center point of the minimum regular graphic and any two non-adjacent first vertices are determined; it is determined whether the line connecting the center point and the two first vertices includes any two non-adjacent second vertices in the contour graphic of the reactor; when it is determined that the line includes any two non-adjacent second vertices in the contour graphic of the reactor, the two first vertices are used as target positions of the microphone array, and an initial audio signal collected by the microphone array is obtained based on the target position; the initial audio signal is processed based on a beam steering algorithm to obtain a three-dimensional sound source model, and the sound source data corresponding to the central symmetric position of the microphone array is determined in the three-dimensional sound source model; the sound source data corresponding to the center position of the microphone array is subjected to inverse Fourier transform processing to obtain a target audio signal. The target position of the microphone array is determined by the contour of the inductor, so that the microphone array can monitor the center position of the inductor more accurately, and the noise outside the monitoring range is greatly attenuated; secondly, the initial audio signal collected by the microphone array can be processed to obtain a directional target audio signal, which not only makes the sound pressure level attenuated more as it deviates from the center position, but also effectively avoids the interference of environmental noise on the detection sound collection, thereby improving the accuracy and reliability of the voiceprint monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0062] Figure 1 An overall flow chart of a noise reduction method based on state monitoring of a directional pickup reactor provided in an embodiment of the present application;
[0063] Figure 2 A schematic diagram of the recognition effect of a plane image of a reactor provided in an embodiment of the present application;
[0064] Figure 3 A schematic diagram of a marking effect of a plane image of a reactor provided in an embodiment of the present application;
[0065] Figure 4 A schematic diagram of the effect of a three-dimensional sound source model provided in an embodiment of the present application;
[0066] Figure 5 A schematic diagram of the structure of a noise reduction device based on state monitoring of a directional pickup reactor provided in an embodiment of the present application;
[0067] Figure 6 A schematic structural diagram of another noise reduction device based on state monitoring of a directional pickup reactor provided in an embodiment of the present application. DETAILED DESCRIPTION
[0068] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.
[0069] In the following introduction, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. The following introduction provides multiple embodiments of the present application, and different embodiments can be replaced or combined, so the present application can also be considered to include all possible combinations of the same and / or different embodiments recorded. Therefore, if one embodiment includes features A, B, and C, and another embodiment includes features B and D, then the present application should also be considered to include embodiments containing one or more of all other possible combinations of A, B, C, and D, although the embodiment may not be clearly recorded in the following text.
[0070] The following description provides examples and does not limit the scope, applicability or examples set forth in the claims. Changes may be made to the functions and arrangements of the elements described without departing from the scope of the present application. Various processes or components may be appropriately omitted, substituted or added to each example. For example, the described method may be performed in an order different from the order described, and various steps may be added, omitted or combined. In addition, features described in some examples may be combined in other examples.
[0071] See also Figure 1 , Figure 1 An overall flow chart of a noise reduction method based on state monitoring of a directional pickup reactor provided in an embodiment of the present application is shown.
[0072] like Figure 1 As shown, the noise reduction method based on the state monitoring of the directional pickup reactor may at least include the following steps:
[0073] Step 102: Acquire a planar image of the reactor based on the top camera, and identify a contour graphic of the reactor in the planar image of the reactor.
[0074] Specifically, in the signal denoising process of voiceprint monitoring of the reactor, the top camera can be used to obtain a top-down plane image of the reactor from above the reactor, and the top-down plane image can be used to more clearly obtain the planar structure of the reactor. Among them, the top camera can be, but is not limited to, pre-set above the reactor or other positions, and moved to the top of the reactor according to the planned path, and the distance between the top camera and the reactor is within a preset distance range, so that the target position of the microphone array can be more accurately determined from the top-down plane image. It can be understood that the position of the microphone array directly determines the signal-to-noise ratio of the collected audio signal, and in the traditional reactor voiceprint recognition system, the signal-to-noise ratio of the audio file collected when there is environmental noise interference is relatively high, and thus the reliability of its determination of whether the equipment is faulty cannot be guaranteed.
[0075] Furthermore, after the top camera acquires the top-down plane image of the reactor, image recognition processing can be performed on the top-down plane image of the reactor, but is not limited to, based on a visual algorithm to determine the contour feature points of the reactor in the top-down plane image of the reactor, and determine the contour graphic of the reactor based on the contour feature points of the reactor.
[0076] See here Figure 2 FIG. 1 is a schematic diagram showing the recognition effect of a plane image of a reactor provided by an embodiment of the present application, such as Figure 2As shown, the upper part is a plane image including the top-view structure of the reactor captured by the top camera, and the lower part is the contour graphic of the reactor identified based on the plane image of the upper part. The contour graphic of the reactor can be obtained by connecting multiple determined contour feature points, and the multiple contour feature points can be but are not limited to the various vertices in the top-view structure of the reactor (the black dots shown in the figure are contour feature points, that is, the various vertices in the top-view structure of the reactor).
[0077] It is understandable that, combined with the above Figure 2 The outline of the reactor shown and the top view of a conventional reactor show that the outline of the reactor is generally irregular (i.e. not symmetrical), which makes it difficult to directly determine the center position of the reactor, and further makes it impossible to effectively and accurately obtain the target position of the microphone array.
[0078] Step 104: Mark a minimum regular graph including the contour graph of the reactor according to the contour graph of the reactor, and determine the center point of the minimum regular graph and any two non-adjacent first vertices.
[0079] Specifically, after determining the contour of the reactor based on the visual algorithm, the minimum regular pattern containing the contour area of the reactor can be marked according to the contour of the reactor. The minimum regular pattern can be understood as the minimum regular pattern that can completely cover the contour area of the reactor. The regular pattern here can be, but is not limited to, a rectangle or a circle. Preferably, in the embodiment of the present application, the regular pattern can be a square or a rectangle. Figure 2 In the contour figure of the reactor shown, the long side of the minimum regular figure can coincide with the outermost long side in the contour figure of the reactor, and the length of the long side of the minimum regular figure is greater than the length of the outermost long side in the contour figure of the reactor; the wide side of the minimum regular figure can coincide with the outermost wide side in the contour figure of the reactor, and the length of the wide side of the minimum regular figure is greater than the length of the outermost wide side in the contour figure of the reactor.
[0080] Furthermore, after marking the minimum regular figure, the center point and any two non-adjacent first vertices corresponding to the minimum regular figure can be determined according to the structure of the minimum regular figure. Here, taking the minimum regular figure as a rectangle as an example, the distances from the center point to the four vertices in the minimum regular figure are the same. The any two non-adjacent first vertices can be understood as the upper left vertex and the lower right vertex in the minimum regular figure, or the lower left vertex and the upper right vertex in the minimum regular figure.
[0081] Step 106: Determine whether the center point and the line connecting the two first vertices include any two non-adjacent second vertices in the contour graph of the reactor.
[0082] Specifically, after determining the center point and any two non-adjacent first vertices in the minimum regular figure, the center point and any two non-adjacent first vertices can be connected, and it can be determined whether the intersection between the connection line and the contour figure of the reactor is any two non-adjacent second vertices in the contour figure of the reactor. The judgment method here can be, but is not limited to, establishing a plane rectangular coordinate system with the minimum regular figure as the plane, taking any vertex in the minimum regular figure as the origin in the plane rectangular coordinate system, taking two adjacent sides in the minimum regular figure as the X-axis and Y-axis in the plane rectangular coordinate system, and determining the center point coordinates in the minimum regular figure, the coordinates of any two non-adjacent first vertices, and all the vertex coordinates in the contour figure of the reactor in the plane rectangular coordinate system, and then establishing a connection function based on the center point coordinates and the coordinates of any two non-adjacent first vertices, and by substituting all the vertex coordinates in the contour figure of the reactor into the connection function in turn, it can be effectively determined whether the connection line contains any two non-adjacent second vertices in the contour figure of the reactor.
[0083] It can be understood that when it is detected that the line connecting the center point and the two first vertices includes any two non-adjacent second vertices in the contour figure of the reactor, it indicates that the contour figure of the reactor is a regular figure, that is, the center position of the contour figure of the reactor is on the line; when it is detected that the line connecting the center point and the two first vertices does not include any two non-adjacent second vertices in the contour figure of the reactor, or any vertex, it indicates that the contour figure of the reactor is an irregular figure.
[0084] Step 108: When it is determined that the connecting line includes any two non-adjacent second vertices in the contour graph of the reactor, the two first vertices are used as target positions of the microphone array, and the initial audio signal collected by the microphone array is obtained based on the target positions.
[0085] Specifically, after determining any two non-adjacent second vertices in the contour graph of the connecting line containing the inductor, in order to maximize the detection range of the microphone array for the inductor and avoid receiving too much ambient noise, the two first vertices in the minimum regular graph can be used as the target positions of the microphone array, that is, the microphone array is placed on each first vertex, and the initial audio signal collected by the microphone array at each first vertex is obtained.
[0086] It can be understood that the effective receiving signal angle range of a general microphone array is about 90 degrees. Taking the minimum regular figure as a rectangle as an example, when the microphone array is set on any two non-adjacent vertices of the minimum regular figure, it can ensure that the receiving signal of the microphone array completely covers the inductor, thereby improving the reliability and effectiveness of the audio signal collected by the microphone array.
[0087] As an optional embodiment of the present application, after determining whether the line connecting the center point and the two first vertices includes any two non-adjacent second vertices in the contour graph of the reactor, the method further includes:
[0088] When it is determined that the connecting line does not include any two non-adjacent second vertices in the contour graph of the reactor, marking the first intersection point between the connecting line and the contour graph of the reactor, and the third vertex closest to each first intersection point in the contour graph of the reactor;
[0089] Mark the midpoint between each first intersection point and the nearest third vertex in the contour graph of the reactor, and determine the second intersection point between the extended line of the connection and the minimum regular graph based on the extended line of the connection of all the midpoints;
[0090] The second intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0091] Specifically, after determining that the connecting line does not include any two non-adjacent second vertices in the contour graph of the inductor, in order to ensure that the detection range of the microphone array for the inductor is maximized and avoid receiving excessive ambient noise, the first intersection points (usually two) between the connecting line and the contour graph of the inductor and the third vertex closest to each first intersection point can be marked in the contour graph of the inductor.
[0092] Since the contour graph of the reactor is an irregular graph, the approximate center position of the reactor can be determined through the first intersection and the third vertex, and the target position of the microphone array can be obtained based on the approximate center position of the reactor. Furthermore, the midpoints between each first intersection and the nearest third vertex (usually two) can be marked in the contour graph of the reactor, and all the midpoints can be connected. At this time, the approximate center position of the reactor is on the line connecting all the midpoints, and the second intersection between the extension line of the connection and the minimum regular graph can be used as the target position of the microphone array. This target position can ensure that the microphone array maximizes the detection range of the reactor. Then, the microphone array can be placed on each second intersection, and the initial audio signal collected by the microphone array at each second intersection can be obtained.
[0093] See also Figure 3 , Figure 3 FIG. 1 is a schematic diagram showing a marking effect of a plane image of a reactor provided by an embodiment of the present application. Figure 3 As shown, the plane image of the reactor contains Figure 2 The contour graph of the reactor is consistent with the minimum regular graph (rectangle) marked by the dotted line. The contour graph of the reactor includes multiple vertices. The lower left vertex and the upper right vertex in the minimum regular graph are used as the first vertex and marked as 301, the center point is marked as 302, and the intersection between the line connecting the center point 301 and the first vertex 302 and the contour graph of the reactor is used as the first intersection and marked as 303 ( Figure 3 Only one first intersection is marked, and the other is not marked), the third vertex closest to the first intersection 303 is marked as 304 (only the third vertex corresponding to 303 is marked, and the other is not marked), and the midpoint between the first intersection 30 and the third vertex 304 closest to the first intersection 303 can be marked as 305 (the other is not marked). It can be understood that the approximate center position of the reactor is on the line connecting all the midpoints 305, and the second intersection between the extended line of the line and the minimum regular figure can be used as the target position of the microphone array.
[0094] As another optional embodiment of the present application, after marking the first intersection point between the connecting line and the contour graph of the reactor and the third vertex closest to each first intersection point in the contour graph of the reactor, the method further includes:
[0095] When it is detected that the distance between any first intersection point and the nearest third vertex is within a preset distance interval, the first intersection point is used as the fourth vertex;
[0096] When it is detected that the distance between any first intersection point and the nearest third vertex exceeds the preset distance interval, the midpoint between the first intersection point and the nearest third vertex is marked in the contour graph of the reactor, and the third intersection point between the extended line of the connection line and the minimum regular graph is determined based on the extended line of the connection line between the midpoint and the fourth vertex;
[0097] The third intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0098] In order to improve the overall calculation efficiency of the noise reduction method, when it is detected that the distance between any one of the two first intersection points and the nearest third vertex is within a preset distance interval, it indicates that the distance between the first intersection point and the third vertex is small (that is, the midpoint between the first intersection point and the third vertex can be approximately taken as the first intersection point), and the first intersection point can be taken as the fourth vertex for ease of calculation. It can be understood that the third vertex closest to the first intersection point can also be taken as the fourth vertex here, and it is not limited thereto.
[0099] Furthermore, when it is detected that the distance between any one of the two first intersections and the nearest third vertex exceeds a preset distance interval, the midpoint between the first intersection and the nearest third vertex is marked in the contour graph of the reactor, and a line connection can be performed based on the midpoint and the fourth vertex mentioned above. At this time, the approximate center position of the reactor is on the line connecting the midpoint and the fourth vertex, and the second intersection between the extended line of the line and the minimum regular figure can be used as the target position of the microphone array. This target position can ensure that the detection range of the microphone array for the reactor is maximized. Then the microphone array can be placed on each second intersection, and the initial audio signal collected by the microphone array at each second intersection can be obtained.
[0100] Step 110: Process the initial audio signal based on a beam steering algorithm to obtain a three-dimensional sound source model, and determine the sound source data corresponding to the central symmetric position of the microphone array in the three-dimensional sound source model.
[0101] After obtaining an initial audio signal with higher reliability and effectiveness based on the target position, since the initial audio signal can be understood as the sum of the receiving signals of each array element and the corresponding weight vector, it is possible but not limited to calculating the target weight vector pointing to the center position (or approximate center position) of the reactor based on the initial audio signal, and obtaining the corresponding three-dimensional sound source model through a beam steering algorithm (also called a DOA algorithm).
[0102] As another option of the embodiment of the present application, the initial audio signal is processed based on a beam steering algorithm to obtain a three-dimensional sound source model, including:
[0103] Based on a preset target array element signal expression and an initial audio signal, an average power expression is constructed; wherein the preset target array element signal expression includes a preset target vector, a preset target signal, and a preset interference signal;
[0104] Calculate the target weight vector according to the average power expression and the preset proportional constant;
[0105] The initial audio signal and the target weight vector are processed based on the beam steering algorithm to obtain a three-dimensional sound source model.
[0106] Specifically, since the initial audio signal can be understood as the sum of the received signals of each array element and the corresponding weight vector, the expression corresponding to the initial audio signal can be converted but not limited to first, which is expressed as follows:
[0107]
[0108] In the above formula, It can be understood as the initial audio signal of all array elements on the microphone array when the time index is n. It can be understood as a weight matrix. It can be understood as the received signal of each array element when the time index is n, m can be understood as the mth array, and M can be understood as the number of all arrays. It can be understood that the beam steering algorithm can be completed by weighting and summing the received signals output by each array element to converge the beams received by the microphone array in one direction, and the convergence direction where the desired signal obtains the maximum output power.
[0109] Here, based on a preset target vector, a preset target signal, and a preset interference signal, a received signal expression of each array element may be constructed, which may be, but is not limited to, expressed as follows:
[0110]
[0111] In the above formula, It can be understood as the received signal of the kth array element at time index t, which corresponds to ; It can be understood that the kth array element comes from The direction vector of It can be understood as a preset target vector; It can be understood that the kth array element comes from The j-th direction vector of It can be understood as the jth preset interference signal when the time index is t, and J can be understood as the number of interference signals. It can be understood as the white noise signal of the kth array element at time index t (also preset).
[0112] Next, the received signal expression of each array element and the initial audio signal may be combined to construct an average power expression, and the average power expression may be converted, but is not limited to, as follows:
[0113]
[0114] In the above formula, It can be understood as the average power, and N can be understood as the preset number of receiving end snapshots.
[0115] Next, when N approaches infinity, the above average power expression can be transformed, and the following transformation process can be referred to but not limited to:
[0116]
[0117] In the above formula, R can be understood as a diagonal matrix, It can be understood as a weight vector.
[0118] Next, by solving the minimum value of the average power expression after the above conversion, the target weight vector expression as shown below is obtained:
[0119]
[0120] In the above formula, Can be understood as receiving direction The optimal weight vector for beamforming, It can be understood as a preset proportional constant. It can be understood that the wave direction comes from The direction vector of .
[0121] It should be noted that the above method of calculating the target weight vector of the initial audio signal can also refer to the well-known DOA technology, which will not be described in detail here.
[0122] Furthermore, after calculating the target weight vector of the initial audio signal, the initial audio signal and the target weight vector can be processed based on a beam steering algorithm to obtain a three-dimensional sound source model. The beam steering algorithm is a common audio processing in the art and will not be described in detail here.
[0123] See also Figure 4 , Figure 4 FIG. 1 is a schematic diagram showing the effect of a three-dimensional sound source model provided by an embodiment of the present application. Figure 4 As shown, the three-dimensional sound source model can be understood as being drawn according to the sound pressure distribution in the space grid after the beamforming confirms the sound incident direction, and its XOY plane can be simulated as the space plane formed by the microphone array facing, and the Z axis can be the sound pressure value of each position detected by the microphone array. It can be understood that the three-dimensional sound source model includes multiple three-dimensional sound source models corresponding to different frequencies, that is, for the same position, it can include sound pressure values corresponding to multiple frequencies.
[0124] Furthermore, after constructing the three-dimensional sound source model, the center position can be determined on the XOY plane in the three-dimensional sound source model. The center position can be understood as the central symmetric position of the microphone array, that is, corresponding to the center position or approximate center position of the reactor, and the sound pressure values corresponding to the center position at different frequencies can be screened out.
[0125] Step 112: Perform inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain a target audio signal.
[0126] Specifically, after the sound pressure values corresponding to the different frequencies of the center position are screened, the different frequencies and the sound pressure values corresponding to the respective frequencies can be subjected to inverse Fourier transform processing to restore the sound data to the time axis and obtain a target audio signal with directionality. It can be understood that the sound pressure level attenuation of the target audio signal is greater for the more deviated center position, which can effectively avoid the interference of environmental noise on the detection target sound collection, and in a complex environment, can effectively avoid the problem of false alarms and noise interference with the detection object.
[0127] As another optional embodiment of the present application, after performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain the target audio signal, the method further includes:
[0128] Inputting the initial audio signal into a deep learning model to obtain a first prediction result; wherein the deep learning model is trained based on sample audio signals of multiple known fault types;
[0129] Inputting the target audio signal into the deep learning model to obtain a second prediction result, and when the first prediction result is a device failure and the second prediction result is a device failure, determining that the reactor has a device failure; or
[0130] When the first prediction result is no fault, and the second prediction result is a device fault, it is determined that the reactor has a device fault.
[0131] In order to further ensure the accuracy of reactor fault detection, the initial audio signal can be input into the deep learning model to obtain a first prediction result. The first prediction result is used to compare the prediction result corresponding to the target audio signal. It is not the only basis for judging whether the reactor is faulty. The first prediction result can be but is not limited to representing two results: no fault or equipment failure. For example, when the first prediction result is character A, it can correspond to no fault; when the first prediction result is character B, it can correspond to equipment failure.
[0132] It is understandable that the deep learning model can also predict the specific type of equipment failure. For example, the corresponding type of equipment failure can be determined based on the different types of characters output. I will not go into details here.
[0133] Furthermore, the target audio signal can also be input into the deep learning model to obtain a second prediction result, and when it is detected that the first prediction result corresponds to a device failure, and the second prediction result corresponds to a device failure, it is determined that there is a device failure in the reactor.
[0134] Possibly, when it is detected that the first prediction result corresponds to no fault and the second prediction result corresponds to an equipment fault, the second prediction result is mainly used to determine that there is an equipment fault in the reactor.
[0135] Possibly, when it is detected that the first prediction result corresponds to a device fault, and the second prediction result corresponds to no fault, it is determined that the reactor has environmental interference.
[0136] It is also understandable that after the problem with the reactor is determined, feedback can be provided to the staff in a timely manner to avoid losses and risks.
[0137] See also Figure 5 , Figure 5 A structural schematic diagram of a noise reduction device based on state monitoring of a directional pickup reactor provided in an embodiment of the present application is shown.
[0138] like Figure 5 As shown, the noise reduction device based on directional pickup reactor state monitoring may include at least an image recognition module 501, an image processing module 502, an image judgment module 503, a signal acquisition module 504, a signal analysis module 505 and a signal processing module 506, wherein:
[0139] An image recognition module 501 is used to obtain a planar image of the reactor based on a top camera, and recognize a contour of the reactor in the planar image of the reactor; wherein the top camera is located directly above the reactor;
[0140] The image processing module 502 is used to mark the minimum regular graph including the contour graph of the reactor according to the contour graph of the reactor, and determine the center point of the minimum regular graph and any two non-adjacent first vertices;
[0141] An image determination module 503 is used to determine whether the center point and the line connecting the two first vertices contain any two non-adjacent second vertices in the contour graph of the reactor;
[0142] The signal acquisition module 504 is used to, when it is determined that the line includes any two non-adjacent second vertices in the contour graph of the reactor, use the two first vertices as target positions of the microphone array, and acquire the initial audio signal collected by the microphone array based on the target positions;
[0143] The signal analysis module 505 is used to process the initial audio signal based on the beam steering algorithm to obtain a three-dimensional sound source model, and determine the sound source data corresponding to the central symmetric position of the microphone array in the three-dimensional sound source model; wherein the sound source data corresponding to the central position of the microphone array includes at least two frequency values and a sound pressure value corresponding to each frequency value;
[0144] The signal processing module 506 is used to perform inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain a target audio signal.
[0145] In some possible embodiments, the device further includes:
[0146] After determining whether the center point and the connecting line between the two first vertices include any two non-adjacent second vertices in the contour graph of the reactor, when it is determined that the connecting line does not include any two non-adjacent second vertices in the contour graph of the reactor, marking the first intersection point between the connecting line and the contour graph of the reactor, and the third vertex closest to each first intersection point in the contour graph of the reactor;
[0147] Mark the midpoint between each first intersection point and the nearest third vertex in the contour graph of the reactor, and determine the second intersection point between the extended line of the connection and the minimum regular graph based on the extended line of the connection of all the midpoints;
[0148] The second intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0149] In some possible embodiments, the device further includes:
[0150] After marking the first intersection point between the connecting line and the contour graph of the reactor and the third vertex closest to each first intersection point in the contour graph of the reactor, when it is detected that the distance between any first intersection point and the third vertex closest to the first intersection point is within a preset distance interval, the first intersection point is used as the fourth vertex;
[0151] When it is detected that the distance between any first intersection point and the nearest third vertex exceeds the preset distance interval, the midpoint between the first intersection point and the nearest third vertex is marked in the contour graph of the reactor, and the third intersection point between the extended line of the connection line and the minimum regular graph is determined based on the extended line of the connection line between the midpoint and the fourth vertex;
[0152] The third intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0153] In some possible embodiments, the signal analysis module includes:
[0154] Based on a preset target array element signal expression and an initial audio signal, an average power expression is constructed; wherein the preset target array element signal expression includes a preset target vector, a preset target signal, and a preset interference signal;
[0155] Calculate the target weight vector according to the average power expression and the preset proportional constant;
[0156] The initial audio signal and the target weight vector are processed based on the beam steering algorithm to obtain a three-dimensional sound source model.
[0157] In some possible embodiments, the device further includes:
[0158] After performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain the target audio signal, the initial audio signal is input into the deep learning model to obtain a first prediction result; wherein the deep learning model is trained based on sample audio signals of multiple known fault types;
[0159] Inputting the target audio signal into the deep learning model to obtain a second prediction result, and when the first prediction result is a device failure and the second prediction result is a device failure, determining that the reactor has a device failure; or
[0160] When the first prediction result is no fault, and the second prediction result is a device fault, it is determined that the reactor has a device fault.
[0161] In some possible embodiments, the device further includes:
[0162] After the target audio signal is input into the deep learning model and the second prediction result is obtained, when the first prediction result is a device failure and the second prediction result is no failure, it is determined that there is environmental interference in the reactor.
[0163] Those skilled in the art can clearly understand that the technical solutions of the embodiments of the present application can be implemented with the help of software and / or hardware. The "unit" and "module" in this specification refer to software and / or hardware that can independently complete or cooperate with other components to complete specific functions, where the hardware can be, for example, a field programmable gate array (FPGA), an integrated circuit (IC), etc.
[0164] See also Figure 6 , Figure 6 A structural schematic diagram of another noise reduction device based on state monitoring of a directional pickup reactor provided in an embodiment of the present application is shown.
[0165] like Figure 6 As shown, the noise reduction device 600 based on state monitoring of a directional pickup reactor may include at least one processor 601 , at least one network interface 604 , a user interface 603 , a memory 605 and at least one communication bus 602 .
[0166] The communication bus 602 may be used to realize the connection and communication among the above-mentioned components.
[0167] The user interface 603 may include buttons, and the optional user interface may also include a standard wired interface or a wireless interface.
[0168] The network interface 604 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, etc.
[0169] Among them, the processor 601 may include one or more processing cores. The processor 601 uses various interfaces and lines to connect various parts in the noise reduction device 600 based on the state monitoring of the directional pickup reactor, and executes various functions and processes data of the noise reduction device 600 based on the state monitoring of the directional pickup reactor by running or executing instructions, programs, code sets or instruction sets stored in the memory 605, and calling the data stored in the memory 605. Optionally, the processor 601 can be implemented in at least one hardware form of DSP, FPGA, and PLA. The processor 601 can integrate one or a combination of CPU, GPU, modem, etc. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 601, but implemented by a single chip.
[0170] The memory 605 may include a RAM or a ROM. Optionally, the memory 605 includes a non-transitory computer-readable medium. The memory 605 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 605 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 605 may optionally be at least one storage device located away from the aforementioned processor 601. As Figure 6 As shown, the memory 605 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a noise reduction application based on state monitoring of a directional pickup reactor.
[0171] Specifically, the processor 601 may be used to call the noise reduction application based on the state monitoring of the directional pickup reactor stored in the memory 605, and specifically perform the following operations:
[0172] A planar image of the reactor is acquired based on a top camera, and a contour graphic of the reactor is identified in the planar image of the reactor; wherein the top camera is located directly above the reactor;
[0173] According to the contour graph of the reactor, a minimum regular graph including the contour graph of the reactor is marked, and a center point of the minimum regular graph and any two non-adjacent first vertices are determined;
[0174] Determine whether the center point and the line connecting the two first vertices include any two non-adjacent second vertices in the contour graph of the reactor;
[0175] When it is determined that the connecting line includes any two non-adjacent second vertices in the contour graph of the reactor, the two first vertices are used as target positions of the microphone array, and the initial audio signal collected by the microphone array is obtained based on the target positions;
[0176] The initial audio signal is processed based on a beam steering algorithm to obtain a three-dimensional sound source model, and sound source data corresponding to a central symmetrical position of the microphone array is determined in the three-dimensional sound source model; wherein the sound source data corresponding to the central position of the microphone array includes at least two frequency values and a sound pressure value corresponding to each frequency value;
[0177] The sound source data corresponding to the center position of the microphone array is processed by inverse Fourier transform to obtain the target audio signal.
[0178] In some possible embodiments, after determining whether the connection line between the center point and the two first vertices includes any two non-adjacent second vertices in the contour graph of the reactor, the method further includes:
[0179] When it is determined that the connecting line does not include any two non-adjacent second vertices in the contour graph of the reactor, marking the first intersection point between the connecting line and the contour graph of the reactor, and the third vertex closest to each first intersection point in the contour graph of the reactor;
[0180] Mark the midpoint between each first intersection point and the nearest third vertex in the contour graph of the reactor, and determine the second intersection point between the extended line of the connection and the minimum regular graph based on the extended line of the connection of all the midpoints;
[0181] The second intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0182] In some possible embodiments, after marking the first intersection points between the connecting line and the contour graph of the reactor and the third vertex closest to each first intersection point in the contour graph of the reactor, the method further includes:
[0183] When it is detected that the distance between any first intersection point and the nearest third vertex is within a preset distance interval, the first intersection point is used as the fourth vertex;
[0184] When it is detected that the distance between any first intersection point and the nearest third vertex exceeds the preset distance interval, the midpoint between the first intersection point and the nearest third vertex is marked in the contour graph of the reactor, and the third intersection point between the extended line of the connection line and the minimum regular graph is determined based on the extended line of the connection line between the midpoint and the fourth vertex;
[0185] The third intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
[0186] In some possible embodiments, the initial audio signal is processed based on a beam steering algorithm to obtain a three-dimensional sound source model, including:
[0187] Based on a preset target array element signal expression and an initial audio signal, an average power expression is constructed; wherein the preset target array element signal expression includes a preset target vector, a preset target signal, and a preset interference signal;
[0188] Calculate the target weight vector according to the average power expression and the preset proportional constant;
[0189] The initial audio signal and the target weight vector are processed based on the beam steering algorithm to obtain a three-dimensional sound source model.
[0190] In some possible embodiments, after performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain the target audio signal, the method further includes:
[0191] Inputting the initial audio signal into a deep learning model to obtain a first prediction result; wherein the deep learning model is trained based on sample audio signals of multiple known fault types;
[0192] Inputting the target audio signal into the deep learning model to obtain a second prediction result, and when the first prediction result is a device failure and the second prediction result is a device failure, determining that the reactor has a device failure; or
[0193] When the first prediction result is no fault, and the second prediction result is a device fault, it is determined that the reactor has a device fault.
[0194] In some possible embodiments, after inputting the target audio signal into the deep learning model to obtain the second prediction result, the method further includes:
[0195] When the first prediction result is that the device is faulty, and the second prediction result is that there is no fault, it is determined that the reactor has environmental interference.
[0196] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a micro drive, and a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.
[0197] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0198] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0199] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.
[0200] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0201] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0202] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, disk or optical disk and other media that can store program codes.
[0203] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by entering a program to instruct related hardware, and the program may be stored in a computer-readable memory, which may include a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.
[0204] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field that are not recorded in the present disclosure. The description and examples are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.
Claims
1. A noise reduction method based on state monitoring of a directional pickup reactor, characterized in that: include: Acquire a planar image of the reactor based on a top camera, and identify a contour of the reactor in the planar image of the reactor; wherein the top camera is located directly above the reactor; According to the contour graph of the reactor, mark a minimum regular graph including the contour graph of the reactor, and determine a center point of the minimum regular graph and any two non-adjacent first vertices; Determine whether the center point and the line connecting the two first vertices include any two non-adjacent second vertices in the contour graph of the reactor; When it is determined that the connecting line includes any two non-adjacent second vertices in the contour graph of the reactor, taking the two first vertices as target positions of the microphone array, and acquiring the initial audio signal collected by the microphone array based on the target positions; The initial audio signal is processed based on a beam steering algorithm to obtain a three-dimensional sound source model, and sound source data corresponding to a central symmetrical position of the microphone array is determined in the three-dimensional sound source model; wherein the sound source data corresponding to the central position of the microphone array includes at least two frequency values and a sound pressure value corresponding to each of the frequency values; Performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain a target audio signal; The processing of the initial audio signal based on a beam steering algorithm to obtain a three-dimensional sound source model includes: Based on a preset target array element signal expression and the initial audio signal, construct an average power expression; wherein the preset target array element signal expression includes a preset target vector, a preset target signal, and a preset interference signal; Calculating a target weight vector according to the average power expression and a preset proportional constant; The initial audio signal and the target weight vector are processed based on a beam steering algorithm to obtain a three-dimensional sound source model.
2. The method according to claim 1, characterized in that: After determining whether the connection line between the center point and the two first vertices includes any two non-adjacent second vertices in the contour graph of the reactor, the method further includes: When it is determined that the connecting line does not include any two non-adjacent second vertices in the contour graph of the reactor, marking the first intersection points between the connecting line and the contour graph of the reactor, and the third vertices closest to each of the first intersection points in the contour graph of the reactor; Marking the midpoint between each of the first intersection points and the nearest third vertex in the contour graph of the reactor, and determining the second intersection point between the extended line of the connection and the minimum regular graph based on the extended line of the connection of all the midpoints; The second intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
3. The method according to claim 2, characterized in that After marking the first intersection points between the connecting line and the contour graph of the reactor and the third vertices closest to each of the first intersection points in the contour graph of the reactor, the method further includes: When it is detected that the distance between any one of the first intersection points and the nearest third vertex is within a preset distance interval, the first intersection point is used as the fourth vertex; When it is detected that the distance between any one of the first intersection points and the nearest third vertex exceeds the preset distance interval, the midpoint between the first intersection point and the nearest third vertex is marked in the contour graph of the reactor, and based on the extension line of the connection line between the midpoint and the fourth vertex, the third intersection point between the extension line and the minimum regular graph is determined; The third intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
4. The method according to claim 1, characterized in that: After performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain the target audio signal, the method further includes: Inputting the initial audio signal into a deep learning model to obtain a first prediction result; wherein the deep learning model is trained based on sample audio signals of multiple known fault types; Inputting the target audio signal into the deep learning model to obtain a second prediction result, and when the first prediction result is a device failure and the second prediction result is a device failure, determining that the reactor has a device failure; or When the first prediction result is no fault, and the second prediction result is a device fault, it is determined that there is a device fault in the reactor.
5. The method according to claim 4, characterized in that After inputting the target audio signal into the deep learning model to obtain a second prediction result, the method further includes: When the first prediction result is a device failure, and the second prediction result is no failure, it is determined that the reactor has environmental interference.
6. A noise reduction device based on state monitoring of a directional pickup reactor, characterized in that: include: An image recognition module, used for acquiring a planar image of the reactor based on a top camera, and identifying a contour of the reactor in the planar image of the reactor; wherein the top camera is located directly above the reactor; An image processing module, used for marking a minimum regular figure including the outline figure of the reactor according to the outline figure of the reactor, and determining a center point of the minimum regular figure and any two non-adjacent first vertices; An image judgment module, used for judging whether the center point and the line connecting the two first vertices contain any two non-adjacent second vertices in the contour graph of the reactor; A signal acquisition module, configured to, when it is determined that the connecting line includes any two non-adjacent second vertices in the contour graph of the reactor, use the two first vertices as target positions of the microphone array, and acquire the initial audio signal collected by the microphone array based on the target positions; A signal analysis module, used to process the initial audio signal based on a beam steering algorithm to obtain a three-dimensional sound source model, and determine the sound source data corresponding to the central symmetric position of the microphone array in the three-dimensional sound source model; wherein the sound source data corresponding to the central position of the microphone array includes at least two frequency values and a sound pressure value corresponding to each of the frequency values; A signal processing module, used for performing inverse Fourier transform processing on the sound source data corresponding to the center position of the microphone array to obtain a target audio signal; The processing of the initial audio signal based on a beam steering algorithm to obtain a three-dimensional sound source model includes: Based on a preset target array element signal expression and the initial audio signal, construct an average power expression; wherein the preset target array element signal expression includes a preset target vector, a preset target signal, and a preset interference signal; Calculating a target weight vector according to the average power expression and a preset proportional constant; The initial audio signal and the target weight vector are processed based on a beam steering algorithm to obtain a three-dimensional sound source model.
7. The device according to claim 6, characterized in that The device also includes: After determining whether the connecting line between the center point and the two first vertices includes any two non-adjacent second vertices in the contour graph of the reactor, when it is determined that the connecting line does not include any two non-adjacent second vertices in the contour graph of the reactor, marking in the contour graph of the reactor a first intersection point between the connecting line and the contour graph of the reactor, and a third vertex closest to each of the first intersection points; Marking the midpoint between each of the first intersection points and the nearest third vertex in the contour graph of the reactor, and determining the second intersection point between the extended line of the connection and the minimum regular graph based on the extended line of the connection of all the midpoints; The second intersection point is used as a target position of the microphone array, and an initial audio signal collected by the microphone array is acquired based on the target position.
8. A noise reduction device based on state monitoring of a directional pickup reactor, characterized in that: including a processor and a memory; The processor is connected to the memory; The memory is used to store executable program code; The processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the steps of the method according to any one of claims 1 to 5.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a computer or a processor, the computer or the processor executes the steps of the method according to any one of claims 1 to 5.
Citation Information
Patent Citations
Sound localization system based on microphone array
CN102707262A
Method and system for detecting state of loudspeaker by using machine auditory sense
CN117119355A
