A method, device and storage medium for monitoring aluminum leakage in deep well casting

Abstract

The present application relates to the technical field of aluminum processing safety monitoring, and in particular to a method, device, and storage medium for monitoring aluminum leakage in deep-well casting. The method comprises: selecting a visual monitoring target area, and performing visual analysis of acquired visual features based on the visual monitoring target area to obtain visual monitoring results, wherein the objects of the visual monitoring results include workers and workpieces; acquiring sound signals and preprocessing the sound signals; performing time-frequency processing on the preprocessed sound signals to obtain a time-frequency graph; calculating the intensity values corresponding to the sound signals and, based on the intensity values, calculating the proportion of high-intensity components in the time-frequency graph, and comparing the proportion of high-intensity components with a preset value to obtain sound monitoring results; and performing aluminum leakage monitoring based on the visual monitoring results and the sound monitoring results, and generating an alarm signal in abnormal situations. The present application is used to improve the monitoring effect of aluminum leakage in human-machine monitoring.

Description

Technical Field

[0001] The present application relates to the technical field of aluminum processing safety monitoring, and in particular to a method, equipment and storage medium for monitoring aluminum leakage in deep well casting. Background Art

[0002] The aluminum processing industry is a high-risk industry, especially in casting workshops, where explosion risks are often significant. Safety hazards in the aluminum processing industry are often caused by aluminum leaks. These leaks are typically caused by two factors: leaks inherent to the production process itself, which are difficult to prevent, and leaks caused by improper worker behavior during production operations.

[0003] Initially, the aluminum processing industry had a low level of informatization in production safety monitoring. Production safety warnings mainly relied on experience, that is, they mainly relied on manual monitoring. Safety warnings in this way were mostly experience-driven and heavily dependent on human monitoring and judgment.

[0004] In current technology, with the continuous improvement of computer technology, the construction of workshop safety tends to be intelligent. The aluminum processing industry generally implements the method of human-machine joint monitoring. The staff monitors and inspects based on experience, and the monitoring equipment determines whether aluminum leakage occurs through visual processing or sound processing. Specifically,

[0005] Chinese invention patent publication CN115423784A discloses a molten aluminum leak monitoring and alarm system and method based on liquid level fluctuation detection. This system uses a camera to capture images of the molten aluminum surface in a diverter tray and uses the presence of vortices or bubbles on the surface as a basis for leak detection. However, during the actual casting process, a layer of aluminum oxide forms on the surface of the molten aluminum, obscuring any vortices or bubbles, leading to missed or false detections.

[0006] A Chinese invention patent, publication number CN117129154A, discloses a method and system for detecting aluminum leakage in a manifold based on sound recognition. This method uses sensors to collect underwater soundprint features, uses deep learning to learn the soundprint characteristics of aluminum leakage, and then uses these soundprint features to determine if the leakage is occurring. However, this method requires collecting a large number of soundprints for algorithm training before implementation, which is costly.

[0007] Therefore, in current technology, although a dual human-machine monitoring method has been developed, the current intelligent monitoring method generally has certain monitoring loopholes and high usage costs. At the same time, manual monitoring by staff may fail to detect aluminum liquid leakage in time due to unauthorized absence of staff, and the overall aluminum leakage monitoring effect is poor. Summary of the Invention

[0008] In order to improve the monitoring effect of aluminum leakage in human-machine monitoring, the present application provides a deep well casting aluminum leakage monitoring method, equipment and storage medium.

[0009] In the first aspect, the present application provides a method for monitoring aluminum leakage in deep well casting, which adopts the following technical solution:

[0010] A method for monitoring aluminum leakage in deep well casting, comprising the following steps:

[0011] Selecting a visual monitoring target area, and performing visual analysis on the acquired visual features based on the visual monitoring target area to obtain a visual monitoring result, wherein the objects of the visual monitoring result include workers and work objects;

[0012] Acquiring a sound signal and preprocessing the sound signal;

[0013] Performing time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph;

[0014] Calculating the intensity value corresponding to the sound signal and counting the proportion of high-intensity components in the time-frequency graph based on the intensity value, and comparing the proportion of high-intensity components with a preset value to obtain a sound monitoring result;

[0015] Aluminum leakage monitoring is performed based on the visual monitoring results and the sound monitoring results, and an alarm signal is generated in an abnormal situation.

[0016] In some embodiments, the workpiece includes an equipment indicator light, selecting a visual monitoring target area, and performing visual analysis based on the visual monitoring target area to obtain a visual monitoring result include the following steps:

[0017] Set the indicator light ROI area;

[0018] Obtaining a video surveillance stream and reading video frames from the video surveillance stream;

[0019] Cropping the video frame based on the indicator light ROI area to obtain an indicator light classification picture;

[0020] Analyzing the classification results of the indicator light classification pictures, and determining the lighting status of the device indicator lights based on the classification results;

[0021] The current casting state is determined based on the lighting state, and the casting state is used as the visual monitoring result.

[0022] In some embodiments, the worker includes a worker, the workpiece includes a safety helmet, selecting a visual monitoring target area, and performing visual analysis based on the visual monitoring target area to obtain a visual monitoring result include the following steps:

[0023] Select the work area to be inspected;

[0024] Acquire a first detection frame for the worker and a second detection frame for the safety helmet in the video frame based on a visual processing algorithm;

[0025] Calculating a first center point corresponding to the first detection frame and a second center point corresponding to the second detection frame;

[0026] Determining the number of the first center points in the work area to be detected to generate an off-duty monitoring result;

[0027] Determine whether the second center point exists in the first detection frame to generate a safety monitoring result;

[0028] The off-duty monitoring results and the safety monitoring results are both included in the visual monitoring results.

[0029] In some embodiments, obtaining a sound signal and preprocessing the sound signal include the following steps:

[0030] Obtaining a preset number of sound signals at regular intervals;

[0031] The sound signal is divided into frames to generate a multi-frame signal including a plurality of frame signal sets, and the multi-frame signal is windowed.

[0032] In some embodiments, performing time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph comprises the following steps:

[0033] Perform fast Fourier transform on the sound signal to obtain a complex value matrix. The specific formula is:

[0034] D=STFT(x,n_fft,overlap,win_lenght),

[0035] Where D represents the complex matrix of Fourier transform coefficients, n_fft represents the window size, overlap represents the window overlap length, and win_lenght represents the window length.

[0036] The time-frequency diagram is obtained based on the time-frequency analysis result of the complex-valued matrix.

[0037] In some embodiments, calculating the intensity value corresponding to the sound signal and calculating the proportion of high-intensity components in the time-frequency graph based on the intensity value includes the following steps:

[0038] The absolute value of the complex matrix is calculated to obtain an amplitude matrix, and the signal strength decibel value matrix is calculated according to the amplitude matrix. The specific formula is:

[0039] magnitude=|D|,

[0040] value_db=10*np.log10(magnitude+(1e-10)),

[0041] Wherein, magnitude represents the amplitude matrix, value_db represents the signal strength decibel value matrix, and np.log represents the natural logarithm operation;

[0042] The number of high-intensity signals is calculated according to the value of each element in the signal strength decibel value matrix, and the ratio of the number of high-intensity signals to the total number of elements in the signal strength decibel value matrix is calculated to obtain the high-intensity component ratio.

[0043] In some embodiments, calculating the number of high-intensity signals according to the values of each element in the signal strength decibel value includes the following steps:

[0044] Obtaining a preset normal signal strength value, and performing Fourier transform on the normal signal strength value to obtain a normal signal decibel value;

[0045] The elements in the signal strength decibel value matrix that are smaller than the normal signal decibel value are set to 0, and the number of the remaining elements in the signal strength decibel value matrix is defined as the number of the high-intensity signals.

[0046] In some embodiments, comparing the high-intensity component ratio with a preset value to obtain a sound monitoring result includes the following steps:

[0047] If the proportion of the high-strength component is greater than a preset value, the system enters a suspected aluminum leakage state and obtains the current time;

[0048] Taking the current time as a starting point, generating a determination time window including a preset time length;

[0049] When the suspected aluminum leakage state is maintained within the determination time window, the sound monitoring result is determined to be aluminum leakage.

[0050] In a second aspect, the present application provides a deep well casting aluminum leakage monitoring device, which adopts the following technical solution:

[0051] A deep well casting aluminum leakage monitoring device, comprising:

[0052] A visual feature acquisition module, used to acquire visual features;

[0053] a visual monitoring determination module, configured to select a visual monitoring target area and perform visual analysis on visual features based on the visual monitoring target area to obtain a visual monitoring result, wherein the objects of the visual analysis include people and objects, and the visual monitoring result includes any one or a combination of people analysis and object analysis;

[0054] A sound signal acquisition module, used to acquire a sound signal and pre-process the sound signal;

[0055] a sound monitoring and determination module, configured to perform time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph, calculate an intensity value corresponding to the sound signal, and based on the intensity value, calculate a proportion of high-intensity components in the time-frequency graph, and compare the proportion of high-intensity components with a preset value to obtain a sound monitoring result;

[0056] An alarm module is used to monitor aluminum leakage based on the visual monitoring results and the sound monitoring results and generate an alarm signal in abnormal situations.

[0057] In a third aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution:

[0058] A computer-readable storage medium stores a computer program that can be run on a processor, and the computer program implements the above method when executed by the processor.

[0059] The technical solutions provided by the embodiments of this application have the following technical effects:

[0060] Through dual monitoring of vision and sound, it is possible to simultaneously monitor the aluminum leakage problems that have occurred and the potential aluminum leakage hazards caused by workers' improper operations. The monitoring results are more accurate and stable. At the same time, only the time frequency and signal strength ratio of the sound signal are calculated and processed. There is no need for a large amount of aluminum leakage data, which greatly reduces costs and has higher economic value. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a schematic diagram of the steps of a method for monitoring aluminum leakage in deep well casting provided in an embodiment of the present application.

[0062] Figure 2 This is a module connection diagram of a deep well casting aluminum leakage monitoring device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0063] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. However, it should be understood by those skilled in the art that the present application can be implemented without these details. In some cases, in order to avoid unnecessary descriptions that make various aspects of the present application obscure, the well-known methods, processes, systems, components and / or circuits that have been described at a higher level will not be described in detail. It is obvious to those skilled in the art that various changes can be made to the embodiments disclosed in the present application, and the general principles defined in the present application can be applied to other embodiments and application scenarios without departing from the principles and scope of the present application. Therefore, the present application is not limited to the embodiments shown, but conforms to the broadest scope consistent with the scope claimed for protection in the present application.

[0064] It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0065] In the description of this application, "several" means one or more, "many" means more than two, "greater than," "less than," and "exceed" are understood to exclude the number itself, while "above," "below," and "within" are understood to include the number itself. The use of "first" and "second" in the description is solely for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance, implicitly specifying the number of the indicated technical features, or implicitly specifying the order of the indicated technical features.

[0066] In the description of this application, reference to the terms "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any combination in one or more embodiments or examples.

[0067] The embodiments of the present application disclose a method for monitoring aluminum leakage in deep well casting.

[0068] like Figure 1 As shown, a method for monitoring aluminum leakage in deep well casting includes the following steps:

[0069] S100, selecting a visual monitoring target area, and performing visual analysis on the acquired visual features based on the visual monitoring target area to obtain a visual monitoring result.

[0070] Among them, the objects of visual monitoring results include workers and work objects.

[0071] After the system is running, the visual feature acquisition module collects video information in the working environment. The visual feature acquisition module is generally a camera.

[0072] Different visual monitoring target areas are set, and the state of the aluminum processing working environment is determined based on the specific status of the workers and workpieces in the visual monitoring target areas and combined with the corresponding visual processing algorithms.

[0073] Visual monitoring is mainly used to monitor the standardization and safety of the processing process.

[0074] S200: Acquire a sound signal and pre-process the sound signal.

[0075] The voiceprint in the video source collected by the visual feature acquisition module or the voiceprint collected by an independent hydrophone is used as the sound signal.

[0076] The collected sound signals are pre-processed before aluminum leakage monitoring.

[0077] Since the current deep well casting process is divided into water casting and waterless manufacturing according to the water level, when aluminum leakage occurs during the casting process in water, it can be distinguished and judged through the sound pattern detected by the hydrophone or video source.

[0078] S300: Perform time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph.

[0079] S400, calculating the intensity value corresponding to the sound signal and counting the proportion of high-intensity components in the time-frequency diagram based on the intensity value, and comparing the proportion of high-intensity components with a preset value to obtain a sound monitoring result.

[0080] The acquired sound signal is converted into a signal strength matrix through time-frequency processing, and whether aluminum leakage occurs is determined based on the change in sound signal strength during the monitoring period and its proportion in the matrix.

[0081] S500: Aluminum leakage monitoring is performed based on the visual monitoring results and the sound monitoring results, and an alarm signal is generated in an abnormal situation.

[0082] Aluminum leakage monitoring is divided into explicit aluminum leakage monitoring and implicit aluminum leakage monitoring. Explicit aluminum leakage monitoring is characterized by judging whether there is objective aluminum leakage that can be monitored through sound monitoring results. Implicit aluminum leakage monitoring is characterized by judging and restricting the behavior of the working environment through visual monitoring of the safety and standardization of workers and work objects, so as to monitor potential hidden dangers of aluminum leakage that cannot be monitored.

[0083] Then, by synchronizing the visual monitoring results with the sound monitoring results, the aluminum leakage problems and hidden dangers in the monitored area can be effectively monitored, and an alarm signal can be generated in time to remind the relevant personnel when an abnormal situation occurs.

[0084] Through the above method, the dual monitoring of vision and sound can realize the simultaneous monitoring of the aluminum leakage problem and the potential aluminum leakage hidden dangers caused by the workers' improper operation. The monitoring results are more accurate and stable. At the same time, only the time frequency and signal strength ratio of the sound signal are calculated and processed. There is no need for a large amount of aluminum leakage data, which greatly reduces the cost and has higher economic value.

[0085] In other embodiments, the workpiece includes an equipment indicator light, selecting a visual monitoring target area, and performing visual analysis based on the visual monitoring target area to obtain a visual monitoring result include the following steps:

[0086] S110, setting the indicator light ROI area.

[0087] The ROI area is also called the region of interest. In machine vision and image processing, it is an artificially set area that needs to be processed and is outlined in the image being processed in the form of a box, circle, ellipse, irregular polygon, etc.

[0088] In this embodiment, after the specific position of the visual feature acquisition module is determined, the image area captured by the module is obtained, and the corresponding indicator light ROI area is set based on the area where the equipment indicator light is located in the video image. Because casting equipment is large and its position does not move for a long time, once the indicator light ROI area is set, it generally does not need to be revised or changed.

[0089] S111: Obtain a video surveillance stream and read video frames from the video surveillance stream.

[0090] The video frame is obtained from the video surveillance stream collected by the visual feature acquisition module.

[0091] S112: crop the video frame based on the indicator light ROI area to obtain an indicator light classification picture.

[0092] The video frame is cropped to the size corresponding to the indicator light ROI area, and after cropping, a picture containing the indicator light is obtained.

[0093] S113: Analyze the classification results of the indicator light classification pictures, and determine the lighting status of the device indicator lights based on the classification results.

[0094] Use the MobileNetV2 algorithm to perform binary classification on the indicator light image. If the classification result is "light on", it indicates that the device's indicator light is currently on. If the classification result is "light off", it indicates that the device's indicator light is currently off.

[0095] S114, judging the current casting state based on the lighting state, and taking the casting state as the visual monitoring result.

[0096] The casting status is characterized by whether the current monitoring environment is in the casting working state, and the subsequent visual monitoring and sound monitoring are combined to judge the aluminum leakage based on the current casting state. If it is not in the casting state, there is no need for aluminum leakage monitoring.

[0097] In other embodiments, the worker includes a worker, the workpiece includes a safety helmet, selecting a visual monitoring target area, and performing visual analysis based on the visual monitoring target area to obtain a visual monitoring result include the following steps:

[0098] S120, selecting a working area to be inspected.

[0099] The working area to be inspected corresponding to the working environment is artificially selected, generally the area around the casting plate. This range is defined as the working range of the casting staff and the monitoring range of the personnel for aluminum leakage.

[0100] S121: Obtain a first detection frame for the worker and a second detection frame for the helmet in the video frame based on a visual processing algorithm.

[0101] The number of categories of the YOLOv5 algorithm is set to 2 to monitor pedestrians (staff) and helmets in the video frame, thereby obtaining the first detection box person_bbox containing the staff and the second detection box helmet_bbox containing the helmet.

[0102] S122: Calculate a first center point corresponding to the first detection frame and a second center point corresponding to the second detection frame.

[0103] The center point position of each detection frame is calculated based on each first detection frame and the second detection frame.

[0104] S123, determining the number of first center points in the work area to be detected to generate an off-post monitoring result.

[0105] First, it is determined whether there is a first center point in the working area to be detected. If so, it indicates that there are workers in the working range of the casting. Then, the number of first center points in the working area to be detected is detected, which represents the number of workers around the casting plate.

[0106] Set the number of staff members in the work area to be detected according to the requirements. For example, if it is set to 2, then when there are more than two first center points in the work area to be detected, it is considered that the work range is not in an off-duty state. When the number of first center points in the work area to be detected is less than 2, it is considered that there is an off-duty state, and an alarm is required.

[0107] S124: Determine whether there is a second center point in the first detection frame to generate a safety monitoring result.

[0108] Secondly, when there is a second center point in the first detection frame, it is considered that the staff is wearing a safety helmet when working. Conversely, if there is no second center point in the first detection frame, it is considered that the staff is not wearing a safety helmet when working, and an alarm signal needs to be issued.

[0109] Among them, off-duty monitoring results and safety monitoring results are both included in the visual monitoring results.

[0110] The above monitoring results are used to monitor the behavior of staff. When no alarm signal appears in the behavior of staff, it is considered that there is no safety hazard of aluminum leakage that cannot be monitored by staff in time, and aluminum leakage monitoring based on sound monitoring can be carried out.

[0111] In some other embodiments, obtaining a sound signal and preprocessing the sound signal include the following steps:

[0112] S210: Obtain a preset number of sound signals at a fixed time.

[0113] 200,000 digital signals are acquired per second from a video source or hydrophone as sound signals.

[0114] S220 , dividing the sound signal into frames to generate a multi-frame signal including a plurality of frame signal sets, and performing windowing on the multi-frame signal.

[0115] Framing the digital signal to divide it into multiple frame signal sets, specifically:

[0116] x→M=[x1,x2,...,x n-1 ,x n ],

[0117] n=(N-overlap) / inc,

[0118] overlap=wlen-inc.

[0119] Among them, overlap represents the overlapping part; wlen represents the frame length, which is 5000; inc represents the frame shift, which is 2500; n represents the number of frames of the entire signal; N represents the length of the entire signal, which is 200000; M represents the set of multi-frame signals.

[0120] When performing a Fast Fourier Transform (FFT) on a signal, if the selected frame is not an integer multiple of the signal period, sample discontinuities will occur when the period is expanded, causing subsequent spectrum distortion. Theoretically, the spectrum only contains the frequencies of the signal being measured. However, when the spectrum is distorted, it contains numerous frequency components, resulting in a spectrum leakage effect.

[0121] Therefore, in order to solve the above phenomenon, it is necessary to window the multi-frame signal after framing so that the left and right ends of the windowed signal are connected to make it more periodic.

[0122] In the embodiment of the present application, the multi-frame signal is windowed by a Hamming window, and the specific formula is as follows:

[0123]

[0124] Among them, x i It represents the signal of each frame input; n represents the number of frames of the entire signal; a is the coefficient, which is 0.46; W i The representation is the signal after windowing each frame.

[0125] In some other embodiments, performing time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph includes the following steps:

[0126] S310 , performing fast Fourier transform on the sound signal to obtain a complex-valued matrix.

[0127] Perform fast Fourier transform on the above frame signal frame by frame. The specific formula is:

[0128] D=STFT(x,n_fft,overlap,win_lenght).

[0129] Where D represents the complex matrix of Fourier transform coefficients; n_fft represents the window size, which is 5000; overlap represents the window overlap length, which is 2500; and win_length represents the window length, which is 5000.

[0130] S320: Obtain a time-frequency diagram based on the time-frequency analysis result of the complex-valued matrix.

[0131] The fast Fourier transform of the sound signal is used to perform time-frequency analysis to obtain an image containing time domain and frequency domain, where the x-axis is the time series and the y-axis is the frequency range.

[0132] The amplitude and magnitude of the sound signal can be represented by complex-valued matrices and time-frequency diagrams.

[0133] In some other embodiments, calculating the intensity value corresponding to the sound signal and counting the proportion of high-intensity components in the time-frequency graph based on the intensity value includes the following steps:

[0134] S410, calculating the absolute value of the complex-valued matrix to obtain an amplitude matrix, and calculating the signal strength decibel value matrix based on the amplitude matrix. The specific formula is:

[0135] magnitude=|D|,

[0136] value_db=10*np.log10(magnitude+(1e-10)).

[0137] Among them, magnitude represents the amplitude matrix, value_db represents the signal strength decibel value matrix, and np.log represents the natural logarithm operation.

[0138] The amplitude is represented by the intensity of the frequency domain signal. After the sound signal undergoes Fourier transform, the size of each frequency component in the spectrum can be regarded as the modulus of the complex-valued spectrum. Therefore, the amplitude matrix is obtained by performing an absolute value operation on the complex-valued matrix.

[0139] A signal strength decibel value matrix is obtained by performing a logarithmic operation on the amplitude matrix, wherein each element represents the sound intensity corresponding to each frame signal.

[0140] S420 , calculating the number of high-intensity signals according to the value of each element in the signal strength decibel value matrix, and obtaining the proportion of high-intensity components according to the ratio of the number of high-intensity signals to the total number of elements in the signal strength decibel value matrix.

[0141] In the embodiment of the present application, high intensity is set to 1000. When the signal intensity is 1000, the corresponding decibel value is 30.

[0142] When aluminum leakage occurs, the decibel value corresponding to the monitored sound will increase. When it is greater than the preset value, it is considered that there is suspected aluminum leakage. Then, a high-intensity component ratio can be obtained by the ratio of the number of high-intensity signals to the number in the signal intensity decibel value matrix.

[0143] The specific formula is:

[0144] M=positive_count / L.

[0145] L represents the numerical size of the frame signal of the complex-valued matrix, which is 2501*79.

[0146] In some other embodiments, calculating the number of high-intensity signals based on the values of each element in the signal strength decibel value includes the following steps:

[0147] S421: Obtain a preset normal signal strength value, and perform Fourier transform on the normal signal strength value to obtain a normal signal decibel value.

[0148] S422: Set the elements in the signal strength decibel value matrix that are smaller than the normal signal decibel value to 0, and define the number of remaining elements in the signal strength decibel value matrix as the number of high-intensity signals.

[0149] When determining the number of high-intensity signals, we can first set the elements in the matrix that do not belong to high-intensity signals to zero. That is, elements with signal strengths below 1000, whose corresponding decibel values are below 30db, are not considered to have aluminum leakage. All non-zero elements in the signal strength decibel value matrix are characterized as high-intensity signals.

[0150] In other embodiments, comparing the high-intensity component ratio with a preset value to obtain a sound monitoring result includes the following steps:

[0151] S430: If the proportion of high-strength components is greater than a preset value, the system enters a suspected aluminum leakage state and obtains the current time.

[0152] S440 , taking the current time as a starting point, generating a determination time window including a preset time length.

[0153] S450: When the suspected aluminum leakage state is maintained within the determination time window, the sound monitoring result is determined to be aluminum leakage.

[0154] If the proportion of high-intensity signals in the signal strength decibel value matrix (M) is greater than 0.3, it is determined that aluminum leakage is currently suspected. At the same time, when the suspected aluminum leakage occurs, the current time is obtained and a 3-second duration is selected as the judgment time window. If the suspected aluminum leakage state persists for three seconds, aluminum leakage is confirmed and the corresponding alarm information is output.

[0155] like Figure 2 As shown, the embodiment of the present application also discloses a deep well casting aluminum leakage monitoring device, comprising:

[0156] The visual feature acquisition module is used to acquire visual features.

[0157] The visual monitoring judgment module is used to select a visual monitoring target area and perform visual analysis on visual features based on the visual monitoring target area to obtain visual monitoring results, wherein the objects of visual analysis include people and objects, and the visual monitoring results include any one or a combination of people analysis and object analysis.

[0158] The sound signal acquisition module is used to acquire the sound signal and pre-process the sound signal.

[0159] The sound monitoring and judgment module is used to calculate the time-frequency processing of the pre-processed sound signal to obtain a time-frequency diagram, calculate the intensity value corresponding to the sound signal, and based on the intensity value, count the proportion of high-intensity components in the time-frequency diagram, and compare the proportion of high-intensity components with the preset value to obtain the sound monitoring result.

[0160] The alarm module is used to monitor aluminum leakage based on visual monitoring results and sound monitoring results and generate alarm signals in abnormal situations.

[0161] In an embodiment of the present application, a computer-readable storage medium is also disclosed, which stores a computer program that can be run on a processor. When the computer program is executed by the processor, the steps of the above method are implemented.

[0162] The implementation principle is:

[0163] Through dual monitoring of vision and sound, it is possible to simultaneously monitor the aluminum leakage problems that have occurred and the potential aluminum leakage hazards caused by workers' improper operations. The monitoring results are more accurate and stable. At the same time, only the time frequency and signal strength ratio of the sound signal are calculated and processed. There is no need for a large amount of aluminum leakage data, which greatly reduces costs and has higher economic value.

[0164] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps and they may be performed in other orders.

[0165] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for monitoring aluminum leakage in deep well casting, characterized in that: The following steps are involved: Selecting a visual monitoring target area, and performing visual analysis on the acquired visual features based on the visual monitoring target area to obtain a visual monitoring result, wherein the objects of the visual monitoring result include workers and work objects, wherein the workers include staff members, and the work objects include safety helmets and equipment indicator lights; Selecting a work area to be inspected; obtaining a first detection frame for the worker and a second detection frame for the safety helmet in a video frame based on a visual processing algorithm; calculating a first center point corresponding to the first detection frame and a second center point corresponding to the second detection frame; determining the number of the first center points in the work area to be inspected to generate an off-post monitoring result; determining whether the second center point exists in the first detection frame to generate a safety monitoring result; both the off-post monitoring result and the safety monitoring result are included in the visual monitoring result; Setting an indicator light ROI area; acquiring a video surveillance stream and reading a video frame from the video surveillance stream; cropping the video frame based on the indicator light ROI area to obtain an indicator light classification image; analyzing a classification result of the indicator light classification image, and determining the lighting state of the equipment indicator light based on the classification result; determining a current casting state based on the lighting state, and using the casting state as the visual monitoring result; Acquiring a sound signal and preprocessing the sound signal; Performing time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph; Calculating the intensity value corresponding to the sound signal and counting the proportion of high-intensity components in the time-frequency graph based on the intensity value, and comparing the proportion of high-intensity components with a preset value to obtain a sound monitoring result; Aluminum leakage monitoring is performed based on the visual monitoring results and the sound monitoring results, and an alarm signal is generated in an abnormal situation.

2. The deep well casting aluminum leakage monitoring method according to claim 1 is characterized in that: Acquiring a sound signal and preprocessing the sound signal comprises the following steps: Obtaining a preset number of sound signals at regular intervals; The sound signal is divided into frames to generate a multi-frame signal including a plurality of frame signal sets, and the multi-frame signal is windowed.

3. The method for monitoring aluminum leakage in deep well casting according to claim 2, characterized in that: Performing time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph comprises the following steps: Perform fast Fourier transform on the sound signal to obtain a complex value matrix. The specific formula is: D=STFT(x,n_fft,overlap,win_lenght), Wherein, x represents the sound signal, D represents the complex matrix of Fourier transform coefficients, n_fft represents the window size, overlap represents the window overlap length, and win_length represents the window length; The time-frequency diagram is obtained based on the time-frequency analysis result of the complex-valued matrix.

4. The method for monitoring aluminum leakage in deep well casting according to claim 3, characterized in that: Calculating the intensity value corresponding to the sound signal and counting the proportion of high-intensity components in the time-frequency graph based on the intensity value, including the following steps: The absolute value of the complex-valued matrix is calculated to obtain an amplitude matrix, and the signal strength decibel value matrix is calculated based on the amplitude matrix. The specific formula is: magnitude=|D|, value_db=10*np.log10(magnitude+(1e-10)), Wherein, magnitude represents the amplitude matrix, value_db represents the signal strength decibel value matrix, and np.log represents the natural logarithm operation; The number of high-intensity signals is calculated according to the value of each element in the signal strength decibel value matrix, and the ratio of the number of high-intensity signals to the total number of elements in the signal strength decibel value matrix is calculated to obtain the high-intensity component ratio.

5. The method for monitoring aluminum leakage in deep well casting according to claim 4, characterized in that: Calculating the number of high-intensity signals according to the numerical value of each element in the signal strength decibel value includes the following steps: Obtaining a preset normal signal strength value, and performing Fourier transform on the normal signal strength value to obtain a normal signal decibel value; The elements in the signal strength decibel value matrix that are smaller than the normal signal decibel value are set to 0, and the number of the remaining elements in the signal strength decibel value matrix is defined as the number of the high-intensity signals.

6. The method for monitoring aluminum leakage in deep well casting according to claim 4, characterized in that: Comparing the high-intensity component ratio with a preset value to obtain a sound monitoring result includes the following steps: If the proportion of the high-strength component is greater than a preset value, the system enters a suspected aluminum leakage state and obtains the current time; Taking the current time as a starting point, generating a determination time window including a preset time length; When the suspected aluminum leakage state is maintained within the determination time window, the sound monitoring result is determined to be aluminum leakage.

7. A deep well casting aluminum leakage monitoring device, applied to the deep well casting aluminum leakage monitoring method according to any one of claims 1 to 6, characterized in that: include: A visual feature acquisition module, used to acquire visual features; A visual monitoring determination module is used to select a visual monitoring target area and perform visual analysis on visual features based on the visual monitoring target area to obtain a visual monitoring result. A sound signal acquisition module, used to acquire a sound signal and pre-process the sound signal; a sound monitoring and determination module, configured to perform time-frequency processing on the pre-processed sound signal to obtain a time-frequency graph, calculate an intensity value corresponding to the sound signal, and based on the intensity value, calculate a proportion of high-intensity components in the time-frequency graph, and compare the proportion of high-intensity components with a preset value to obtain a sound monitoring result; An alarm module is used to monitor aluminum leakage based on the visual monitoring results and the sound monitoring results and generate an alarm signal in abnormal situations.

8. A computer-readable storage medium, characterized in that A computer program that can be run on a processor is stored thereon, and when the computer program is executed by the processor, the method according to any one of claims 1 to 6 is implemented.

Images