Infrared lamp bead anomaly detection device and method for monitor

Abstract

The invention provides an infrared lamp bead anomaly detection device and method for a monitor. The method comprises the following steps: S101, acquiring image data containing infrared lamp beads of the monitor; s102, carrying out image optimization on the image data; s103, segmenting the image data into a plurality of segmented images based on the position information; s104, carrying out edge contour and physical damage identification operation; carrying out abnormal marking on the infrared lamp beads which are physically damaged; fourier transform is carried out on the segmented image of which the physical damage is not recognized, frequency domain information is input into a pre-trained anomaly recognition model, a frequency domain waveform is analyzed, and anomaly marking is carried out on an abnormal infrared lamp bead; s105, outputting a detection result and sending the detection result to a detection center; therefore, when the lamp beads of the monitor are abnormal, specific abnormity can be positioned to take targeted repair measures, and the problem that a traditional lamp bead abnormity detection method can only judge whether the lamp beads have defects or not is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of lamp bead detection, and more specifically, particularly relates to an infrared lamp bead abnormality detection device and method for a monitor. Background Art

[0002] From important places with strict security to various commercial office areas, monitors always protect people's lives and property safety and social stability and order; infrared lamp beads are components that realize the infrared night vision function of the monitor, allowing the monitor to still work normally at night or in low light environments.

[0003] When an abnormality occurs in the infrared lamp beads, the infrared night vision function of the monitor will be greatly reduced, the monitoring screen will become blurred and the noise will increase, and the traditional manual detection method requires a lot of manpower and time, resulting in low efficiency; such as the Chinese invention patent with patent number 202310765566.2, which provides a lamp bead defect detection method and system, which collects images of the lamp beads to be detected through a camera, and then performs image analysis on the image to obtain a global correlation feature vector for lamp bead detection, and finally, based on the global correlation feature vector for lamp bead detection, determines whether the lamp bead to be detected has defects; however, the traditional lamp bead abnormality detection method is only used to determine whether the lamp bead has defects, but cannot analyze and identify the specific abnormal conditions of the lamp beads, so that the monitor cannot quickly take countermeasures when lamp bead abnormalities occur. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an infrared lamp bead abnormality detection device and method for a monitor, so as to solve the technical problem in the prior art that the traditional lamp bead abnormality detection method is only used to determine whether the lamp bead has defects, but cannot analyze and identify the specific abnormal conditions of the lamp bead, making it impossible for the monitor to quickly take countermeasures when the lamp bead abnormality occurs.

[0005] The purpose and effect of the infrared lamp bead abnormality detection device and method for monitoring device of the present invention are achieved by the following specific technical means: A method for detecting abnormality of infrared lamp beads for a monitor includes the following steps: S101: placing a monitor in a closed detection box, continuously photographing the infrared lamp beads in the monitor based on an image acquisition device to obtain image data containing the infrared lamp beads of the monitor, and the image acquisition device is installed at a position where the light-emitting state of the infrared lamp beads can be captured; S102: using median filtering to remove salt and pepper noise in the image data, and using an unsharp mask algorithm to sharpen the image data to highlight the edge features of the infrared lamp beads; S103: performing matching based on the information database of the monitor, determining the position information of each infrared lamp bead on the monitor, and dividing the image data into multiple segmented images based on the position information of each infrared lamp bead on the monitor; S104: Identify edge contour information of the infrared lamp beads in the segmented image, and determine whether the infrared lamp beads are physically damaged based on the edge contour information; The infrared lamp beads that are identified to be physically damaged will be marked as abnormal; The segmented images without physical damage are Fourier transformed, and the converted frequency domain information is input into the pre-trained abnormality recognition model. The abnormality recognition model identifies the abnormal information of the lamp beads by analyzing the frequency domain waveform characteristics, and marks the abnormal infrared lamp beads as abnormal. S105: Output the test results and send the test results to the testing center; Test results include: The location information of abnormal infrared lamp beads is marked in the original image information, and the specific location of the abnormal lamp beads on the monitor is displayed in the form of coordinate positioning; Abnormal type of abnormal infrared lamp beads, including physical damage and abnormal information; Generate an exception report, which includes the number of the abnormal lamp bead in the monitor information database, location coordinates, abnormality type, abnormality severity assessment and possible repair suggestions.

[0006] As a further solution of the present invention, in S103, dividing the image data into a plurality of divided images includes: The image data is laid out as a three-dimensional matrix. Based on the position information of each infrared lamp bead on the monitor, each infrared lamp bead is segmented into a cylinder. Then, the segmented cylindrical image is layered and cut in three dimensions to obtain a multi-layer two-dimensional slice image of each infrared lamp bead.

[0007] As a further solution of the present invention, the three-dimensional matrix is ​​expressed as: ; Where x and y represent two-dimensional coordinates, and z represents the depth coordinate of the layered cutting.

[0008] As a further solution of the present invention, the cylinder segmentation formula for performing cylinder segmentation on each infrared lamp bead is: ; In the formula, Represents the center coordinates of the lamp bead, r represents the radius of the cylinder, which is dynamically adjusted according to the actual size and distribution density of the infrared lamp bead.

[0009] As a further solution of the present invention, in S104, identifying the edge contour information of the infrared lamp beads in the segmented image includes: The edge contour information of the infrared lamp beads in each two-dimensional slice image is identified by the edge algorithm. The edge algorithm formula is: ; In the formula, represents the gradient of the image in the x direction, Represents the gradient of the image in the y direction; Based on the edge profile information, determine whether the infrared lamp beads are physically damaged.

[0010] As a further aspect of the present invention, physical damage includes external damage and internal damage; The surface damage is based on the curvature change of multiple edge contour information. If the curvature change exceeds a preset threshold, it is determined as surface damage; Internal damage is based on calculating the mean and standard deviation of the pixel values ​​in the internal area of ​​the edge contour. If the mean pixel value is lower than the preset threshold or the standard deviation exceeds the preset threshold, it is determined to be internal damage.

[0011] As a further solution of the present invention, Fourier transform includes: The two-dimensional slice image without physical damage is Fourier transformed. The Fourier transform formula is: ; In the formula, represents a two-dimensional slice image, Represents frequency domain information; The converted frequency domain information is input into the pre-trained anomaly recognition model to identify the abnormal information of the lamp beads.

[0012] As a further solution of the present invention, the abnormal information includes brightness abnormality, no light abnormality and flickering abnormality, and the abnormal information of identifying the lamp beads specifically includes: Abnormal brightness recognition: By analyzing the amplitude changes of the low-frequency components in the frequency domain waveform, it is determined whether the brightness is abnormal. The abnormal brightness recognition formula is: ; In the formula, represents the sum of the amplitudes of the low-frequency components, and Indicates the low frequency range; Abnormal identification of no light: By detecting whether the amplitude of all frequency components in the frequency domain waveform is close to zero, it is judged whether the lamp bead is not emitting light. The abnormal identification formula of no light is: ; In the formula, represents the total amplitude; Flicker anomaly identification: By analyzing the amplitude changes of high-frequency components in the frequency domain waveform, it is determined whether the lamp bead flickers. The flicker anomaly identification formula is: ; In the formula, represents the sum of the amplitudes of high-frequency components, and Indicates the high frequency range.

[0013] As a further solution of the present invention, the anomaly recognition model is a deep learning model based on a convolutional neural network, and its training process includes: Collect image data of multiple normal and abnormal infrared lamp beads as training sets; Label the training set, including the location of the infrared lamp beads and the type of anomaly; Use the labeled training set to train the anomaly recognition model and optimize the model parameters.

[0014] An infrared lamp bead abnormality detection device for a monitor includes: The image acquisition unit is used to acquire the image data of the infrared lamp beads of the monitor and transmit it to the image processing unit; the image acquisition unit includes a detection box and an image acquisition device, the detection box is used to place the monitor to be detected, and its interior can shield external light interference and electromagnetic interference to ensure a stable shooting environment; the image acquisition device is installed in a position where it can capture the luminous state of the infrared lamp beads, and continuously shoots the infrared lamp beads in the monitor; The image processing unit includes an image optimization module and an image segmentation module. The image optimization module uses a median filter to remove salt and pepper noise in the image data, and uses an unsharp mask algorithm to sharpen the image data to highlight the edge features of the infrared lamp beads. The image segmentation module determines the position information of each infrared lamp bead on the monitor based on the information database of the monitor, and divides the image data into multiple segmented images and transmits them to the abnormality detection unit. The abnormality detection unit includes an edge recognition module and an abnormality recognition module. The edge recognition module is used to recognize the edge contour information of the infrared lamp beads in the segmented image, and judge whether the infrared lamp beads are physically damaged based on the edge contour information, and mark the infrared lamp beads that are physically damaged as abnormal; the abnormality recognition module performs Fourier transform on the segmented image that is not physically damaged, and then inputs the converted frequency domain information into a pre-trained abnormality recognition model, and recognizes the abnormal information of the lamp beads by analyzing the frequency domain waveform characteristics, including brightness abnormality, non-luminous abnormality and flickering abnormality, and marks the infrared lamp beads that are abnormal as abnormal; The output unit is used to output the test results and send them to the test center, mark the location information of the abnormal infrared lamp beads in the original image, display the specific location of the abnormal lamp beads on the monitor by coordinate positioning, and generate an abnormality report, which includes the number of the abnormal lamp beads, location coordinates, abnormality type, abnormality severity assessment and possible repair suggestions; The storage unit is used to store the information database of the monitor, the parameters of the abnormality recognition model and the judgment criteria of different abnormality types; the storage unit includes the location information database of infrared lamp beads, the frequency domain waveform feature threshold, the physical damage judgment criteria and the abnormality type judgment criteria, so that each unit can call it during operation.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. The image quality is optimized by applying median filtering and unsharp masking algorithms to image data. Edge detection and frequency domain analysis are used to comprehensively detect various abnormal conditions of infrared lamp beads, including physical damage (external damage and internal damage) and abnormal working conditions (brightness abnormality, non-luminous abnormality and flickering abnormality). This solves the problem that traditional lamp bead abnormality detection methods can usually only determine whether the lamp beads are defective. This method can identify specific abnormality types by performing multi-level analysis on the edge contour information and frequency domain waveform characteristics of infrared lamp beads. It can determine whether it is external damage or internal damage by calculating the mean and standard deviation of pixel values ​​in the internal area of ​​the edge contour and analyzing the curvature change of the edge contour information. For lamp beads that have no physical damage, brightness abnormality, non-luminous abnormality and flickering abnormality are identified through Fourier transform and frequency domain analysis, so that the monitor can quickly locate the problem and take targeted repair measures when lamp bead abnormalities occur.

[0016] 2. When the monitor shows abnormal lamp beads, countermeasures can be taken quickly; in the output link of the detection results, the location information of the abnormal infrared lamp beads is marked in the original image, and its specific location on the monitor is displayed in the form of coordinate positioning. At the same time, the generated abnormal report contains the number of the abnormal lamp beads, location coordinates, abnormality type, abnormality severity assessment and possible repair suggestions, which enables the staff to obtain detailed abnormality information at the first time, quickly locate the problem lamp beads, and make repairs based on the severity of the abnormality and repair suggestions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of the steps of a method for detecting abnormality of infrared lamp beads for a monitor of the present invention; Figure 2 The present invention is a flowchart of the steps of abnormal classification in a method for detecting abnormalities of infrared lamp beads for a monitor. DETAILED DESCRIPTION

[0018] The following is a further detailed description of the embodiments of the present invention in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the technical solutions of the present invention, but cannot be used to limit the scope of protection of the present invention.

[0019] Example: Please refer to Figure 1 and Figure 2As shown, the present invention provides a method for detecting abnormalities of infrared lamp beads for a monitor, which is characterized by comprising the following steps:

[0020] S101: Place the monitor in a closed detection box, which has good sealing performance, and the inner wall of the detection box is provided with light-absorbing materials and electromagnetic shielding materials, which can shield the interference of external light and electromagnetic interference, create a stable environment for subsequent image acquisition work, prevent external interference, and avoid the reflection of natural light on the surface of infrared lamp beads, resulting in inaccurate image data; based on the image acquisition device, the infrared lamp beads in the monitor are continuously photographed, and the image acquisition device is installed in a position where the luminous state of the infrared lamp beads can be clearly and comprehensively captured to ensure that the acquired image data can accurately reflect the actual luminous state of the infrared lamp beads; during the shooting process, the image acquisition device shoots at a stable frame rate to ensure the continuity and integrity of the image data, collects the image data of the infrared lamp beads, and transmits the image data to the image optimization module, providing a reliable data basis for subsequent abnormality detection.

[0021] S102: Optimize the collected image data. First, use median filtering to process the acquired image data to remove the salt and pepper noise. The median filtering sorts the pixel values ​​in the neighborhood of each pixel in the image and selects the middle value as the new value of the pixel, thereby eliminating the salt and pepper noise and retaining the image detail information as much as possible. After removing the salt and pepper noise, use the unsharp mask algorithm to sharpen the image data. The unsharp mask algorithm enhances the edge and details of the image and highlights the edge features of the infrared lamp beads, making the outline of the infrared lamp beads clearer, facilitating the subsequent recognition and analysis of its edge contour information, and improving the accuracy. The sharpened image data is transmitted to the image segmentation module.

[0022] S103: performing a matching operation based on the information database of the monitor, the information database is stored in the storage unit, and the database contains detailed position information of infrared lamp beads on various models of monitors. The specific position information of each infrared lamp bead on the current monitor to be detected is determined through the matching operation. Based on the position information, the image data is segmented into multiple segmented images through the image segmentation module, so as to independently analyze and detect each infrared lamp bead; Furthermore, the specific operation of segmenting the image data into multiple segmented images is as follows: First, the image data is laid out as a three-dimensional matrix. This three-dimensional matrix representation can more comprehensively reflect the spatial information of the image data. Among them, x and y represent two-dimensional coordinates, which are used to determine the position on the image plane; z represents the depth coordinate of layered cutting, which is used to represent different image layers. Based on the position information of each infrared lamp bead on the monitor, each infrared lamp bead is cylindrically segmented, which can accurately separate each infrared lamp bead from the image. Then, the segmented cylindrical image is layered and cut in three dimensions to obtain multi-layer two-dimensional slice images of each infrared lamp bead, and these two-dimensional slice images are transmitted to the edge recognition module. These two-dimensional slice images can provide more detailed infrared lamp bead information and provide data for subsequent abnormality detection. Specifically, the three-dimensional matrix is ​​expressed as: ; Where x and y represent two-dimensional coordinates, and z represents the depth coordinate of the layered cutting.

[0023] The cylindrical segmentation formula for each infrared lamp bead is: ; In the formula, Represents the center coordinates of the lamp bead, r represents the radius of the cylinder. The radius r of the cylinder is not a fixed value, but is dynamically adjusted according to the actual size and distribution density of the infrared lamp bead. When the infrared lamp bead is large or sparsely distributed, the radius of the cylinder is appropriately increased; otherwise, the radius of the cylinder is reduced to ensure that each infrared lamp bead can be accurately segmented.

[0024] S104: Identify edge contour information of infrared lamp beads in the segmented image to find possible problems in the appearance and internal structure of the infrared lamp beads; based on the edge contour information, determine whether the infrared lamp beads are physically damaged, and mark the infrared lamp beads that are identified to be physically damaged as abnormal. The marking content includes information such as the location of the lamp beads and possible damage types, so as to facilitate subsequent processing and analysis; Furthermore, the specific method for identifying the edge contour information of the infrared lamp beads in the segmented image is as follows: The edge recognition module uses the edge algorithm to identify the edge contour information of the infrared lamp beads in each two-dimensional slice image. The edge algorithm uses the gradient changes of the image in different directions to determine the position of the edge. The specific edge algorithm formula is: ; In the formula, represents the gradient of the image in the x direction, Represents the gradient of the image in the y direction; by calculating the amplitude of the gradient, the edge position of the infrared lamp bead can be accurately determined, and based on this edge contour information, it can be judged whether the infrared lamp bead is physically damaged.

[0025] Specifically, physical damage includes external damage and internal damage: The appearance damage is judged based on the curvature change of multiple edge profile information. The curvature change reflects the bending degree of the edge profile. If the curvature change exceeds the preset threshold, it means that the appearance of the infrared lamp bead may be damaged, such as the shell is cracked or deformed. Internal damage is judged based on the mean and standard deviation of the pixel values ​​in the inner area of ​​the edge contour. The mean pixel value reflects the average brightness of the area, and the standard deviation reflects the discrete degree of the pixel value. If the mean pixel value is lower than the preset threshold or the standard deviation exceeds the preset threshold, it means that there may be problems inside the infrared lamp bead, such as damage to the light-emitting chip, loss of internal components, deformation of internal components, etc. Perform Fourier transform on the segmented image without physical damage to separate the different frequency components in the image. Through Fourier transform, the frequency domain characteristics of the image can be more conveniently analyzed to identify possible abnormal information of the infrared lamp beads. The converted frequency domain information is then input into the pre-trained abnormality recognition model, which is a deep learning model based on convolutional neural network. The training process of the anomaly recognition model includes: First, we collect image data of multiple normal and abnormal infrared lamp beads as training sets. These image data cover various types of abnormal situations to ensure that the model can learn rich feature information. Then, we annotate the training set, including the location and abnormal type of the infrared lamp beads. The annotation process requires manual analysis and judgment of each image to ensure the accuracy of the annotation information. Finally, we use the annotated training set to train the abnormality recognition model. During the training process, we continuously adjust the parameters of the model so that the model can better identify and classify the abnormal information of the infrared lamp beads. The anomaly recognition model identifies the abnormal information of the lamp beads by analyzing the frequency domain waveform characteristics, and marks the abnormal infrared lamp beads.

[0026] Furthermore, Fourier transform includes: The two-dimensional slice image without physical damage is Fourier transformed. The Fourier transform formula is: ; In the formula, represents a two-dimensional slice image, Represents frequency domain information; through this formula, the two-dimensional slice image can be converted from the spatial domain to the frequency domain, and the converted frequency domain information is input into the pre-trained abnormality recognition model to identify the abnormal information of the lamp beads; Specifically, the abnormal information includes abnormal brightness, abnormal non-luminescence and abnormal flickering. The abnormal information of identifying lamp beads specifically includes: Abnormal brightness recognition: By analyzing the amplitude changes of low-frequency components in the frequency domain waveform, it is determined whether the brightness is abnormal. The low-frequency components mainly reflect the overall brightness information of the image. The abnormal brightness recognition formula is: ; In the formula, represents the sum of the amplitudes of the low-frequency components, and Indicates the low-frequency range; by calculating the sum of the amplitudes of the low-frequency components and comparing it with the amplitude under normal conditions, it can be determined whether the brightness of the infrared lamp beads is abnormal; Abnormal identification of no light: By detecting whether the amplitude of all frequency components in the frequency domain waveform is close to zero, it is judged whether the lamp bead is not emitting light. If the amplitude of all frequency components is close to zero, it means that the infrared lamp bead does not emit light; the abnormal identification formula of no light is: ; In the formula, Indicates the total amplitude; when the total amplitude is close to zero, it can be determined that the infrared lamp beads are not emitting light.

[0027] Flicker anomaly recognition: By analyzing the amplitude changes of high-frequency components in the frequency domain waveform, it is determined whether the lamp beads are flickering. The high-frequency components mainly reflect the rapid changes in the image. The flicker phenomenon will cause the amplitude of the high-frequency components to increase. The flicker anomaly recognition formula is: ; In the formula, represents the sum of the amplitudes of high-frequency components, and Indicates the high-frequency range; by calculating the sum of the amplitudes of the high-frequency components and comparing it with the amplitude under normal conditions, it can be determined whether the infrared lamp beads have abnormal flickering.

[0028] S105: Output the test results and send them to the test center; the test center can handle the abnormal conditions of the infrared lamp beads of the monitor according to these test results; Test results include: The location information of abnormal infrared lamp beads is marked in the original image information, and the specific location of the abnormal lamp beads on the monitor is displayed in the form of coordinate positioning. This coordinate positioning method enables the staff to quickly and accurately find the location of the abnormal lamp beads and improve the processing efficiency; The abnormal types of abnormal infrared lamp beads are listed in detail, including physical damage and abnormal information. The types of physical damage include external damage and internal damage; the types of abnormal information include brightness abnormality, non-luminous abnormality, and flickering abnormality. Clear abnormal type information helps staff to develop targeted treatment plans.

[0029] Generate an abnormality report, which is a comprehensive summary and analysis of the test results; the abnormality report includes the number of the abnormal lamp bead in the monitor information database, the location coordinates, the abnormality type, the abnormality severity assessment and possible repair suggestions. The abnormality severity assessment can be comprehensively judged according to the type and manifestation of the abnormality, and is divided into different levels such as mild, moderate and severe. Possible repair suggestions are provided according to the type and severity of the abnormality, such as replacing the lamp bead, repairing the line, etc., to provide specific operational guidance for the staff.

[0030] An infrared lamp bead abnormality detection device for a monitor includes: an image acquisition unit, an image processing unit, an abnormality detection unit, an output unit and a storage unit; The image acquisition unit is used to acquire the image data of the infrared lamp beads of the monitor and transmit it to the image processing unit; the image acquisition unit includes a detection box and an image acquisition device, the detection box is used to place the monitor to be detected, and the detection box is provided with shielding materials and light-absorbing materials that can shield external light interference and electromagnetic interference. These materials can reduce external ambient light interference and electromagnetic interference to ensure a stable shooting environment; the image acquisition device is installed in a position where it can capture the luminous state of the infrared lamp beads, and continuously shoots the infrared lamp beads in the monitor; The image processing unit includes an image optimization module and an image segmentation module. The image optimization module uses a median filter to remove salt and pepper noise in the image data, and uses an unsharp mask algorithm to sharpen the image data to highlight the edge features of the infrared lamp beads. The image segmentation module determines the position information of each infrared lamp bead on the monitor based on the information database of the monitor, and divides the image data into multiple segmented images and transmits them to the abnormality detection unit. The abnormality detection unit includes an edge recognition module and an abnormality recognition module. The edge recognition module is used to recognize the edge contour information of the infrared lamp beads in the segmented image, and judge whether the infrared lamp beads are physically damaged based on the edge contour information, and mark the infrared lamp beads that are physically damaged as abnormal; the abnormality recognition module performs Fourier transform on the segmented image that is not physically damaged, and then inputs the converted frequency domain information into a pre-trained abnormality recognition model, and recognizes the abnormal information of the lamp beads by analyzing the frequency domain waveform characteristics, including brightness abnormality, non-luminous abnormality and flickering abnormality, and marks the infrared lamp beads that are abnormal as abnormal; The output unit is used to output the test results and send them to the test center, mark the location information of the abnormal infrared lamp beads in the original image, display the specific location of the abnormal lamp beads on the monitor by coordinate positioning, and generate an abnormality report, which includes the number of the abnormal lamp beads, location coordinates, abnormality type, abnormality severity assessment and possible repair suggestions; The storage unit is used to store the information database of the monitor, the parameters of the abnormality recognition model and the judgment criteria of different abnormality types; including the location information database of infrared lamp beads, the frequency domain waveform feature threshold, the physical damage judgment criteria and the abnormality type judgment criteria, so that each unit can call them during operation.

[0031] An infrared lamp bead abnormality detection device for a monitor proposed in an embodiment of the present invention, first, the image acquisition unit starts working, and the monitor to be detected is placed in a detection box. The detection box uses shielding materials and light-absorbing materials that shield external light interference and electromagnetic interference, which can reduce external ambient light and electromagnetic interference and ensure a stable shooting environment. Based on an image acquisition device installed in a position that can capture the luminous state of the infrared lamp beads, the infrared lamp beads in the monitor are continuously photographed to obtain image data containing the infrared lamp beads of the monitor, and transmit it to the image processing unit.

[0032] Next, the image optimization module of the image processing unit processes the received image data. The median filter is used to remove the salt and pepper noise in the image data, and then the unsharp mask algorithm is used to sharpen the image data to highlight the edge features of the infrared lamp beads. Subsequently, the image segmentation module determines the position information of each infrared lamp bead on the monitor based on the information database of the monitor, divides the image data into multiple segmented images and transmits them to the abnormality detection unit. During segmentation, the image data will be laid out as a three-dimensional matrix, and cylindrical segmentation will be performed based on the position information of the infrared lamp beads. Then, the segmented cylindrical image will be layered and cut in three dimensions to obtain a multi-layer two-dimensional slice image of each infrared lamp bead.

[0033] After that, the abnormality detection unit starts the detection work. The edge recognition module identifies the edge contour information of the infrared lamp beads in the segmented image, determines the edge position through the edge algorithm, and judges whether the infrared lamp beads are physically damaged based on the edge contour information. Physical damage includes external damage and internal damage. External damage is judged based on the curvature change of multiple edge contour information. If the curvature change exceeds the preset threshold, it is judged as external damage; internal damage is judged by calculating the pixel value mean and standard deviation of the internal area of ​​the edge contour. If the pixel value mean is lower than the preset threshold or the standard deviation exceeds the preset threshold, it is judged as internal damage. Infrared lamp beads that are identified to be physically damaged will be marked as abnormal.

[0034] For the segmented images that do not identify physical damage, the abnormality recognition module performs Fourier transform on them and inputs the converted frequency domain information into the pre-trained abnormality recognition model. The abnormality recognition model is based on a convolutional neural network. It identifies abnormal information of lamp beads by analyzing the characteristics of frequency domain waveforms, such as brightness abnormality, non-luminous abnormality, and flickering abnormality, and marks abnormal infrared lamp beads with abnormalities. Brightness abnormality is determined by analyzing the amplitude change of low-frequency components in the frequency domain waveform; non-luminous abnormality is determined by detecting whether the amplitude of all frequency components in the frequency domain waveform is close to zero; flickering abnormality is determined by analyzing the amplitude change of high-frequency components in the frequency domain waveform.

[0035] After that, the output unit outputs the test results and sends them to the test center. The position information of the abnormal infrared lamp beads is marked in the original image, and the specific position of the abnormal lamp beads on the monitor is displayed by coordinate positioning. At the same time, an abnormality report is generated. The abnormality report contains the number of the abnormal lamp beads in the monitor information database, the position coordinates, the abnormality type, the abnormality severity assessment, and possible repair suggestions. If there are other monitors that need to be tested after one test is completed, the above image acquisition, processing, abnormality detection, and result output process are repeated, so that the infrared lamp beads of multiple monitors can be abnormally detected, and abnormal infrared lamp beads can be found and marked in time, providing guarantee for the maintenance and normal operation of the monitor.

[0036] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for detecting abnormality of infrared lamp beads for a monitor, characterized in that: The following steps are included: S101: placing a monitor in a closed detection box, continuously photographing the infrared lamp beads in the monitor based on an image acquisition device to obtain image data containing the infrared lamp beads of the monitor, and the image acquisition device is installed at a position where the light-emitting state of the infrared lamp beads can be captured; S102: using median filtering to remove salt and pepper noise in the image data, and using an unsharp mask algorithm to sharpen the image data to highlight the edge features of the infrared lamp beads; S103: performing matching based on the information database of the monitor, determining the position information of each infrared lamp bead on the monitor, and dividing the image data into multiple segmented images based on the position information of each infrared lamp bead on the monitor; S104: Identify edge contour information of the infrared lamp beads in the segmented image, and determine whether the infrared lamp beads are physically damaged based on the edge contour information; The infrared lamp beads that are identified to be physically damaged will be marked as abnormal; The segmented images without physical damage are Fourier transformed, and the converted frequency domain information is input into the pre-trained abnormality recognition model. The abnormality recognition model identifies the abnormal information of the lamp beads by analyzing the frequency domain waveform characteristics, and marks the abnormal infrared lamp beads as abnormal. S105: Output the test results and send the test results to the testing center; Test results include: The location information of abnormal infrared lamp beads is marked in the original image information, and the specific location of the abnormal lamp beads on the monitor is displayed in the form of coordinate positioning; Abnormal type of abnormal infrared lamp beads, including physical damage and abnormal information; Generate an exception report, which includes the number of the abnormal lamp bead in the monitor information database, location coordinates, abnormality type, abnormality severity assessment and possible repair suggestions.

2. A method for detecting abnormalities of infrared lamp beads for a monitor according to claim 1, characterized in that: In S103, dividing the image data into a plurality of divided images includes: The image data is laid out as a three-dimensional matrix. Based on the position information of each infrared lamp bead on the monitor, each infrared lamp bead is segmented into a cylinder. Then, the segmented cylindrical image is layered and cut in three dimensions to obtain a multi-layer two-dimensional slice image of each infrared lamp bead.

3. A method for detecting abnormality of infrared lamp beads for a monitor according to claim 2, characterized in that: The three-dimensional matrix is ​​represented as: Where x and y represent two-dimensional coordinates, and z represents the depth coordinate of the layered cutting.

4. A method for detecting abnormalities of infrared lamp beads for a monitor according to claim 3, characterized in that: The cylindrical segmentation formula for each infrared lamp bead is: In the formula, Represents the center coordinates of the lamp bead, r represents the radius of the cylinder, which is dynamically adjusted according to the actual size and distribution density of the infrared lamp bead.

5. A method for detecting abnormalities of infrared lamp beads for a monitor according to claim 2, characterized in that: In S104, identifying the edge contour information of the infrared lamp beads in the segmented image includes: The edge contour information of the infrared lamp beads in each two-dimensional slice image is identified by the edge algorithm. The edge algorithm formula is: In the formula, represents the gradient of the image in the x direction, Represents the gradient of the image in the y direction; Based on the edge profile information, determine whether the infrared lamp beads are physically damaged.

6. A method for detecting abnormalities of infrared lamp beads for a monitor according to claim 5, characterized in that: Physical damage includes external damage and internal damage; The surface damage is based on the curvature change of multiple edge contour information. If the curvature change exceeds a preset threshold, it is determined as surface damage; Internal damage is based on calculating the mean and standard deviation of the pixel values ​​in the internal area of ​​the edge contour. If the mean pixel value is lower than the preset threshold or the standard deviation exceeds the preset threshold, it is determined to be internal damage.

7. A method for detecting abnormalities of infrared lamp beads for a monitor according to claim 2, characterized in that: Fourier transform includes: The two-dimensional slice image without physical damage is Fourier transformed. The Fourier transform formula is: In the formula, represents a two-dimensional slice image, Represents frequency domain information; The converted frequency domain information is input into the pre-trained anomaly recognition model to identify the abnormal information of the lamp beads.

8. A method for detecting abnormalities of infrared lamp beads for a monitor according to claim 7, characterized in that: Abnormal information includes abnormal brightness, abnormal non-luminescence and abnormal flickering. The abnormal information of identifying lamp beads specifically includes: Abnormal brightness recognition: By analyzing the amplitude changes of the low-frequency components in the frequency domain waveform, it is determined whether the brightness is abnormal. The abnormal brightness recognition formula is: In the formula, represents the sum of the amplitudes of the low-frequency components, and Indicates the low frequency range; Abnormal identification of no light: By detecting whether the amplitude of all frequency components in the frequency domain waveform is close to zero, it is judged whether the lamp bead is not emitting light. The abnormal identification formula of no light is: In the formula, represents the total amplitude; Flicker anomaly identification: By analyzing the amplitude changes of high-frequency components in the frequency domain waveform, it is determined whether the lamp bead flickers. The flicker anomaly identification formula is: In the formula, represents the sum of the amplitudes of high-frequency components, and Indicates the high frequency range.

9. A method for detecting abnormalities of infrared lamp beads for a monitor according to claim 1, characterized in that: The anomaly recognition model is a deep learning model based on convolutional neural networks. Its training process includes: Collect image data of multiple normal and abnormal infrared lamp beads as training sets; Label the training set, including the location of the infrared lamp beads and the type of anomaly; Use the labeled training set to train the anomaly recognition model and optimize the model parameters.

10. An infrared lamp bead abnormality detection device for a monitor, characterized in that: Included are: The image acquisition unit is used to acquire the image data of the infrared lamp beads of the monitor and transmit it to the image processing unit; the image acquisition unit includes a detection box and an image acquisition device, the detection box is used to place the monitor to be detected, and its interior can shield external light interference and electromagnetic interference to ensure a stable shooting environment; the image acquisition device is installed in a position where it can capture the luminous state of the infrared lamp beads, and continuously shoots the infrared lamp beads in the monitor; The image processing unit includes an image optimization module and an image segmentation module. The image optimization module uses a median filter to remove salt and pepper noise in the image data, and uses an unsharp mask algorithm to sharpen the image data to highlight the edge features of the infrared lamp beads. The image segmentation module determines the position information of each infrared lamp bead on the monitor based on the information database of the monitor, and divides the image data into multiple segmented images and transmits them to the abnormality detection unit; The abnormality detection unit includes an edge recognition module and an abnormality recognition module. The edge recognition module is used to recognize the edge contour information of the infrared lamp beads in the segmented image, and judge whether the infrared lamp beads are physically damaged based on the edge contour information, and mark the infrared lamp beads that are physically damaged as abnormal; The abnormality recognition module performs Fourier transform on the segmented images that do not identify physical damage, and then inputs the converted frequency domain information into the pre-trained abnormality recognition model. By analyzing the frequency domain waveform characteristics, the abnormal information of the lamp beads is identified, including brightness abnormality, non-luminous abnormality and flickering abnormality, and abnormal infrared lamp beads with abnormalities are marked as abnormal; The output unit is used to output the test results and send them to the test center, mark the location information of the abnormal infrared lamp beads in the original image, display the specific location of the abnormal lamp beads on the monitor by coordinate positioning, and generate an abnormality report, which includes the number of the abnormal lamp beads, location coordinates, abnormality type, abnormality severity assessment and possible repair suggestions; The storage unit is used to store the information database of the monitor, the parameters of the abnormality recognition model and the judgment criteria of different abnormality types; the storage unit includes the location information database of infrared lamp beads, the frequency domain waveform feature threshold, the physical damage judgment criteria and the abnormality type judgment criteria, so that each unit can call it during operation.

Images