A method, apparatus, medium, and program product for monitoring status of an indicator light

Abstract

The application discloses a monitoring method, device, medium and program product of an indicator light state, relates to the technical field of computers, and the method comprises the following steps: after obtaining target observation data of an indicator light, target environment data of an environment where the indicator light is located and a first estimated state determined in advance, constructing a joint probability distribution according to a causal relationship among observation variables, environment variables and the indicator light state, and solving a conditional probability obtained based on the joint probability distribution by using a maximum posterior probability estimation algorithm. In the process, the influence of the environment on the observation variables is considered in the link of the indicator light state monitoring, the problem that environment factors are ignored in related technologies is solved, and the accuracy of the monitoring result is improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, medium, and program product for monitoring indicator light status. Background Art

[0002] Indicator lights are widely used in various fields as status indication components.

[0003] Currently, there are two main methods for detecting indicator light status: single-modal detection and static fusion detection. Single-modal detection primarily determines the actual status of an indicator light based on a single visual or electrical data set. Static fusion detection, on the other hand, uses a fixed weight to perform a weighted summation of the visual and electrical data to determine the indicator light status based on the final result.

[0004] However, these detection methods all ignore the impact of the environment on the detection results, resulting in a high misjudgment rate. Summary of the Invention

[0005] The present application provides a method, device, medium and program product for monitoring the status of an indicator light, so as to at least solve the problem of high misjudgment rate in related technologies.

[0006] This application provides a method for monitoring the status of an indicator light, comprising:

[0007] Obtain target observation data of the indicator light, target environmental data of the environment in which the indicator light is located, and a predetermined first estimated state, wherein the first estimated state is determined based on the target observation data; construct a joint probability distribution among observation variables, environmental variables, and indicator light states based on a pre-constructed causal relationship diagram, wherein the causal relationship diagram includes: causal relationships among observation variables, environmental variables, and indicator light states; obtain the conditional probability of the indicator light state when the observation variable is the target observation data based on the joint probability distribution; based on the target observation data, target environmental data, and conditional probability, use the maximum a posteriori probability estimation algorithm with the first estimated state as the initial value to solve the conditional probability, and obtain a second estimated state when the conditional probability takes the maximum value, and a first probability value corresponding to the second estimated state; determine the target state of the indicator light based on the second estimated state and the first probability value.

[0008] The present application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for implementing the steps of any of the above-mentioned methods for monitoring the status of an indicator light when executing the computer program.

[0009] The present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods for monitoring the status of an indicator light are implemented.

[0010] The present application also provides a computer program product, including a computer program, which implements the steps of any of the above-mentioned methods for monitoring the status of an indicator light when the computer program is executed by a processor.

[0011] Through this application, a joint probability distribution is constructed based on the causal relationship between observed variables, environmental variables and indicator light status, and the impact of the environment on the observed variables is taken into account in the indicator light status monitoring process, which solves the problem of ignoring environmental factors in related technologies and improves the accuracy of monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0013] Figure 1 A flow chart of a method for monitoring the status of an indicator light provided in an embodiment of the present application;

[0014] Figure 2 A cause-effect relationship diagram provided in an embodiment of the present application;

[0015] Figure 3 A schematic diagram of a process for verifying the credibility of a second estimated state provided in an embodiment of the present application;

[0016] Figure 4 A causal relationship diagram after the elimination operation provided in an embodiment of the present application;

[0017] Figure 5 A schematic diagram of a process for diagnosing observed variables provided in an embodiment of the present application;

[0018] Figure 6 This is a structural connection diagram of the indicator light status monitoring device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0020] It should be noted that in the description of the present application, the terms "comprising", "containing" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or equipment. The terms "first", "second" and the like in the present application are used to distinguish similar objects, not to describe a specific order or sequence.

[0021] In conjunction with the specific application environment architecture or specific hardware architecture on which the monitoring method of the indicator light state depends, the specific application environment architecture or specific hardware architecture is described herein.

[0022] In the related art, the detection of the indicator light state mainly includes two ways, namely single-mode detection and static fusion detection.

[0023] Single-mode detection mainly determines the actual state of the indicator light through single visual data or electrical data. Common pure visual detection schemes include light characteristic detection and automatic optical detection. Light characteristic detection is to measure parameters such as luminous flux and color temperature by integral method, spectral method and variable angle photometer method. Automatic optical detection is to automatically detect the quality of the indicator light through a camera system and machine learning. The pure visual detection scheme is greatly disturbed by light and shielding, resulting in low accuracy of the recognized visual data. Common pure electrical detection schemes include electrical characteristic detection. Electrical characteristic detection is to measure parameters such as voltage and current by a constant current and constant voltage source. The pure electrical detection scheme is only used to monitor voltage / current overrun and cannot identify the color or flashing state of the indicator light.

[0024] Static fusion detection mainly adopts fixed weights to weight and sum the visual and electrical scores to determine the state of the indicator light according to the final result. Although static fusion detection combines visual data and electrical data, it still cannot consider the influence of environmental factors, resulting in a high misjudgment rate of the final detection result.

[0025] Therefore, the embodiments of the present application provide a monitoring method, device, medium and program product of an indicator light state, which constructs a joint probability distribution according to the causal relationship among observation variables, environmental variables and the state of the indicator light, considers the influence of the environment on the observation variables in the link of the monitoring of the state of the indicator light, solves the problem of ignoring environmental factors in the related art, and improves the accuracy of the monitoring result.

[0026] The embodiments of the present application provide a monitoring method of an indicator light state, as shown in Figure 1 The monitoring method is described in detail in conjunction with the execution process shown in Figure 1 Figure 1 ​As shown, the method includes the following steps:

[0027] S101, obtaining target observation data of the indicator light, target environment data of the environment in which the indicator light is located, and a predetermined first estimated state.

[0028] Among them, the indicator light refers to a lamp with a status indication function, such as a light emitting diode (LED). The target observation data refers to the value corresponding to the variable that can characterize the state of the indicator light. The variables that can characterize the state of the indicator light include but are not limited to the color, flashing state, current, voltage, value of the status register, etc. of the indicator light. The target environment data is the value of the variable that affects the target observation data in the environment where the indicator light is located. The variables that affect the target observation data in the environment where the indicator light is located include but are not limited to temperature, light, voltage fluctuation signals, etc. The first estimated state is a state obtained by preliminarily estimating the actual state of the indicator light based on the target observation data, that is, the first estimated state is determined based on the target observation data. The target observation data and target environment data involved in the embodiments of the present disclosure can be the original data directly collected by the acquisition device, or they can be data obtained after preprocessing the original data. The preprocessing process of the original data and the determination process of the first estimated state have been described in the following embodiments and will not be repeated here.

[0029] Specifically, target observation data of the indicator light and target environment data of the environment in which the indicator light is located are obtained, and then, based on the target observation data of the indicator light, a preliminary estimate is made of the actual state of the indicator light to obtain a first estimated state.

[0030] In some optional embodiments, the target observation data includes target visual data, target electrical data and target state value. The target visual data includes color probability distribution, flicker state and hue characteristic value. The target electrical data includes the voltage signal after removing the trend item, the current signal after removing the trend item and the electrical anomaly index, and the target environmental data includes at least one of the target temperature value, the target light intensity and the target electrical fluctuation value. In the subsequent embodiments corresponding to the present disclosure, unless otherwise specified, the target observation data and target environmental data in this embodiment will be used as an example for explanation. However, in actual application scenarios, the target observation data and target environmental data are not limited to the above examples, and those skilled in the art can redefine them based on actual conditions.

[0031] S102: Construct a joint probability distribution among observation variables, environmental variables, and indicator light states based on a pre-constructed causal relationship graph.

[0032] The causal relationship diagram includes the causal relationships between observed variables, environmental variables, and indicator light status. This diagram can be generated using a pre-trained causal relationship recognition model or algorithm. The diagram corresponds to the application scenario. Observed variables are variables that represent the indicator light status. Environmental variables are variables in the indicator light's environment that affect the target observed data.

[0033] Specifically, the causal relationship between each variable is obtained from the causal relationship graph, and based on each causal relationship, the joint probability distribution among the observed variables, environmental variables, and indicator light status is constructed.

[0034] For example, Figure 2 The causal relationship diagram shown includes the visual variable V, electrical variable E, and flag state variable F under the observation variables; the lighting variable L and electrical fluctuation variable under the environmental variables, and the indicator light state S. The pre-built causal relationship diagram includes the causal relationship between the visual variable V, electrical variable E, flag state variable F, lighting variable L, electrical fluctuation variable D, and indicator light state S. The causal relationship is specifically expressed as:

[0035] S V: The indicator light state S determines the visual variable V;

[0036] S E: The indicator light status S determines the electrical variable E;

[0037] F S: Flag state variable F affects indicator light state S;

[0038] L V: The value of the illumination variable L affects the value of the visual variable V;

[0039] D E: The value of the electrical fluctuation variable D affects the value of the electrical variable E.

[0040] According to the above causal relationship, the following joint probability distribution is constructed:

[0041]

[0042] in, is the joint probability distribution, is the visual likelihood term, is the electrical likelihood term, is the sign bit likelihood term, State prior probability.

[0043] S103, obtaining the conditional probability of the indicator light state when the observed variable is the target observation data according to the joint probability distribution.

[0044] Specifically, the Bayesian algorithm is used to convert the joint probability distribution into the conditional probability of the indicator light state when the observed variable is the target observation data. The conditional probability is as follows:

[0045]

[0046] S104, based on the target observation data, target environment data and conditional probability, with the first estimated state as the initial value, the conditional probability is solved using the maximum a posteriori probability estimation algorithm to obtain the second estimated state when the conditional probability takes the maximum value, and the corresponding first probability value of the second estimated state.

[0047] Specifically, the target observation data, target environment data, and first estimated state are all substituted into the above conditional probability. With the first estimated state as the initial value of the indicator light state, the conditional probability is solved using the maximum a posteriori probability estimation algorithm to obtain the second estimated state that maximizes the conditional probability, as well as the first probability value corresponding to the second estimated state. The second estimated state here can be the same as or different from the first estimated state.

[0048] In the solution process, the first estimated state is used as the initial value of the indicator light state, all possible states of the indicator light are traversed, and then the posterior probability of each state is calculated. Finally, the state when the conditional probability takes the maximum value is selected, that is, the second estimated state.

[0049] Maximize the posterior probability estimate:

[0050]

[0051] During the solution process, the visual likelihood term The specific value of can be obtained from the pre-built visual probability table based on the estimated indicator light state (such as the initial value first estimated state) and the target light intensity in the target environment data. The visual probability table is constructed as follows:

[0052] Step a1, data collection

[0053] For compliance indicators, Under different lighting conditions (such as "red light is always on" or "green light is flashing"), a large number of frame images (such as 1000 frames) corresponding to different light intensities are collected, and the collected large number of images are input into a pre-built recognition model (such as the YOLO-Lite model) to obtain the visual recognition results corresponding to each frame image.

[0054] Step a2, probability calculation

[0055] For each indicator light status Corresponding 1000 frames of images, from the visual recognition results corresponding to the 1000 frames of images, statistical visual recognition results and indicator light status The number of consistent frame images will be consistent with the indicator status The consistent visual recognition results are considered to be correct and will be compared with the indicator light status. Inconsistent visual recognition results are considered recognition errors. For example, if the indicator light state S is solid red, and the result obtained after recognizing a corresponding frame of image is flashing green, flashing red, or solid green, then the visual recognition result is considered inconsistent with the indicator light state S, indicating a recognition failure. If the result obtained after recognizing a corresponding frame of image is solid red, then the visual recognition result is considered consistent with the indicator light state S, indicating a successful recognition.

[0056]

[0057] in, is the visual recognition result; The indicator light status (any of all the states corresponding to the indicator light under normal circumstances); Light intensity level (divided into four levels, the first level L1 corresponds to a light intensity range of 0lux-200lux, the second level L2 corresponds to a light intensity range of 200lux-500lux, the third level L3 corresponds to a light intensity range of 500lux-1000lux, and the fourth level L4 corresponds to a light intensity range of 1000lux-2000lux); To identify the correct number of Total number of times.

[0058] According to steps a1 and a2, through a large amount of experimental data, we can obtain the visual probability table shown in Table 1:

[0059] Table 1

[0060]

[0061] Step a3: Update mechanism

[0062] The data in the visual probability table is updated incrementally on a regular basis (e.g. monthly) in the following manner:

[0063]

[0064] During the solution process, the electrical likelihood term The specific value of is determined by the following formula based on the estimated indicator light state (such as the initial value first estimated state) and the electrical anomaly index ETA in the target electrical data:

[0065]

[0066] wherein, is the expected value of the electrical abnormality index ETA when the indicator light status is S, is the standard deviation of the electrical abnormality index ETA when the indicator light status is S.

[0067] In the solving process, the flag likelihood term is obtained according to the following manner:

[0068]

[0069] is the probability that the status in the status register is consistent with the indicator light status;

[0070] The above formula is obtained according to the basic verification and the field verification:

[0071] The basic verification is: according to the error rate of the status register in the indicator light hardware specification book <0.1%>, therefore, the probability of the basic verification is determined as: .

[0072] The field verification is: through the random sampling method, the number of times that the status value stored in the status register in the sample under inspection matches the indicator light status is counted, so as to determine the probability of the field verification as:

[0073]

[0074] wherein, is the probability of the field verification, is the number of matches, is the number of samples under inspection.

[0075] Finally, according to the probability of the basic verification and the probability of the field verification, the probability that the status in the status register is consistent with the indicator light status is determined as:

[0076]

[0077] In the solving process, the status prior probability is obtained according to the following manner:

[0078]

[0079] wherein, is the probability that the indicator light status is under normal circumstances, is any status in the indicator light status. It can be obtained by randomly reading the server running log, and the fields that can be read are [timestamp, device ID, LED status]. In addition, It can also be obtained through the following statistical methods:

[0080]

[0081] in, The indicator light status is The number of occurrences of is the indicator status number, is the number of occurrences of all states.

[0082] S105: Determine a target state of the indicator light according to the second estimated state and the first probability value.

[0083] Specifically, after obtaining the second estimated state and its corresponding first probability value, the target state of the indicator light is determined based on the difference between the first probability value and the preset probability threshold. When the first probability value is greater than or equal to the preset probability threshold, it indicates that the current second estimated state is credible, and the second estimated state is determined as the target state. When the first probability value is less than the preset probability threshold, it indicates that it is impossible to determine whether the second estimated state is credible. Therefore, it is necessary to verify the credibility of the second estimated state by other means, and further determine the target state of the indicator light based on the verification result. The preset probability threshold can be a reference value for evaluating the credibility of the second estimated state obtained through a large number of experiments, such as 0.8.

[0084] The indicator light status monitoring method provided in this application constructs a joint probability distribution based on the causal relationship between the observed variables, environmental variables and the indicator light status, and takes the impact of the environment on the observed variables into consideration in the indicator light status monitoring process, thereby solving the problem of ignoring environmental factors in related technologies and improving the accuracy of the monitoring results.

[0085] In some optional implementations, a step is added to verify the credibility of the second estimated state when the first probability value is less than a preset probability threshold, such as Figure 3 As shown, the method further includes:

[0086] S301: When the first probability value is less than a preset probability threshold, the causal relationship pointing to the indicator light state in the causal relationship graph is removed to obtain a causal relationship graph after removal.

[0087] For example, the causal relationship in the above embodiment is still used as an example for illustration. When the first probability value is less than the preset probability threshold, the causal relationship pointing to the indicator light state S in the causal graph is removed, that is, F S is eliminated, and we get Figure 4 The causal relationship diagram after elimination is shown.

[0088] S302 : In the causal relationship diagram after elimination, for each observed variable, an environmental variable or other observed variable pointing to the observed variable is determined as an associated variable of the observed variable.

[0089] Among them, other observed variables refer to all observed variables except the observed variables currently being discussed.

[0090] For example, in Figure 4 In the causal relationship diagram after elimination shown, the observed variables include the flag state variable F, the visual variable V, and the electrical variable E. For the flag state variable F, there are no environmental variables or other observed variables pointing to it. Therefore, subsequent operations on the flag state variable F are abandoned. For the visual variable V, there are no other observed variables pointing to it, but there are environmental variables pointing to it. Therefore, the illumination variable L pointing to the visual variable V is determined as the associated variable of the visual variable V. In a similar manner, for the electrical variable E, the electrical fluctuation variable D is used as the associated variable of the electrical variable E.

[0091] S303 , for each observed variable, determine a first deviation corresponding to the observed variable according to at least two of the second estimated state, target observed data corresponding to the observed variable, and target associated data corresponding to the associated variable.

[0092] The target associated data is the target data corresponding to the associated variable. If the associated variable is an environmental variable, the target associated data is the environmental data corresponding to the environmental variable; if the associated variable is an observation variable, the target associated data is the observation data corresponding to the observation variable. The first deviation is used to indicate the credibility of the target observation data. The smaller the first deviation, the higher the credibility; the larger the first deviation, the lower the credibility.

[0093] In some optional implementations, S303 includes: determining, according to a determination method corresponding to the observed variable, that when the indicator light state is the second estimated state and the associated variable takes the target associated data, the observed variable takes the second probability value of the target observed data; and determining the first deviation based on the second probability value.

[0094] For example, for the visual variable V, the corresponding determination method is as follows:

[0095]

[0096] in, is the first deviation corresponding to the visual variable V, It indicates that when the indicator light state is the second estimated state and the illumination variable L takes the target illumination intensity, the visual variable V takes the probability of the target visual data (ie, the second probability value). It can be obtained by looking up the visual probability table.

[0097] In some optional embodiments, the first deviation degree of the observation variable can also be determined by the following manner:

[0098] Taking the electrical variable E as an example, the determination manner corresponding to the electrical variable E is as follows:

[0099]

[0100] wherein, is the first deviation degree corresponding to the electrical variable E, is the voltage signal after removing the trend term, is the current signal after removing the trend term, is the voltage mean value in the target electrical fluctuation value, is the voltage standard deviation in the target electrical fluctuation value, is the current mean value in the target electrical fluctuation value, is the current standard deviation in the target electrical fluctuation value.

[0101] S304, determining the target state of the indicator light according to the first deviation degree corresponding to each observation variable.

[0102] S304 specifically includes:

[0103] S3041, obtaining a preset deviation degree threshold corresponding to each observation variable.

[0104] Specifically, there is a preset deviation degree threshold corresponding to each observation variable, and the preset deviation degree threshold is obtained by statistical analysis of a large number of test results. For example, the preset deviation degree threshold of the visual variable V is 0.5, and the preset deviation degree threshold of the electrical variable E is 3.

[0105] S3042, comparing the first deviation degree with the preset deviation degree threshold for each observation variable to obtain a comparison result.

[0106] For example, the first deviation degree corresponding to the visual variable V is compared with the preset deviation degree threshold corresponding to the visual variable V to obtain the comparison result corresponding to the visual variable V. For example, the first deviation degree corresponding to the electrical variable E is compared with the preset deviation degree threshold corresponding to the electrical variable E to obtain the comparison result corresponding to the electrical variable E.

[0107] S3043, when the comparison result corresponding to each observation variable is that the first deviation degree is less than or equal to the preset deviation degree threshold, the second estimated state is determined as the target state.

[0108] For example, when the first deviation degree corresponding to the visual variable V is less than or equal to the preset deviation degree threshold, the second estimated state is determined as the target state. ​Less than or equal to the preset deviation threshold 0.5 corresponding to the visual variable V, and the first deviation corresponding to the electrical variable E When the value is less than or equal to the preset deviation threshold of 3 corresponding to the electrical variable E, it indicates that the credibility of the target visual data and the credibility of the target electrical data are both high. Based on this, it can be determined that the credibility of the second estimated state obtained based on the target visual data and the target electrical data is also high. Therefore, the second estimated state can be determined as the target state.

[0109] The embodiment of the present disclosure determines the target state of the indicator light by the first deviation of the observed variable, which can ensure that when the first probability value is small, the first probability value is re-evaluated by calculating the first deviation of the observed variable, thereby further improving the accuracy of the result.

[0110] If at least one of the observed variables has a first deviation greater than a preset deviation threshold, the target observation data corresponding to the observed variable with a first deviation greater than the preset deviation threshold is considered to have low credibility, i.e., the target observation data is considered to have low accuracy. In this case, further verification of the specific cause of the low credibility is necessary. Based on this, the following implementation method is proposed.

[0111] In some optional embodiments, such as Figure 5 As shown, the method further includes the following steps:

[0112] S501 : When, among the observed variables, there is at least one observed variable whose comparison result is that the first deviation is greater than a preset deviation threshold, obtain the standard value of the associated variable corresponding to each observed variable.

[0113] Specifically, when the comparison result of at least one observed variable among the observed variables is that the first deviation is greater than the preset deviation threshold, it is necessary to obtain the standard values ​​of the associated variables corresponding to all the observed variables participating in the calculation of the first deviation. Still taking the above embodiment as an example, if the first deviation corresponding to the visual variable V is greater than the preset deviation threshold, Less than or equal to the preset deviation threshold 0.5 corresponding to the visual variable V, but the first deviation corresponding to the electrical variable E If the deviation is greater than the preset deviation threshold of 3 corresponding to the electrical variable E, the target electrical data corresponding to the electrical variable is considered unreliable. At this point, it is necessary to obtain the standard value of the illumination variable L, the associated variable corresponding to the visual variable V, and the standard value of the electrical fluctuation variable D, the associated variable corresponding to the electrical variable E.

[0114] S502 : For each observed variable, determine a second deviation corresponding to the observed variable based on at least two of the second estimated state, target observed data corresponding to the observed variable, and a standard value of the associated variable.

[0115] In some optional implementations, S502 includes: determining, according to a determination method corresponding to the observed variable, that when the indicator light state is the second estimated state and the associated variable takes the standard value, the observed variable takes the third probability value of the target observation data; and determining the second deviation based on the third probability value.

[0116] For example, for the visual variable V, the corresponding determination method is as follows:

[0117]

[0118] in, is the second deviation corresponding to the visual variable V, Indicates that the indicator light status is the second estimated status And the illumination variable L takes the standard value When , the visual variable V takes the probability of the target visual data (ie, the third probability value). It can be obtained by looking up the visual probability table.

[0119] In some other optional implementations, the second deviation can also be determined in the following manner:

[0120] For the electrical variable E, the second deviation is determined as follows:

[0121]

[0122] in, is the second deviation corresponding to the electrical variable E, is the voltage signal after removing the trend term, is the current signal after removing the trend term, is the standard mean voltage (the mean voltage given in the indicator light specification), is the voltage standard deviation (the current standard deviation given in the indicator light specification), is the standard mean current (the mean current given in the indicator light specification), is the current standard deviation in the second estimated state (the current standard deviation obtained based on sample data statistics).

[0123] S503: Determine the target state of the indicator light according to the second deviation and the standard value corresponding to each observed variable.

[0124] S503 specifically includes:

[0125] S5031 , for each observed variable, diagnose the observed variable according to the second deviation corresponding to the observed variable and the standard value of the associated variable to obtain a diagnosis result corresponding to the observed variable.

[0126] Specifically, for each observed variable, a pre-built diagnostic mapping table exists. This table lists the value range of the second deviation, the value range of the target associated data corresponding to the associated variable, and the diagnostic result. Therefore, for each observed variable, the diagnostic result corresponding to the second deviation corresponding to the observed variable and the standard value of the associated variable can be found by searching the mapping relationship.

[0127] For example, the mapping relationship corresponding to the visual variable V is shown in Table 2:

[0128] Table 2

[0129]

[0130] For example, the mapping relationship corresponding to the electrical variable E is shown in Table 3:

[0131] Table 3

[0132]

[0133] S5032: When the diagnosis results corresponding to the observed variables are all normal, the second estimated state is determined as the target state.

[0134] Specifically, when the diagnostic results corresponding to each observed variable are normal, the second estimated state is determined as the target state. For example, based on the table in the above embodiment, if the diagnostic results corresponding to the visual variable are normal visual fluctuations and the diagnostic results corresponding to the electrical variable are normal electrical fluctuations, the diagnostic results corresponding to the visual variable are considered normal, and the diagnostic results corresponding to the electrical variable are considered normal. In this case, the second estimated state can be directly determined as the target state.

[0135] In some optional embodiments, the method further comprises:

[0136] When the diagnosis result corresponding to at least one observed variable among the observed variables is abnormal, the target causal effect intensity corresponding to the causal relationship between the associated variable and the observed variable is obtained.

[0137] Specifically, in addition to the causal relationships among the observed variables, environmental variables, and indicator light states, the causal relationship graph also includes the causal effect strength corresponding to each causal relationship. When the diagnosis result corresponding to at least one observed variable among the observed variables is abnormal, the target causal effect strength corresponding to the causal relationship between the associated variable and the observed variable is first obtained. For example, the causal effect strength corresponding to the causal relationship between the illumination variable L and the visual variable V (i.e., the target causal effect strength) is obtained. ). Obtain the causal effect intensity corresponding to the causal relationship between the electrical fluctuation variable D and the electrical variable E (i.e., the target causal effect intensity). ).

[0138] For each observed variable, the weight of the observed variable is determined based on the target causal effect strength corresponding to each observed variable, the target correlation data of each associated variable, the standard value of each associated variable, the second deviation corresponding to each observed variable, and the preset parameter set;

[0139] Specifically, the weights of the observed variables are determined as follows:

[0140] Step b1: For each observed variable, calculate the confidence level corresponding to the observed variable based on the target causal effect strength corresponding to the observed variable, the target correlation data of the associated variables corresponding to the observed variable, and the standard values ​​of each associated variable. The confidence level determination method is as follows:

[0141]

[0142] in, is the confidence level corresponding to the observed variable, is the target causal effect strength, is the associated variable corresponding to the observed variable, is the observed variable, For target associated data, is the maximum value of the associated variable, It can be any one of the visual variable V, electrical variable E and flag state variable F.

[0143] Step b2: Determine the weight of the observed variable based on the confidence level corresponding to each observed variable, the second deviation level corresponding to each observed variable, and the preset parameter set. The weight determination method is as follows:

[0144]

[0145] in, is the weight corresponding to the observed variable, is the initial weight, is 0.6, is 0.3, is 0.1, is the sensitivity coefficient (empirical value = 0.5, which can be adjusted through online learning). The preset parameters include and , is the confidence level corresponding to the observed variable.

[0146] Select the maximum and minimum values ​​from the weights of each observed variable and calculate the difference between the maximum and minimum values.

[0147] Specifically, the weights of all observed variables are sorted, the maximum and minimum values ​​are selected, and the difference between the maximum and minimum values ​​is calculated.

[0148] When the difference is greater than the preset weight threshold, the diagnosis result of the observed variable corresponding to the maximum weight is determined as the target state.

[0149] Specifically, the preset weight threshold is obtained based on a large number of experiments, such as 0.7. When the difference is greater than the preset weight threshold, it indicates that the accuracy of the observed variable is high. In this case, the diagnosis result corresponding to the observed variable with the maximum weight can be determined as the target state.

[0150] When the difference is less than the preset weight threshold, the operation ends and a notification message indicating that the target state determination failed is returned.

[0151] Specifically, when the difference is less than the preset weight threshold, it means that the weights of the currently observed variables are similar, and it is impossible to determine which variable is more reliable based on the weights. In this case, the operation will end and a notification message will be returned indicating that the target state determination failed. This notification message allows the user to manually intervene.

[0152] This embodiment determines the target state of the indicator light through the second deviation, which can ensure that when the accuracy of the second estimated state cannot be determined through the first deviation, the associated variables corresponding to the observed variables are adjusted to the ideal state through counterfactual reasoning, and the deviation of the observed data is calculated again, avoiding the problem of low credibility of the recognition result caused by environmental interference, and further determining the credibility of the recognition result.

[0153] In some optional embodiments, a description of the method for obtaining target observation data and target environmental data is added, that is, before S101, the method also includes: obtaining the original observation data of the indicator light and the original environmental data of the environment in which the indicator light is located; preprocessing the original observation data to obtain target observation data; preprocessing the original environmental data to obtain target environmental data.

[0154] Specifically, both the raw observation data and the raw environmental data are collected by an acquisition device. After obtaining the raw observation data and the raw environmental data, the raw observation data is preprocessed to obtain the target observation data, and the raw environmental data is preprocessed to obtain the target environmental data. It should be noted that different data may correspond to different preprocessing methods. The embodiments of this disclosure do not specifically limit the preprocessing method. Those skilled in the art can determine the data preprocessing method according to actual needs in actual applications.

[0155] For example, a camera captures multiple frames of images of the indicator light at a fixed frequency (e.g., 120 fps) over a preset duration, and uses the captured frames as raw visual data. A voltage sensor captures multiple voltage values ​​of the indicator light over a preset duration at a fixed frequency (e.g., 0.5 Hz), and a current sensor captures multiple current values ​​of the indicator light over a preset duration. These captured sets of voltage and current values ​​are used as raw electrical data. The status value of the indicator light over a preset duration is read from a status register, and this status value is used as the raw status value. Generally speaking, raw status values ​​do not change over a short period of time. Therefore, if the status value of the indicator light does not change over a preset duration, only one raw status value is captured. A temperature sensor captures multiple temperature values ​​of the environment surrounding the indicator light over a preset duration, and a light sensor captures multiple light intensities of the environment surrounding the indicator light over a preset duration. These multiple temperature values, multiple light intensities, and the electrical fluctuation value obtained by processing the raw electrical data are used as raw environmental data.

[0156] When preprocessing the original observation data, different processing methods are used for different types of data in the original observation data.

[0157] The preprocessing method for raw visual data is as follows:

[0158] After obtaining n frames of images captured by the camera, the RGB values ​​of each frame are obtained. The hue value (H value) corresponding to each frame of image in the HSV space is then determined based on the RGB values ​​corresponding to each frame of image. After obtaining the H value corresponding to each frame of image, it is determined whether the H value corresponding to each frame of image is within a preset range. If the H value corresponding to at least one frame of image is not within the preset range, the image corresponding to the H value not within the preset range is determined as an abnormal image. If the number m of abnormal images exceeds half of all images, the image is recaptured. If the number m of abnormal images does not exceed half of all images, the same number of restored images as the abnormal images is generated by averaging each pixel of the remaining images. The H values ​​of the restored images are then recalculated. Thus, an H value sequence consisting of the H values ​​corresponding to n frames of image (nm frames of normal image and / or m frames of restored image) is obtained. After obtaining the H value sequence, the hue average and hue standard deviation are calculated based on the multiple H values ​​in the H value sequence, and these hue average and hue standard deviation are used as hue feature values. After determining the hue feature value, n frames of images (nm frames of normal images and / or m frames of repaired images) are input into the pre-trained recognition model to obtain a set of color probability distributions and flickering states. Finally, the color probability distribution, flickering state and hue feature values ​​after preprocessing of the original visual data are used as the target visual data. Among them, the number of images n captured by the shooting device is an odd number, such as 5. The preset range for screening the H value can be the standard range ±10% in the indicator light specification. The pre-built recognition model includes but is not limited to the YOLO-Lite model. The "color" in the aforementioned color probability distribution refers to the color of the indicator light. For example, there are three colors of indicator lights, namely red, yellow and green. Then, the color probability distribution here refers to the probability distribution of red, yellow and green lights obtained by recognizing n frames of images, such as 0.9 for red light, 0.05 for yellow light, 0.05 for green light, etc. The aforementioned method of determining the H value based on RGB is as follows:

[0159]

[0160] in, is the hue value, 、 、 are the three primary color component values ​​corresponding to each frame image.

[0161] The preprocessing method for raw electrical data is as follows:

[0162] After obtaining the raw electrical data (multiple sets of electrical data, each set of electrical data includes voltage and current values), the raw electrical data needs to be processed in two parts. The first part is to remove the trend item, and the second part is to calculate the electrical anomaly index.

[0163] Part 1: Detrending the raw electrical data using a Hodrick-Prescott filter to separate high-frequency noise and low-frequency drift in the data. This results in detrended electrical data, including detrended voltage and detrended current signals. The detrending process is divided into three steps:

[0164] Step c1: Perform sliding average filtering on the original electrical data. The sliding window and step size can be set according to actual needs. The sliding average filtering is implemented as follows:

[0165]

[0166] in, After sliding average filtering Electrical data corresponding to the moment; is a sliding window; express The electrical data at a given moment can be either current or voltage.

[0167] Step c2: low-pass filter the electrical data after the sliding average filtering.

[0168]

[0169] in, is the electrical data after low-pass filtering (representing the low-frequency trend term), For low-pass filter function, the cutoff frequency is usually used (For example, 0.1Hz) filter (such as Butterworth filter). The mathematical representation of low-pass filter is as follows:

[0170]

[0171] in, and is the filter coefficient, is the filter order, the filter order Determines the steepness of the filter (the slope of the transition band). The higher the order, the steeper the transition band.

[0172] Filter order Determined as follows:

[0173]

[0174] in, Stopband attenuation (dB) is the minimum attenuation (dB) required within the stopband, for example 40dB (i.e. the signal attenuates to 1% of the original amplitude at the stopband frequency); is the passband attenuation (dB), the maximum attenuation allowed in the passband (dB), usually taken as 3dB (i.e., the signal attenuates to about 70.7% of the original amplitude at the passband boundary); is the stopband frequency (Hz), the frequency (Hz) required to achieve stopband attenuation; is the cutoff frequency (Hz), the frequency (Hz) at which the filter begins to attenuate. The frequency components above The frequency component of the signal is attenuated. For example, 0.1 Hz.

[0175] For example, if =20dB, =3 dB, =0.3Hz, =0.1Hz, then, calculated by the above formula, The value is taken as 2. When the filter order is known =2, cutoff frequency =0.1Hz, sampling frequency = 1Hz, calculated by the function signal.butter used to design Butterworth filters =[0.0201, 0.0402, 0.0201], =[1.0, -1.561, 0.641].

[0176] Step c3: Subtract the trend term from the low-pass filtered electrical data.

[0177]

[0178] in, Represents electrical data after removing the trend term (retaining high-frequency fluctuations).

[0179] It should be noted here that if there are multiple groups of electrical data after removing the trend items according to the processing of steps a to c, a group of electrical data after removing the trend items can be obtained by taking the average, and this group of data includes the voltage signal and current signal after removing the trend items.

[0180] Part 2: After obtaining the original electrical data, the electrical anomaly index is calculated based on the original electrical data. The electrical anomaly index is determined as follows:

[0181]

[0182] in, is the electrical anomaly index, is the voltage value corresponding to the last acquisition moment in the original electrical data, a standard working voltage (hereinafter referred to as a nominal voltage) of the indicator light, a current standard deviation calculated according to a plurality of current values in the original electrical data, a current average value calculated according to a plurality of current values in the original electrical data.

[0183] Through the above two parts of the original electrical data processing, the voltage signal after removing the trend term, the current signal after removing the trend term and the electrical anomaly index are obtained, and the voltage signal after removing the trend term, the current signal after removing the trend term and the electrical anomaly index are taken as the target electrical data.

[0184] The preprocessing method for the original state value is as follows:

[0185] Since the original state value generally does not change in a short period of time, the original state value obtained from the state register is directly determined as the target state value.

[0186] At this point, the target visual data, the target electrical data and the target state value obtained by the above method are taken as the target observation data.

[0187] When preprocessing the original environmental data, the same preprocessing method can be used for different types of data in the original environmental data, or different preprocessing methods can be used, and the preprocessing method is not limited in the embodiments of the present disclosure.

[0188] For example, the average of the collected plurality of temperature values can be taken as the target temperature value; the average of the collected plurality of light intensities can be taken as the target light intensity; the mean and standard deviation of the voltage can be calculated according to the plurality of voltage values collected by the voltage sensor; the mean and standard deviation of the current can be calculated according to the plurality of current values collected by the current sensor; and the mean and standard deviation of the voltage and the mean and standard deviation of the current can be taken as the target electrical fluctuation value. At this point, the target temperature value, the target light intensity and the target electrical fluctuation value obtained are taken as the target environmental data.

[0189] In some optional embodiments, a determination step of the first estimated state is added, and the predetermined first estimated state is obtained, including: determining a third estimated state according to at least one data in the target observation data; when the third estimated state is an abnormal state, the third estimated state is determined as the target state; and when the third estimated state is a normal state, the third estimated state is determined as the first estimated state.

[0190] Exemplarily, the target observation data includes target visual data, target electrical data, and a target state value. During the process of determining the first estimated state, the first estimated state may be determined based solely on the target visual data, the target electrical data, or the target state value. Alternatively, the first estimated state may be determined based on any two of the target visual data, the target electrical data, and the target state value. Alternatively, the first estimated state may be determined based on the target visual data, the target electrical data, and the target state value.

[0191] This embodiment merely provides an exemplary description of the process of determining the first estimated state according to the target visual data, the target electrical data, and the target state value.

[0192] Obtain a target state value. Determine whether the target state value is valid by determining whether it is within the state value range defined in the indicator light's corresponding specification. If the target state value is valid, parse the target state value to obtain the indicator light state corresponding to the target state value, and determine the indicator light state as the third estimated state.

[0193] If the target state value is determined to be invalid, it indicates that a valid indicator light state cannot be obtained based on the target state value. At this point, an electrical anomaly index is read from the target electrical data and compared with a preset electrical anomaly threshold (e.g., 0.3). If the electrical anomaly index exceeds the preset electrical anomaly threshold, this indicates that the target electrical data of the indicator light may be abnormal. In this case, the voltage signal after removing the trend item is read from the target electrical data and compared with a preset voltage threshold (e.g., 1.5 times the nominal voltage). If the voltage signal after removing the trend item is greater than the preset voltage threshold, this indicates a voltage fault has occurred in the indicator light. In this case, the voltage fault is considered the third estimated state. If the voltage signal after removing the trend item is less than or equal to the preset voltage threshold, the voltage signal of the indicator light is normal. At this point, the current signal after removing the trend item is read from the target electrical data and compared with a preset current threshold (e.g., 0.2 times the nominal current). If the current signal after removing the trend item is greater than or equal to the preset current threshold, this indicates an electrical anomaly in the indicator light. In this case, the electrical anomaly is directly considered the third estimated state. When the current signal after removing the trend item is less than the preset current threshold, it indicates that the indicator light has a circuit breaker fault. At this time, the circuit breaker fault is used as the third estimated state.

[0194] When the electrical anomaly index is less than or equal to the preset electrical anomaly threshold, the target electrical data of the indicator light is normal. At this point, the target visual data needs to be determined for anomalies. A color probability distribution is obtained from the target visual data, and the maximum probability value is compared with a preset probability threshold (e.g., 0.8). If the maximum probability value is less than or equal to the preset probability threshold, the hue average value is obtained from the hue feature values ​​in the target visual data and a determination is made as to whether the hue average value is within a first preset range (e.g., -5 to 15). If the hue average value is within the first preset range, it indicates a physical anomaly with the red indicator light. In this case, the red light physical anomaly is considered the third estimated state. If the hue average value is not within the first preset range, a determination is made as to whether the hue average value is within a second preset range (e.g., 100 to 140). If the hue average value is within the second preset range, it indicates a physical anomaly with the green indicator light. The green light physical anomaly is considered the third estimated state. If the hue average value is not within the second preset range, it indicates that the indicator light state cannot be estimated based on the current situation. The operation ends. If the maximum probability value is greater than the preset probability threshold, a determination is made as to whether an electrical-optical conflict exists based on the maximum probability value and the electrical anomaly index. If it is determined that there is no electrical-optical conflict, the color corresponding to the flickering state and the maximum probability value is obtained from the target visual data, and the color corresponding to the flickering state and the maximum probability value is used as the third estimated state. If it is determined that there is an electrical-optical conflict, it means that the indicator light state cannot be estimated based on the current situation, and the operation is terminated. It should be added here that the method for judging the electrical-optical conflict is: the electrical anomaly index points to the indicator light being abnormal, but the maximum probability value in the color distribution points to the indicator light being normal, then it can be regarded as a conflict. For example, if the electrical anomaly index is in the range of [0.25, 0.30), the indicator light is close to being abnormal, and the maximum probability value in the color distribution is greater than 0.85, indicating that the indicator light is normal. At this time, it can be regarded as an electrical-optical conflict.

[0195] After obtaining the third estimated state, determine whether the third estimated state is normal. Normal means that the third estimated state is within the normal state range, such as having a certain color and flashing state. For example, the third estimated state finally outputted is a physical abnormality of the green light. Since there is no clear color and flashing state in the third estimated state, the third estimated state is determined to be an abnormal state. After the above judgment, if the third estimated state is a normal state, the third estimated state is determined as the first estimated state. If the third estimated state is an abnormal state, the third estimated state is used as the target state of the indicator light. This embodiment obtains the first estimated state through the above-mentioned estimation method, avoids the situation where the target state of the indicator light is still determined when an obvious fault has occurred, and saves computing resources. In addition, the reliability of the final result is improved by judging the target state value, target electrical data and target visual data in turn.

[0196] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0197] For the description of the features in the embodiment corresponding to the device for monitoring the status of an indicator light, reference can be made to the relevant description of the embodiment corresponding to the method for monitoring the status of an indicator light, which will not be described in detail here.

[0198] A monitoring device for the status of an indicator light, such as Figure 6 As shown, it includes the following modules:

[0199] The first acquisition module 601 is used to acquire target observation data of the indicator light, target environment data of the environment where the indicator light is located, and a predetermined first estimated state, where the first estimated state is determined according to the target observation data.

[0200] The construction module 602 is used to construct a joint probability distribution among observation variables, environmental variables, and indicator light states based on a pre-constructed causal relationship graph, where the causal relationship graph includes the causal relationship among observation variables, environmental variables, and indicator light states.

[0201] The first determining module 603 is configured to obtain, according to the joint probability distribution, the conditional probability of the indicator light state when the observed variable is the target observation data.

[0202] Solution module 604 is used to solve the conditional probability based on the target observation data, target environmental data and conditional probability, with the first estimated state as the initial value, using the maximum a posteriori probability estimation algorithm to obtain the second estimated state when the conditional probability takes the maximum value, and the corresponding first probability value of the second estimated state.

[0203] The second determining module 605 is configured to determine a target state of the indicator light according to the second estimated state and the first probability value.

[0204] In some optional implementations, the second determining module includes:

[0205] The first determining submodule is configured to determine the second estimated state as the target state when the first probability value is greater than or equal to a preset probability threshold.

[0206] In some optional implementations, the second determining module further includes:

[0207] The elimination submodule is used to eliminate the causal relationship pointing to the indicator light state in the causal relationship diagram when the first probability value is less than the preset probability threshold, so as to obtain the causal relationship diagram after elimination; the second determination submodule is used to, in the causal relationship diagram after elimination, determine the environmental variable or other observation variable pointing to the observation variable as the associated variable of the observation variable for each observation variable; the third determination submodule is used to determine the first deviation corresponding to the observation variable for each observation variable based on at least two of the second estimated state, the target observation data corresponding to the observation variable, and the target associated data corresponding to the associated variable; the fourth determination submodule is used to determine the target state of the indicator light according to the first deviation corresponding to each observation variable.

[0208] In some optional implementations, the third determining submodule includes:

[0209] The first determination unit is used to determine, based on a determination method corresponding to the observed variable, a second probability value that the observed variable takes the target observed data when the indicator light state is the second estimated state and the associated variable takes the target associated data; the second determination unit is used to determine the first deviation based on the second probability value.

[0210] In some optional implementations, the fourth determining submodule includes:

[0211] The first acquisition unit is used to obtain the preset deviation threshold corresponding to each observed variable; the comparison unit is used to compare the first deviation with the preset deviation threshold for each observed variable to obtain a comparison result; the third determination unit is used to determine the second estimated state as the target state when the comparison results corresponding to each observed variable are that the first deviation is less than or equal to the preset deviation threshold.

[0212] In some optional implementations, the fourth determining submodule further includes:

[0213] The second acquisition unit is used to obtain the standard value of the associated variable corresponding to each observation variable when the comparison result of at least one observation variable among the observation variables is that the first deviation is greater than the preset deviation threshold; the fourth determination unit is used to determine the second deviation corresponding to each observation variable according to at least two of the second estimated state, the target observation data corresponding to the observation variable and the standard value of the associated variable; the fifth determination unit is used to determine the target state of the indicator light according to the second deviation and the standard value corresponding to each observation variable.

[0214] In some optional implementations, the fourth determining unit includes:

[0215] The first determination subunit is used to determine, in accordance with a determination method corresponding to the observed variable, that when the indicator light state is the second estimated state and the associated variable takes the standard value, the observed variable takes the third probability value of the target observation data; the second determination subunit is used to determine the second deviation based on the third probability value.

[0216] In some optional implementations, the fifth determining unit includes:

[0217] The third determination subunit is used to diagnose each observed variable according to the second deviation corresponding to the observed variable and the standard value of the associated variable to obtain a diagnosis result corresponding to the observed variable; the fourth determination subunit is used to determine the second estimated state as the target state when the diagnosis results corresponding to each observed variable are normal.

[0218] In some optional implementations, the causal relationship graph further includes the causal effect strength corresponding to each causal relationship, and the fifth determining unit further includes:

[0219] The acquisition subunit is used to obtain the target causal effect strength corresponding to the causal relationship between the associated variable and the observed variable when there is at least one observed variable among the observed variables and the corresponding diagnosis result is abnormal; the fifth determination subunit is used to determine the weight of the observed variable for each observed variable according to the target causal effect strength corresponding to each observed variable, the target association data of each associated variable, the standard value of each associated variable, the second deviation corresponding to each observed variable and the preset parameter set; the sixth determination subunit is used to select the maximum and minimum values ​​from the weights of each observed variable and calculate the difference between the maximum and minimum values; the seventh determination subunit is used to determine the diagnostic result of the observed variable corresponding to the maximum weight as the target state when the difference is greater than the preset weight threshold; the eighth determination subunit is used to end the operation when the difference is less than the preset weight threshold and return a notification message that the target state determination failed.

[0220] In some optional embodiments, the target observation data includes target visual data, target electrical data and target state values, the target visual data includes color probability distribution, flicker state and hue characteristic value, the target electrical data includes the voltage signal after removing the trend item, the current signal after removing the trend item and the electrical anomaly index, and the target environmental data includes at least one of the target temperature value, target light intensity and target electrical fluctuation value.

[0221] In some optional implementations, the first acquisition module 601 includes:

[0222] The fifth determination submodule is used to determine the third estimated state based on at least one data in the target observation data; the sixth determination submodule is used to determine the third estimated state as the target state when the third estimated state is an abnormal state; and the seventh determination submodule is used to determine the third estimated state as the first estimated state when the third estimated state is a normal state.

[0223] In some optional embodiments, the device further comprises:

[0224] The second acquisition module is used to obtain the original observation data of the indicator light and the original environmental data of the environment in which the indicator light is located before obtaining the target observation data of the indicator light and the target environmental data of the environment in which the indicator light is located; the first preprocessing module is used to preprocess the original observation data to obtain the target observation data; the second preprocessing module is used to preprocess the original environmental data to obtain the target environmental data.

[0225] An embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any of the above-mentioned embodiments of the method for monitoring the status of an indicator light.

[0226] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above-mentioned indicator light status monitoring method embodiments when running.

[0227] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0228] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any of the above-mentioned methods for monitoring the status of an indicator light are implemented.

[0229] An embodiment of the present application also provides another computer program product, including a non-volatile computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps in any of the above-mentioned indicator light status monitoring method embodiments.

[0230] Those skilled in the art will further realize that the mere concepts, teachings, and embodiments described herein are merely meant to provide an enabling description of the claimed application. Accordingly, modifications and / or additions, other than those explicitly described herein, can be obvious to those skilled in the art in the light of this disclosure. The claimed application is intended to embrace all such modifications and / or additions.

[0231] The above describes in detail a method for monitoring the state of an indicator light. The principles and implementation modes of the present application are described by using specific examples. The above description of the examples is only used to help understand the method of the present application and its core idea. It should be pointed out that, for those skilled in the art, some improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A method for monitoring the status of an indicator light, characterized in that: include: Acquire target observation data of the indicator light, target environment data of the environment in which the indicator light is located, and a predetermined first estimated state, where the first estimated state is determined based on the target observation data; Constructing a joint probability distribution among the observed variables, the environmental variables, and the indicator light status according to a pre-constructed causal relationship graph, wherein the causal relationship graph includes: the causal relationship among the observed variables, the environmental variables, and the indicator light status; Obtaining, according to the joint probability distribution, the conditional probability of the indicator light state when the observed variable is the target observation data; Based on the target observation data, the target environment data, and the conditional probability, using the first estimated state as an initial value, solving the conditional probability using a maximum a posteriori probability estimation algorithm to obtain a second estimated state when the conditional probability is maximized, and a first probability value corresponding to the second estimated state; A target state of the indicator light is determined according to the second estimated state and the first probability value.

2. The monitoring method according to claim 1, characterized in that: Determining the target state of the indicator light according to the second estimated state and the first probability value includes: When the first probability value is greater than or equal to a preset probability threshold, the second estimated state is determined as the target state.

3. The monitoring method according to claim 2, characterized in that: The determining the target state of the indicator light according to the second estimated state and the first probability value further includes: When the first probability value is less than the preset probability threshold, the causal relationship pointing to the indicator light state in the causal relationship diagram is removed to obtain a causal relationship diagram after removal; In the causal relationship diagram after elimination, for each of the observed variables, an environmental variable or other observed variable pointing to the observed variable is determined as an associated variable of the observed variable; For each of the observed variables, determining a first deviation corresponding to the observed variable based on at least two of the second estimated state, the target observed data corresponding to the observed variable, and the target associated data corresponding to the associated variable; The target state of the indicator light is determined according to the first deviation corresponding to each of the observed variables.

4. The monitoring method according to claim 3, characterized in that: The determining, based on at least two of the second estimated state, the target observation data corresponding to the observation variable, and the target associated data corresponding to the associated variable, of a first deviation corresponding to the observation variable includes: Determining, according to a determination method corresponding to the observed variable, that when the indicator light state is the second estimated state and the associated variable takes the target associated data, the observed variable takes the second probability value of the target observed data; The first degree of deviation is determined according to the second probability value.

5. The monitoring method according to claim 3, characterized in that: Determining the target state of the indicator light according to the first deviation corresponding to each of the observed variables includes: Obtaining a preset deviation threshold corresponding to each of the observed variables; For each of the observed variables, comparing the first deviation with the preset deviation threshold to obtain a comparison result; When the comparison results corresponding to the observed variables are all that the first deviation is less than or equal to the preset deviation threshold, the second estimated state is determined as the target state.

6. The monitoring method according to claim 5, characterized in that: The determining the target state of the indicator light according to the first deviation corresponding to each of the observed variables further includes: When, among the observed variables, there is at least one observed variable whose comparison result is that the first deviation is greater than the preset deviation threshold, obtaining a standard value of the associated variable corresponding to each observed variable; For each of the observed variables, determining a second deviation corresponding to the observed variable based on at least two of the second estimated state, the target observed data corresponding to the observed variable, and a standard value of the associated variable; The target state of the indicator light is determined according to the second deviation corresponding to each of the observed variables and the standard value.

7. The monitoring method according to claim 6, characterized in that: The determining, based on at least two of the second estimated state, the target observation data corresponding to the observation variable, and the standard value of the associated variable, a second deviation corresponding to the observation variable includes: Determining, according to a determination method corresponding to the observed variable, that when the indicator light state is the second estimated state and the associated variable takes the standard value, the observed variable takes the third probability value of the target observed data; The second degree of deviation is determined according to the third probability value.

8. The monitoring method according to claim 6, characterized in that: Determining the target state of the indicator light according to the second deviation corresponding to each of the observed variables and the standard value includes: For each of the observed variables, diagnose the observed variable according to the second deviation corresponding to the observed variable and the standard value of the associated variable to obtain a diagnosis result corresponding to the observed variable; When the diagnosis results corresponding to the observed variables are all normal, the second estimated state is determined as the target state.

9. The monitoring method according to claim 8, characterized in that: The causal relationship diagram also includes the causal effect strength corresponding to each causal relationship, and determining the target state of the indicator light based on the second deviation corresponding to each observed variable and the standard value further includes: When the diagnosis result corresponding to at least one observed variable among the observed variables is abnormal, obtaining the target causal effect strength corresponding to the causal relationship between the associated variable and the observed variable; For each of the observed variables, determining a weight of the observed variable according to the target causal effect strength corresponding to each of the observed variables, the target association data of each of the associated variables, the standard value of each of the associated variables, the second deviation corresponding to each of the observed variables, and a preset parameter set; Selecting a maximum value and a minimum value from the weights of each of the observed variables, and calculating a difference between the maximum value and the minimum value; When the difference is greater than a preset weight threshold, the diagnosis result of the observation variable corresponding to the maximum weight is determined as the target state; When the difference is less than the preset weight threshold, the operation is terminated and notification information indicating that the target state determination has failed is returned.

10. The monitoring method according to claim 1, characterized in that: The target observation data includes target visual data, target electrical data and target state value. The target visual data includes color probability distribution, flicker state and hue characteristic value. The target electrical data includes voltage signal after removing trend item, current signal after removing trend item and electrical anomaly index. The target environmental data includes at least one of target temperature value, target light intensity and target electrical fluctuation value.

11. The monitoring method according to claim 10, characterized in that: Obtaining a predetermined first estimated state, including: determining a third estimated state based on at least one of the target observation data; When the third estimated state is an abnormal state, determining the third estimated state as the target state; When the third estimated state is a normal state, the third estimated state is determined as the first estimated state.

12. The monitoring method according to claim 1, characterized in that: Before acquiring target observation data of the indicator light and target environmental data of the environment in which the indicator light is located, the method further includes: Acquiring original observation data of the indicator light and original environmental data of the environment in which the indicator light is located; Preprocessing the original observation data to obtain the target observation data; The original environmental data is preprocessed to obtain the target environmental data.

13. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the method for monitoring the indicator light status as claimed in any one of claims 1 to 12 when executing the computer program.

14. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of the method for monitoring the indicator light status according to any one of claims 1 to 12.

15. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for monitoring the indicator light status according to any one of claims 1 to 12 are implemented.

Images