Dam safety monitoring instrument equipment identification method for automatic image recognition in patrol inspection

Abstract

The invention discloses a dam safety monitoring instrument and equipment identification method for automatic image recognition in patrol inspection, and the method comprises the steps: obtaining an original image of a field identifier and a reading command of a field equipment variable, and carrying out the preprocessing of the original image; carrying out binarization processing on pixel points in the pixel color blocks by adopting a self-adaptive threshold value to obtain a square image, and obtaining first identification according to the on-off state of a field identifier of the square image; if the first identification is a fault state, inputting the original image into a generative adversarial network with a low light enhancement function for secondary identification to obtain second identification; retrieving a rule from a rule set according to the field device identifier and the semantic identifier to obtain an action rule, obtaining a parameter value based on the action rule, and calculating a variable value; and integrating and outputting the second identification and the variable value as an identification result. The method improves the accuracy of dam safety monitoring instrument equipment identification, and can be directly used in a monitoring instrument equipment identification system.

Description

Technical Field

[0001] The invention relates to the field of identification and marking, and in particular to an identification method for dam safety monitoring instrument equipment through automatic image recognition for patrol inspection. Background Art

[0002] By monitoring the dam, managers can understand the operating status and safety issues of the dam. In general, the safety monitoring of reservoir dams is carried out through instrument observation. Regular inspections are used to systematically grasp the main structure of the reservoir dam, the slopes on both sides and the dam foot foundation, related facilities and the surrounding environment. The operation of the dam can be judged by analyzing the observation data. During the on-site inspection, the safety monitoring personnel can initially grasp the overall safety status of the dam through the operation of the instruments and equipment, and can also promptly discover the instruments and equipment that may fail, record them and carry out maintenance and replacement in time. Usually, the safety monitoring personnel will take photos and record the status of the inspected instruments and equipment. The photos are saved as important on-site monitoring record materials, providing an important basis for subsequent work. For industry personnel and scientific researchers who are not familiar with the on-site conditions, these on-site photos can quickly help them have an overall grasp of the engineering structure, so as to facilitate the development of subsequent work. At the same time, for some old projects, due to time and conditions, there is no database and intelligent monitoring system. The on-site photos can also serve as important information to provide data support for possible system transformation in the future.

[0003] However, due to the harsh on-site environment, the limitations of handheld devices for monitoring personnel, and the fact that many instruments are buried in places that cannot be directly observed, the on-site identification of domestic dam safety monitoring instruments lacks a lot of background information in the photos taken on site, which is not conducive to the interpretation of the photos. It is necessary to add more obvious and camera-friendly identification to the instruments and equipment on site. To this end, the present invention proposes an on-site identification method for dam safety monitoring instruments and equipment that is convenient for patrol inspection and automatic image recognition. Summary of the invention

[0004] The purpose of the present invention is to provide a dam safety monitoring instrument and equipment identification method for automatic image recognition during patrol inspection.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] The present invention comprises the following steps:

[0007] Acquire an original image of the on-site identification and a read command of the on-site device variable, and pre-process the original image; the pre-processing includes decomposing the original image into a plurality of pixel blocks, graying the pixel blocks, removing noise, and enhancing contrast;

[0008] Adopting an adaptive threshold to binarize the pixel points in the pixel color block to obtain a square image, judging the convergence and dispersion state of the binary numbers from the edge to the center according to the square image, preliminarily identifying the on and off state of the on-site sign according to the convergence and dispersion state, and obtaining a first recognition according to the on and off state; the first recognition includes a fault state and a light-off state;

[0009] If the first recognition is a fault state, the original image is input into a generative adversarial network of a low-light enhancement function for secondary recognition to obtain a second recognition;

[0010] Parse the read command, identify the field device identifier and the semantic identifier, retrieve the rules from the rule set according to the field device identifier and the semantic identifier to obtain the action rules, use natural language processing technology to generate computer executable code based on the action rules, obtain parameter values ​​and calculate variable values; integrate the second identification with the variable value and output it as an identification result.

[0011] Furthermore, the method for extracting the pixel point includes:

[0012] The original image is obtained by edge detection and represented by the Zernike moment. The expression is:

[0013]

[0014] The Zernike moment of the original image is H(u,y), the pixel of the original image is (u,y), the original image is K(u,y), the order of the Zernike moment is c, the distance from the pixel to the center position in the target image is l, the angle between the pixel and the center position in the original image is φ, the conjugate matrix of the Zernike moment c-order polynomial is M, and the adjoint matrix of the conjugate matrix M is M * ;

[0015] Determine the conjugate matrix, the expression is:

[0016] M=U zw e i·z·φ

[0017] The order of the Zernike moment polynomial is w, and the z-order polynomial of the Zernike moment is U zw ;

[0018] According to the conjugate matrix, the real-valued polynomial is obtained as follows:

[0019]

[0020] The real-valued polynomial is From 0 to The integer is a, and the real-valued polynomial is used to update the Zernike moment of the original image. The expression is:

[0021]

[0022] The Zernike moment of the updated original image is Calculate the distance from the original image pixel to the center and the angle of the original image pixel relative to the center position:

[0023]

[0024] The moment of the pixel in the original image is R, the real part of the moment R is Re[R], and the imaginary part of the moment R is lm[R];

[0025] The updated Zernike moment, the distance from the original image pixel to the center, and the angle of the original image pixel relative to the center position are output as pixel extraction information.

[0026] Further, the preliminary identification of the on / off status of the on-site identification includes:

[0027] If the pixel points marked as 1 converge from the edge to the center, the preliminary identification scene is marked as a fault state; if the pixel points marked as 1 do not converge from the edge to the center and are in a scattered state, the preliminary identification scene is marked as a light-off state.

[0028] Furthermore, the method of generating an adversarial network for low-light enhancement function includes:

[0029] The original image of the fault state is judged to be low light according to the ISO value and exposure time. When the judgment result is low light, the total value of the channel pixel points of the original image is calculated:

[0030]

[0031] The spatial resolution width of the original image is N, the spatial resolution height of the original image is P, the channel is b, and the total pixel value of the b channel is The value of the b channel pixel (x, s) is I b (x,s);

[0032] Calculate the pixel average:

[0033]

[0034] The number of channels is The average pixel value is The channel is equalized and the expression is:

[0035]

[0036] The updated pixel channel value is Calculate the variance of low-light enhancement:

[0037]

[0038] The feature map index is k, the original image index is t, and the mean of the feature map index k and the original image index t is h tk , the original image with width v and height r is K tk,vr , the variance of low-light enhancement between feature map index k and original image index t is

[0039] The enhancement result is given according to the variance of low-light enhancement, and the expression is:

[0040]

[0041] The enhancement result is θ, and edge detection is used to perform secondary recognition based on the enhancement result to obtain the second recognition;

[0042] The loss function of the generated adversarial network is expressed as:

[0043]

[0044] The loss function is The discriminant network is f(·), the original image is K(x,s), and the mathematical expectation is The adjustment coefficient is σ, and the generated loss is The cycle consistency loss is

[0045] Adjust the accuracy of the generative adversarial network according to the loss function, and stop adjusting when the loss function value is lower than 0.108.

[0046] Furthermore, the action rule indicates that at least one sought physical quantity as the sought value of the variable is calculated based on one or more measured values.

[0047] Furthermore, the action rule indicates a mapping and conversion relationship between the corresponding engineering unit of the field device parameter value and the engineering unit of the sought value of the variable.

[0048] Further, the engineering units are stored as a reference to a predetermined table of engineering units.

[0049] Furthermore, the low-light enhanced generative adversarial network includes four branches for processing on-site identification images that are initially identified as fault states, and the loss function includes discrimination loss, generation loss, and cycle consistency loss.

[0050] The beneficial effects of the present invention are:

[0051] The present invention is a dam safety monitoring instrument identification method for automatic image recognition during patrol inspection. Compared with the prior art, the present invention has the following technical effects:

[0052] The present invention can improve the accuracy of the identification of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections through preprocessing, obtaining the first identification and obtaining the second identification steps, thereby improving the accuracy of the identification of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections, optimizing the identification of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections, greatly saving resources, and improving work efficiency, and can realize the intelligent identification of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections, and can identify and correct the identification of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections in real time, which is of great significance to the identification of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections, can adapt to the identification of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections of different standards, and the identification requirements of dam safety monitoring instruments and equipment through automatic image recognition of patrol inspections of different standards, and has a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flowchart of the steps of the dam safety monitoring instrument and equipment identification method for patrol inspection and automatic image recognition of the present invention;

[0054] Figure 2 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0055] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0056] The dam safety monitoring instrument and equipment identification method for patrol inspection and automatic image recognition of the present invention comprises the following steps:

[0057] like Figure 1 As shown, in this embodiment, the following steps are included:

[0058] Acquire an original image of the on-site identification and a read command of the on-site device variable, and pre-process the original image; the pre-processing includes decomposing the original image into a plurality of pixel blocks, graying the pixel blocks, removing noise, and enhancing contrast;

[0059] Adopting an adaptive threshold to binarize the pixel points in the pixel color block to obtain a square image, judging the convergence and dispersion state of the binary numbers from the edge to the center according to the square image, preliminarily identifying the on and off state of the on-site sign according to the convergence and dispersion state, and obtaining a first recognition according to the on and off state; the first recognition includes a fault state and a light-off state;

[0060] If the first recognition is a fault state, the original image is input into a generative adversarial network of a low-light enhancement function for secondary recognition to obtain a second recognition;

[0061] Parse the read command, identify the field device identifier and the semantic identifier, retrieve the rule from the rule set according to the field device identifier and the semantic identifier to obtain the action rule, generate a computer executable code based on the action rule using natural language processing technology, obtain parameter values ​​and calculate variable values; integrate the second identification with the variable value and output it as an identification result;

[0062] In the actual evaluation, Case 1: Substation Intelligent Patrol System: involves an intelligent patrol system that integrates image recognition and equipment parameter reading, which is used for intelligent patrol of substations in the power industry;

[0063] Processing steps: Use a high-definition camera to capture the image of the sign in the substation; pre-process the image through the intelligent patrol system; apply the adaptive threshold method to perform binarization processing to identify the on and off status of the sign; if the initial identification is a fault state, use the low-light enhanced generative adversarial network for secondary identification; at the same time, the system receives and parses the device variable reading command, and uses natural language processing technology to retrieve relevant action rules from the rule set; generates cross-platform computer executable code, automatically compiles and executes it to calculate the device parameter value; integrates the image recognition result with the device parameter calculation result, and outputs the final patrol report;

[0064] In the actual evaluation, Case 1 Industrial Automation Monitoring System: used in the field of industrial automation to monitor the status of the production line in real time;

[0065] Processing steps: Install sensors and cameras to capture various identifications and equipment status on the production line in real time; the system automatically obtains the original image and the read command of the equipment variable; pre-processes the image; uses the adaptive threshold method to binarize the image and identify the key identification status on the production line; for the identification initially identified as a fault, a low-light enhancement network is used for secondary identification; parses the read command, retrieves the rule set, and generates code to calculate and update the equipment parameters; integrates the image recognition results with the equipment parameters to provide real-time production status reports and alarms.

[0066] In this embodiment, the method for extracting the pixel point includes:

[0067] The original image is obtained by edge detection and represented by the Zernike moment. The expression is:

[0068]

[0069] The Zernike moment of the original image is H(u,y), the pixel of the original image is (u,y), the original image is K(u,y), the order of the Zernike moment is c, the distance from the pixel to the center position in the target image is l, the angle between the pixel and the center position in the original image is φ, the conjugate matrix of the Zernike moment c-order polynomial is M, and the adjoint matrix of the conjugate matrix M is M * ;

[0070] Determine the conjugate matrix, the expression is:

[0071] M=U zw e i·z·φ

[0072] The order of the Zernike moment polynomial is w, and the z-order polynomial of the Zernike moment is U zw ;

[0073] According to the conjugate matrix, the real-valued polynomial is obtained as follows:

[0074]

[0075] The real-valued polynomial is From 0 to The integer is a, and the real-valued polynomial is used to update the Zernike moment of the original image. The expression is:

[0076]

[0077] The Zernike moment of the updated original image is Calculate the distance from the original image pixel to the center and the angle of the original image pixel relative to the center position:

[0078]

[0079] The moment of the pixel in the original image is R, the real part of the moment R is Re[R], and the imaginary part of the moment R is lm[R];

[0080] The updated Zernike moment, the distance from the original image pixel to the center, and the angle of the original image pixel relative to the center position are output as pixel extraction information.

[0081] In this embodiment, the preliminary identification of the on / off status of the on-site identification includes:

[0082] If the pixel points marked as 1 converge from the edge to the center, the preliminary identification scene is marked as a fault state; if the pixel points marked as 1 do not converge from the edge to the center and are in a scattered state, the preliminary identification scene is marked as a light-off state.

[0083] In this embodiment, the method for generating an adversarial network for low-light enhancement function includes:

[0084] The original image of the fault state is judged to be low light according to the ISO value and exposure time. When the judgment result is low light, the total value of the channel pixel points of the original image is calculated:

[0085]

[0086] The spatial resolution width of the original image is N, the spatial resolution height of the original image is P, the channel is b, and the total pixel value of the b channel is The value of the b channel pixel (x, s) is I b (x,s);

[0087] Calculate the pixel average:

[0088]

[0089] The number of channels is The average pixel value is The channel is equalized and the expression is:

[0090]

[0091] The updated pixel channel value is Calculate the variance of low-light enhancement:

[0092]

[0093] The feature map index is k, the original image index is t, and the mean of the feature map index k and the original image index t is h tk , the original image with width v and height r is K tk,vr , the variance of low-light enhancement between feature map index k and original image index t is

[0094] The enhancement result is given according to the variance of low-light enhancement, and the expression is:

[0095]

[0096] The enhancement result is θ, and edge detection is used to perform secondary recognition based on the enhancement result to obtain the second recognition;

[0097] The loss function of the generated adversarial network is expressed as:

[0098]

[0099] The loss function is The discriminant network is f(·), the original image is K(x,s), and the mathematical expectation is The adjustment coefficient is σ, and the generated loss is The cycle consistency loss is

[0100] Adjust the accuracy of the generative adversarial network according to the loss function, and stop adjusting when the loss function value is lower than 0.108.

[0101] In this embodiment, the action rule indicates that at least one sought physical quantity is calculated as the sought value of the variable from one or more measured values.

[0102] In this embodiment, the action rule indicates a mapping and conversion relationship between the corresponding engineering unit of the field device parameter value and the engineering unit of the sought value of the variable.

[0103] In this embodiment, the engineering units are stored as a reference to a predetermined table of engineering units.

[0104] In this embodiment, the low-light enhanced generative adversarial network includes four branches for processing the scene identification image that is initially identified as a fault state, and the loss function includes the discriminant loss, the generation loss and the cycle consistency loss.

[0105] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0106] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0107] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0108] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a user intention recognition device for a multi-dimensional human-computer interaction scene at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the aforementioned user intention recognition methods for a multi-dimensional human-computer interaction scene.

[0109] The above application Figure 1 The user intention recognition method for a multi-dimensional human-computer interaction scenario disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in a decoding processor. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0110] The electronic device may also perform Figure 1 A user intention recognition method for multi-dimensional human-computer interaction scenarios is implemented Figure 1 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0111] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple applications, they execute any of the user intention recognition methods for the aforementioned multi-dimensional human-computer interaction scenarios.

[0112] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0113] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0116] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0117] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0118] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0119] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0120] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0121] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying dam safety monitoring instruments and equipment by automatic image recognition during patrol inspection, characterized in that: The following steps are involved: Acquire an original image of the on-site identification and a read command of on-site device variables, and pre-process the original image; The preprocessing includes decomposing the original image into a plurality of pixel blocks, graying the pixel blocks, removing noise and enhancing contrast; Adopting an adaptive threshold to binarize the pixel points in the pixel color block to obtain a square image, judging the convergence and dispersion state of the binary numbers from the edge to the center according to the square image, preliminarily identifying the on and off state of the on-site sign according to the convergence and dispersion state, and obtaining a first recognition according to the on and off state; the first recognition includes a fault state and a light-off state; If the first recognition is a fault state, the original image is input into a generative adversarial network of a low-light enhancement function for secondary recognition to obtain a second recognition; Parse the read command, identify the field device identifier and the semantic identifier, retrieve the rule from the rule set according to the field device identifier and the semantic identifier to obtain the action rule, generate a computer executable code based on the action rule by using natural language processing technology, obtain parameter values ​​and calculate variable values; The second identification is integrated with the variable value and output as an identification result.

2. According to the method for identifying dam safety monitoring instruments and equipment by patrol inspection and automatic image recognition according to claim 1, it is characterized in that: The method for extracting the pixel point comprises: The original image is obtained by edge detection and represented by the Zernike moment. The expression is: The Zernike moment of the original image is H(u,y), the pixel of the original image is (u,y), the original image is K(u,y), the order of the Zernike moment is c, the distance from the pixel to the center position in the target image is l, the angle between the pixel and the center position in the original image is φ, the conjugate matrix of the Zernike moment c-order polynomial is M, and the adjoint matrix of the conjugate matrix M is M * ; Determine the conjugate matrix, the expression is: M=U zw have been i·z·φ The order of the Zernike moment polynomial is w, and the z-order polynomial of the Zernike moment is U zw ; According to the conjugate matrix, the real-valued polynomial is obtained as follows: The real-valued polynomial is From 0 to The integer is a, and the real-valued polynomial is used to update the Zernike moment of the original image. The expression is: The Zernike moment of the updated original image is Calculate the distance from the original image pixel to the center and the angle of the original image pixel relative to the center position: The moment of the pixel in the original image is R, the real part of the moment R is Re[R], and the imaginary part of the moment R is lm[R]; The updated Zernike moment, the distance from the original image pixel to the center, and the angle of the original image pixel relative to the center position are output as pixel extraction information.

3. The on-site identification method for dam safety monitoring equipment that facilitates patrol inspection and automatic image recognition according to claim 1 is characterized in that: The initial identification of the on / off status of the on-site identification includes: If the pixel points marked as 1 converge from the edge to the center, the preliminary identification scene is marked as a fault state; if the pixel points marked as 1 do not converge from the edge to the center and are in a scattered state, the preliminary identification scene is marked as a light-off state.

4. According to the method for identifying dam safety monitoring instruments and equipment by patrol inspection and automatic image recognition according to claim 1, it is characterized in that: The method of generating an adversarial network for low-light enhancement function comprises: The original image of the fault state is judged to be low light according to the ISO value and exposure time. When the judgment result is low light, the total value of the channel pixel points of the original image is calculated: The spatial resolution width of the original image is N, the spatial resolution height of the original image is P, the channel is b, and the total pixel value of the b channel is The value of the b channel pixel (x, s) is I b (x,s); Calculate the pixel average: The number of channels is The pixel average is The channel is equalized and the expression is: The updated pixel channel value is Calculate the variance of low-light enhancement: The feature map index is k, the original image index is t, and the mean of the feature map index k and the original image index t is h tk , the original image with width v and height r is K tk,vr , the variance of low-light enhancement between feature map index k and original image index t is The enhancement result is given according to the variance of low-light enhancement, and the expression is: The enhancement result is θ, and edge detection is used to perform secondary recognition based on the enhancement result to obtain the second recognition; The loss function of the generated adversarial network is expressed as: The loss function is The discriminant network is f(·), the original image is K(x,s), and the mathematical expectation is The adjustment coefficient is σ, and the generated loss is The cycle consistency loss is Adjust the accuracy of the generative adversarial network according to the loss function, and stop adjusting when the loss function value is lower than 0.

108.

5. The dam safety monitoring instrument and equipment identification method for patrol inspection and automatic image recognition according to claim 1 is characterized in that: The action rule indicates the calculation of at least one sought physical quantity as the sought value of the variable from one or more measured values.

6. The dam safety monitoring instrument and equipment identification method for patrol inspection and automatic image recognition according to claim 1 is characterized in that: The action rule indicates a mapping and conversion relationship between the corresponding engineering units of the field device parameter value and the engineering units of the sought value of the variable.

7. The dam safety monitoring instrument and equipment identification method of patrol inspection automatic image recognition according to claim 6 is characterized in that: The engineering units are stored as a reference to a predetermined table of engineering units.

8. The dam safety monitoring instrument and equipment identification method for patrol inspection and automatic image recognition according to claim 1 is characterized in that: The low-light enhanced generative adversarial network includes four branches for processing scene identification images that are initially identified as fault states, and the loss function includes discrimination loss, generation loss and cycle consistency loss.

Images