Detection system and detection method for intelligent instruments and meters
Through the design of intelligent instrument detection system, image acquisition and visual detection analysis are used to identify instrument readings, the problem that existing systems cannot recognize pointer instruments is solved, and the rapid and efficient detection of various instruments and meters is achieved, which improves detection efficiency .
Patent Information
- Application Number
- CN202411230237.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing instrument detection system cannot effectively identify pointer instruments, and has a narrow application range and cannot widely and effectively test a variety of instruments and instruments.
An intelligent instrument detection system was designed, including an environment creation module, image acquisition module, visual detection and analysis module, working parameter monitoring module, comprehensive safety factor analysis and evaluation module, communication module and fault warning module, through image acquisition and visual detection Analyze and identify the instrument readings, compare them with the working parameters, evaluate the instrument performance, and issue a fault warning.
It realizes fast and efficient detection of multiple instruments and meters, improves detection efficiency, and can simultaneously and effectively detect multiple instruments and meters in the system, suitable for systems that use pointer and digital instruments.
Smart Images

Figure CN119022982A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of instrument detection, and in particular relates to an intelligent instrument detection system and a detection method. Background Art
[0002] With the advancement of industrial intelligence, more and more work in the industry is replaced by machines and equipment, but machines and equipment do not have autonomous capabilities. If there is a problem with the operation of the machine, these machines will not actively report the problem like people do. Operation and maintenance technicians are required to discover the problems of these machines in time. Operation and maintenance technicians mainly judge the machine status through the instruments installed on the machine. In some large machine systems, multiple instruments are often installed. Some machine points use traditional pointer instruments, and some machine points use digital display instruments instead of pointer instruments. Due to the large number of instruments on the system, it is difficult for operation and maintenance technicians to detect the instruments quickly and effectively.
[0003] In the Chinese patent with publication number CN104618035B, a system and method for testing test instruments are mentioned, including: an environment creation unit, a data processing unit, a data transmission unit, a data storage unit and an intelligent terminal; the method includes: step 1: the environment creation unit creates different test environments and detects the test instruments; step 2: the data processing unit receives the detection results transmitted by the environment creation unit and converts the detection results into data; step 3: the data transmission unit transmits the data converted by the data processing unit to the data storage unit and the intelligent terminal; step 4: the data storage unit stores the data converted by the data processing unit; step 5: the intelligent terminal analyzes the data converted by the data processing unit. The technical problems that cannot be achieved by the prior art, such as batch testing of test instruments, saving time and labor, thereby improving work efficiency, reducing the time consumed by switching back and forth between test environments, and detecting the relevant performance of test instruments are solved.
[0004] During the application process, the applicant discovered that the above application had defects during use. The instrument test system disclosed in the application mainly focused on measuring the performance of intelligent instruments. However, the instrument systems of many machines mixed pointer instruments and digital instruments. Only using this test system, some pointer instruments could not be effectively identified, which resulted in a narrow application scope of the test system and the inability to widely and effectively test a variety of instruments. Summary of the invention
[0005] The purpose of the present invention is to provide a detection system and a detection method for intelligent instruments and meters to solve the problems raised in the above background technology.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] An intelligent instrument detection system, specifically comprising:
[0008] An environment creation module, which is used to create different test environments to facilitate placing instruments and meters in different test environments for testing;
[0009] An image acquisition module, which is used to acquire image information of the instrument to be detected;
[0010] A visual inspection and analysis module, the visual inspection and analysis module is used to obtain and analyze the reading information contained in the image information collected by the image acquisition module;
[0011] A working parameter monitoring module, which is used to monitor the output parameters of the comparison test environment and the analysis parameters of the visual detection analysis module;
[0012] A comprehensive safety factor analysis and evaluation module, which is used to evaluate and judge the performance of the instrument based on the analysis results of the visual detection analysis module and the monitoring comparison results of the working parameter monitoring module;
[0013] A communication module, the communication module is used to communicate with the terminal to transmit the analysis and evaluation results;
[0014] A fault warning module is used to issue a fault warning prompt according to the analysis and evaluation results.
[0015] Preferably, the environment creation module is connected to the image acquisition module and the working parameter monitoring module, the image acquisition module is connected to the visual detection and analysis module, the working parameter monitoring module is connected to the visual detection and analysis module, the working parameter monitoring module and the visual detection and analysis module are connected to the comprehensive safety factor analysis and evaluation module, and the comprehensive safety factor analysis and evaluation module is connected to the communication module and the fault warning module.
[0016] Preferably, the hardware part of the environment creation module includes a chassis, a radiator, an auxiliary controller, an ARI NC429 module, a 1553B module and an RS232 / 422 / 485 serial port module.
[0017] Preferably, the image acquisition module includes a visual inspection box, a multi-degree-of-freedom positioning fixture and an industrial camera.
[0018] Preferably, the environment creation module and the image acquisition module are both connected to a computer terminal, and the visual detection and analysis module, the working parameter monitoring module, the comprehensive safety factor analysis and evaluation module, the communication module and the fault warning module are all placed inside the computer terminal.
[0019] Preferably, the visual detection and analysis module detection step specifically includes:
[0020] S1, the image acquisition module transmits the acquired instrument image to the visual inspection and analysis module, which analyzes it;
[0021] S2, the visual inspection and analysis module has built-in pointer instrument reading recognition algorithms and digital instrument reading recognition algorithms, which perform pointer instrument reading recognition and digital instrument reading recognition on instrument images respectively;
[0022] S3, passing the identified readings to the comprehensive safety factor analysis and evaluation module for analysis.
[0023] Preferably, in S2, the pointer instrument reading recognition algorithm uses the yo l ov5s model to identify the pointer instrument from the captured image, then uses the yo l ov8x-pose model to detect the scale lines and key points of the pointer in the instrument, then uses the DBNetpp model to detect the value box and uses the SATRN model to perform text recognition, and finally performs post-processing to obtain the reading result.
[0024] Preferably, when the yo l ov8x-pose model identifies the pointer key points, the first and last points of the pointer are taken as the key points, and the rectangular frame composed of the key points is taken as the target detection frame. When identifying the scale lines of the pointer instrument, each numerical frame is regarded as the same category for target detection, and each numerical frame carries the corresponding scale points as numerical key points. When post-processing, opencv is used to post-process the SATRN model recognition output, distinguish the inner and outer diameters of the pointer instrument, determine the order of the scale points, and correct the OCR recognition results, and judge whether the recognition results meet the perspective transformation. If the perspective transformation is met, the points blocked by the pointer are supplemented and the readings are calculated. If the perspective transformation is not met, the perspective transformation is performed, and then the blocked points are supplemented and the readings are calculated.
[0025] Preferably, in S2, the digital meter reading recognition algorithm identifies the position of the dial from the acquired image through target detection, segments the dial position, further identifies the numbers in the dial through target detection, and intelligently reads the numbers in the dial based on the identified dial numbers.
[0026] An intelligent instrument detection method is applied to an intelligent instrument detection system, specifically comprising the following steps:
[0027] S4, creating an instrumentation simulation working environment, outputting a working environment simulation signal, and transmitting the working environment simulation signal to the instrumentation under test;
[0028] S5, the instrument under test receives the working environment simulation signal and displays the instrument reading according to the simulation signal;
[0029] S6, capturing instrument readings, and comparing the instrument readings with the working environment simulation signal to determine the accuracy of the instrument readings;
[0030] S7, the comprehensive safety factor analysis and evaluation module outputs a fault analysis report based on the judgment result, reports a fault warning, gives an explanation and a solution report, and transmits the report to the communication terminal.
[0031] Compared with the prior art, the present invention has the following beneficial effects:
[0032] The present invention sets an environment creation module, an image acquisition module, a visual detection and analysis module, a working parameter monitoring module, a comprehensive safety factor analysis and evaluation module, a communication module and a fault warning module. When performing instrument detection, the instrument is connected to the detection system, and the environment creation environment builds a virtual system working environment. The instrument is used to perceive and reflect the operating parameters of the virtual system working environment. The visual detection and analysis module obtains the display screen of the instrument and obtains the reading of the instrument according to the display screen. Then, the displayed reading is compared with the virtual system working environment parameters identified by the working parameter monitoring module. The comprehensive safety factor analysis and evaluation module determines the working status of the instrument. Through digital identification and detection, multiple instruments can be detected quickly and efficiently, so that a detection and operation and maintenance technician can effectively detect multiple instruments of the system at the same time, which helps to improve the detection efficiency of the detection technician. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 is a block diagram of the detection system of the present invention;
[0034] Figure 2 This is a block diagram of the detection method of the present invention. DETAILED DESCRIPTION
[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0036] Embodiment 1:
[0037] See also Figure 1 As shown, an intelligent instrument detection system specifically includes:
[0038] An environment creation module, which is used to create different test environments to facilitate placing instruments and meters in different test environments for testing;
[0039] An image acquisition module, which is used to acquire image information of the instrument to be detected;
[0040] A visual inspection and analysis module, the visual inspection and analysis module is used to obtain and analyze the reading information contained in the image information collected by the image acquisition module;
[0041] A working parameter monitoring module, which is used to monitor the output parameters of the comparison test environment and the analysis parameters of the visual detection analysis module;
[0042] A comprehensive safety factor analysis and evaluation module, which is used to evaluate and judge the performance of the instrument based on the analysis results of the visual detection analysis module and the monitoring comparison results of the working parameter monitoring module;
[0043] A communication module, the communication module is used to communicate with the terminal to transmit the analysis and evaluation results;
[0044] A fault warning module is used to issue a fault warning prompt according to the analysis and evaluation results.
[0045] As can be seen from the above, the environment creation module creates a test environment and transmits the created test environment signal to the instrument. The image acquisition module captures the changing image information of the instrument, obtains high-definition instrument images, and transmits the images to the visual inspection and analysis module for inspection and analysis. The visual inspection and analysis module determines the readings of the instrument, and compares the displayed readings with the test environment values monitored by the working parameter monitoring module. The comprehensive safety factor analysis and evaluation module evaluates the comparison results to determine whether there are errors in the displayed values. The fault warning module can have a built-in instrument display database to match the instrument display values with the fault conditions. When relevant fault conditions occur, the relevant fault conditions and fault solutions are output, and the communication module transmits the evaluation results to the terminal.
[0046] The environment creation module is connected to the image acquisition module and the working parameter monitoring module, the image acquisition module is connected to the visual detection and analysis module, the working parameter monitoring module is connected to the visual detection and analysis module, the working parameter monitoring module and the visual detection and analysis module are connected to the comprehensive safety factor analysis and evaluation module, and the comprehensive safety factor analysis and evaluation module is connected to the communication module and the fault warning module.
[0047] The hardware part of the environment creation module includes a chassis, a radiator, an auxiliary controller, an ARI NC429 module, a 1553B module and an RS232 / 422 / 485 serial port module.
[0048] The chassis has built-in PCI cards and multiple slots, and can use DC fans and specially designed ventilation slots in a straight-line air duct to increase the chassis's heat dissipation capacity. The controller uses existing chipsets, memory, and hard disks. The ARI NC429 module sets two channels for sending and receiving, supports two sending modes, FIFO and Schedule, two receiving filter modes, SDI and Label, and supports Time Stamping, interrupts, and external triggers. The 1533B module complies with MI L-STD-1553B protocol, with dual channels and BC, RT and CM functions, supports 1M / 2M / 4Mbps and user-defined baud rate, each channel is dual redundant A and B channels, supports time stamp, resolution 1μs, supports self-test and error injection, supports interrupt and external trigger, RS232 / 422 / 485 serial port module has 8 channels of RS232 / 422 / 485, the baud rate is customized according to user needs, supports two data receiving modes: protocol reception and transparent reception, and the receiving frame header can be set when receiving the protocol.
[0049] The image acquisition module includes a visual inspection box, a multi-degree-of-freedom positioning fixture and an industrial camera. The industrial camera is fixed on the multi-degree-of-freedom positioning fixture, which is convenient for users to adjust the direction of the industrial camera so that the industrial camera can be aimed at instruments to accurately obtain instrument readings.
[0050] The environment creation module and the image acquisition module are both connected to a computer terminal, and the visual detection and analysis module, the working parameter monitoring module, the comprehensive safety factor analysis and evaluation module, the communication module and the fault warning module are all placed inside the computer terminal.
[0051] The visual inspection and analysis module detection steps specifically include:
[0052] S1, the image acquisition module transmits the acquired instrument image to the visual inspection and analysis module, which analyzes it;
[0053] S2, the visual inspection and analysis module has built-in pointer instrument reading recognition algorithms and digital instrument reading recognition algorithms, which perform pointer instrument reading recognition and digital instrument reading recognition on instrument images respectively;
[0054] S3, passing the identified readings to the comprehensive safety factor analysis and evaluation module for analysis.
[0055] In S2, the pointer instrument reading recognition algorithm uses the yo l ov5s model to identify the pointer instrument from the captured image, then uses the yo l ov8x-pose model to detect the scale lines and key points of the pointer in the instrument, then uses the DBNetpp model to detect the value box and uses the SATRN model for text recognition, and finally performs post-processing to obtain the reading result.
[0056] When the yo l ov8x-pose model recognizes the pointer key points, the first and last points of the pointer are taken as the key points, and the rectangular frame composed of the key points is taken as the target detection frame. If the rectangular frame composed of the two points is too narrow, the rectangular frame can be appropriately extended. Since the number of numerical frames of each instrument is different, when identifying the scale lines of the pointer instrument, each numerical frame is regarded as the same category for target detection, and each numerical frame carries the corresponding scale point as the numerical key point. Text detection based on DBNetpp is selected. A multi-level feature aggregation module is introduced on the basis of the DBNet model. It consists of stage attention and spatial attention sub-modules, which strengthens the fusion of features of different scales and improves the robustness of processing images of different sizes. The SATRN model uses the se l f-attent ion mechanism to model the 2D spatial relationship of characters, and introduces convolutional layers in the FFN module to enhance the model's ability to capture global and local features. The SATRN model has sufficient recognition ability for texts with large curvature and large angle rotation, and can achieve SOTA on multiple irregular text datasets.
[0057] During post-processing, opencv is used to post-process the SATRN model recognition output, distinguish the inner and outer diameters of the pointer instrument, determine the order of the scale points and correct the OCR recognition results. The inner diameter and outer diameter are distinguished according to the distance between the scale point and the value box and the pointer origin, and subsequent processing is performed separately. The angle difference between the starting point and the end point of the scale point is the largest angle difference between two adjacent points. Only the largest angle difference needs to be found to determine the starting point. First, the scale point coordinates are converted into polar coordinates centered on the pointer origin, and the angle difference between two adjacent scale points is calculated. The next scale point with the largest difference is selected as the starting point, and the scale data order is adjusted to determine whether the recognition result satisfies the perspective transformation. If the perspective transformation is satisfied, the points blocked by the pointer are supplemented. And calculate the readings. When the perspective transformation is not satisfied, after the perspective transformation, the occlusion points are supplemented and the readings are calculated. When performing the perspective transformation, the coordinates of the last point are first determined. For example, it is known that the polar coordinate angle of the last point after the perspective transformation is 45°. Assuming that the distance between all scale points and the center of the circle is R, the coordinates of the last point can be expressed as (45°, R). The coordinates in the Cartesian coordinate system can be obtained based on the polar coordinates, and the coordinates of the other three points can be determined. For example, there are 6 intervals that bisect the entire 270° range, so the angle of each scale is 45°. Based on this, the coordinates of the other three points after the change can be calculated in turn. In this way, the coordinates of the four scale points in the image after the perspective transformation can be determined, and these coordinates can be used for perspective transformation.
[0058] When filling in the points blocked by the pointer, first determine the front and rear points of the missing point, as well as their average distance from the center of the circle. This information will be used to infer the polar coordinates of the missing point. Then, based on the angle and average distance of the front and rear points, calculate the polar coordinates of the missing point. The polar coordinates are composed of the polar diameter and the polar angle. The polar diameter represents the distance from the point to the circle, and the polar angle represents the angle of the point in the polar coordinate system. Convert the polar coordinates of the missing point to Cartesian coordinates. If a perspective transformation is performed, apply the converted coordinates to the perspective matrix to obtain the coordinates in the original image.
[0059] When calculating the reading, first determine the positions of the front and rear scale points of the pointer, and record their angle values in the polar coordinate system. Assume that the angle of the front scale point is θ1, the angle of the rear scale point is θ2, and the angle of the pointer is θr. Assume that the reading of the front scale point of the pointer is r1, and the reading of the rear scale point is r2. The pointer reading is r=r1+Δr*Δθ, where Δr=r2-r1. Through perspective transformation, the deformed instrument image is corrected, and the meter readings can be read more accurately
[0060] In S2, the digital meter reading recognition algorithm is to identify the position of the dial from the acquired image through target detection, segment the dial position, further identify the numbers in the dial through target detection, and intelligently read the numbers in the dial based on the identified dial numbers.
[0061] When building a digital instrument reading recognition algorithm, first collect multiple digital instrument display images, build a training data set, load the pre-trained model yo l ov8n.pt, use the training data set to train the model, calculate the dial position and displayed numbers, and use the training model to quickly identify the digital instrument readings.
[0062] Embodiment 2:
[0063] See also Figure 2 As shown, an intelligent instrument detection method specifically includes the following steps:
[0064] S4, creating an instrumentation simulation working environment, outputting a working environment simulation signal, and transmitting the working environment simulation signal to the instrumentation under test;
[0065] S5, the instrument under test receives the working environment simulation signal and displays the instrument reading according to the simulation signal;
[0066] S6, capturing instrument readings, and comparing the instrument readings with the working environment simulation signal to determine the accuracy of the instrument readings;
[0067] S7, the comprehensive safety factor analysis and evaluation module outputs a fault analysis report based on the judgment result, reports a fault warning, gives an explanation and a solution report, and transmits the report to the communication terminal.
[0068] Embodiment three:
[0069] See also Figure 1 - Figure 2 As shown, an intelligent instrument detection system specifically includes:
[0070] An environment creation module, which is used to create different test environments to facilitate placing instruments and meters in different test environments for testing;
[0071] An image acquisition module, which is used to acquire image information of the instrument to be detected;
[0072] A visual inspection and analysis module, the visual inspection and analysis module is used to obtain and analyze the reading information contained in the image information collected by the image acquisition module;
[0073] A working parameter monitoring module, which is used to monitor the output parameters of the comparison test environment and the analysis parameters of the visual detection analysis module;
[0074] A comprehensive safety factor analysis and evaluation module, which is used to evaluate and judge the performance of the instrument based on the analysis results of the visual detection analysis module and the monitoring comparison results of the working parameter monitoring module;
[0075] A communication module, the communication module is used to communicate with the terminal to transmit the analysis and evaluation results;
[0076] A fault warning module is used to issue a fault warning prompt according to the analysis and evaluation results.
[0077] As can be seen from the above, the environment creation module creates a test environment and transmits the created test environment signal to the instrument. The image acquisition module captures the changing image information of the instrument, obtains high-definition instrument images, and transmits the images to the visual inspection and analysis module for inspection and analysis. The visual inspection and analysis module determines the readings of the instrument, and compares the displayed readings with the test environment values monitored by the working parameter monitoring module. The comprehensive safety factor analysis and evaluation module evaluates the comparison results to determine whether there are errors in the displayed values. The fault warning module can have a built-in instrument display database to match the instrument display values with the fault conditions. When relevant fault conditions occur, the relevant fault conditions and fault solutions are output, and the communication module transmits the evaluation results to the terminal.
[0078] The environment creation module is connected to the image acquisition module and the working parameter monitoring module, the image acquisition module is connected to the visual detection and analysis module, the working parameter monitoring module is connected to the visual detection and analysis module, the working parameter monitoring module and the visual detection and analysis module are connected to the comprehensive safety factor analysis and evaluation module, and the comprehensive safety factor analysis and evaluation module is connected to the communication module and the fault warning module.
[0079] The hardware part of the environment creation module includes a chassis, a radiator, an auxiliary controller, an ARI NC429 module, a 1553B module and an RS232 / 422 / 485 serial port module.
[0080] The chassis has built-in PCI cards and multiple slots, and can use DC fans and specially designed ventilation slots in a straight-line air duct to increase the chassis's heat dissipation capacity. The controller uses existing chipsets, memory, and hard disks. The ARI NC429 module sets two channels for sending and receiving, supports two sending modes, FIFO and Schedule, two receiving filter modes, SDI and Label, and supports Time Stamping, interrupts, and external triggers. The 1533B module complies with MI L-STD-1553B protocol, with dual channels and BC, RT and CM functions, supports 1M / 2M / 4Mbps and user-defined baud rate, each channel is dual redundant A and B channels, supports time stamp, resolution 1μs, supports self-test and error injection, supports interrupt and external trigger, RS232 / 422 / 485 serial port module has 8 channels of RS232 / 422 / 485, the baud rate is customized according to user needs, supports two data receiving modes: protocol reception and transparent reception, and the receiving frame header can be set when receiving the protocol.
[0081] The image acquisition module includes a visual inspection box, a multi-degree-of-freedom positioning fixture and an industrial camera. The industrial camera is fixed on the multi-degree-of-freedom positioning fixture, which is convenient for users to adjust the direction of the industrial camera so that the industrial camera can be aimed at instruments to accurately obtain instrument readings.
[0082] The environment creation module and the image acquisition module are both connected to a computer terminal, and the visual detection and analysis module, the working parameter monitoring module, the comprehensive safety factor analysis and evaluation module, the communication module and the fault warning module are all placed inside the computer terminal.
[0083] The visual inspection and analysis module detection steps specifically include:
[0084] S1, the image acquisition module transmits the acquired instrument image to the visual inspection and analysis module, which analyzes it;
[0085] S2, the visual inspection and analysis module has built-in pointer instrument reading recognition algorithms and digital instrument reading recognition algorithms, which perform pointer instrument reading recognition and digital instrument reading recognition on instrument images respectively;
[0086] S3, passing the identified readings to the comprehensive safety factor analysis and evaluation module for analysis.
[0087] In S2, the pointer instrument reading recognition algorithm uses the yo l ov5s model to identify the pointer instrument from the captured image, then uses the yo l ov8x-pose model to detect the scale lines and key points of the pointer in the instrument, then uses the DBNetpp model to detect the value box and uses the SATRN model for text recognition, and finally performs post-processing to obtain the reading result.
[0088] When the yo l ov8x-pose model recognizes the pointer key points, the first and last points of the pointer are taken as the key points, and the rectangular frame composed of the key points is taken as the target detection frame. If the rectangular frame composed of the two points is too narrow, the rectangular frame can be appropriately extended. Since the number of numerical frames of each instrument is different, when identifying the scale lines of the pointer instrument, each numerical frame is regarded as the same category for target detection, and each numerical frame carries the corresponding scale point as the numerical key point. Text detection based on DBNetpp is selected. A multi-level feature aggregation module is introduced on the basis of the DBNet model. It consists of stage attention and spatial attention sub-modules, which strengthens the fusion of features of different scales and improves the robustness of processing images of different sizes. The SATRN model uses the se l f-attent ion mechanism to model the 2D spatial relationship of characters, and introduces convolutional layers in the FFN module to enhance the model's ability to capture global and local features. The SATRN model has sufficient recognition ability for texts with large curvature and large angle rotation, and can achieve SOTA on multiple irregular text datasets.
[0089] During post-processing, opencv is used to post-process the SATRN model recognition output, distinguish the inner and outer diameters of the pointer instrument, determine the order of the scale points and correct the OCR recognition results. The inner diameter and outer diameter are distinguished according to the distance between the scale point and the value box and the pointer origin, and subsequent processing is performed separately. The angle difference between the starting point and the end point of the scale point is the largest angle difference between two adjacent points. Only the largest angle difference needs to be found to determine the starting point. First, the scale point coordinates are converted into polar coordinates centered on the pointer origin, and the angle difference between two adjacent scale points is calculated. The next scale point with the largest difference is selected as the starting point, and the scale data order is adjusted to determine whether the recognition result satisfies the perspective transformation. If the perspective transformation is satisfied, the points blocked by the pointer are supplemented. And calculate the readings. When the perspective transformation is not satisfied, after the perspective transformation, the occlusion points are supplemented and the readings are calculated. When performing the perspective transformation, the coordinates of the last point are first determined. For example, it is known that the polar coordinate angle of the last point after the perspective transformation is 45°. Assuming that the distance between all scale points and the center of the circle is R, the coordinates of the last point can be expressed as (45°, R). The coordinates in the Cartesian coordinate system can be obtained based on the polar coordinates, and the coordinates of the other three points can be determined. For example, there are 6 intervals that bisect the entire 270° range, so the angle of each scale is 45°. Based on this, the coordinates of the other three points after the change can be calculated in turn. In this way, the coordinates of the four scale points in the image after the perspective transformation can be determined, and these coordinates can be used for perspective transformation.
[0090] When filling in the points blocked by the pointer, first determine the front and rear points of the missing point, as well as their average distance from the center of the circle. This information will be used to infer the polar coordinates of the missing point. Then, based on the angle and average distance of the front and rear points, calculate the polar coordinates of the missing point. The polar coordinates are composed of the polar diameter and the polar angle. The polar diameter represents the distance from the point to the circle, and the polar angle represents the angle of the point in the polar coordinate system. Convert the polar coordinates of the missing point to Cartesian coordinates. If a perspective transformation is performed, apply the converted coordinates to the perspective matrix to obtain the coordinates in the original image.
[0091] When calculating the reading, first determine the positions of the front and rear scale points of the pointer, and record their angle values in the polar coordinate system. Assume that the angle of the front scale point is θ1, the angle of the rear scale point is θ2, and the angle of the pointer is θr. Assume that the reading of the front scale point of the pointer is r1, and the reading of the rear scale point is r2. The pointer reading is r=r1+Δr*Δθ, where Δr=r2-r1. Through perspective transformation, the deformed instrument image is corrected, and the meter readings can be read more accurately
[0092] In S2, the digital meter reading recognition algorithm is to identify the position of the dial from the acquired image through target detection, segment the dial position, further identify the numbers in the dial through target detection, and intelligently read the numbers in the dial based on the identified dial numbers.
[0093] When building a digital instrument reading recognition algorithm, first collect multiple digital instrument display images, build a training data set, load the pre-trained model yo l ov8n.pt, use the training data set to train the model, calculate the dial position and displayed numbers, and use the training model to quickly identify the digital instrument readings.
[0094] An intelligent instrument detection method is applied to an intelligent instrument detection system, specifically comprising the following steps:
[0095] S4, creating an instrumentation simulation working environment, outputting a working environment simulation signal, and transmitting the working environment simulation signal to the instrumentation under test;
[0096] S5, the instrument under test receives the working environment simulation signal and displays the instrument reading according to the simulation signal;
[0097] S6, capturing instrument readings, and comparing the instrument readings with the working environment simulation signal to determine the accuracy of the instrument readings;
[0098] S7, the comprehensive safety factor analysis and evaluation module outputs a fault analysis report based on the judgment result, reports a fault warning, gives an explanation and a solution report, and transmits the report to the communication terminal.
Claims
1. An intelligent instrument detection system, characterized in that: Specifically include: An environment creation module, which is used to create different test environments to facilitate placing instruments and meters in different test environments for testing; An image acquisition module, which is used to acquire image information of the instrument to be detected; A visual inspection and analysis module, the visual inspection and analysis module is used to obtain and analyze the reading information contained in the image information collected by the image acquisition module; A working parameter monitoring module, which is used to monitor the output parameters of the comparison test environment and the analysis parameters of the visual detection analysis module; A comprehensive safety factor analysis and evaluation module, which is used to evaluate and judge the performance of the instrument based on the analysis results of the visual detection analysis module and the monitoring comparison results of the working parameter monitoring module; A communication module, the communication module is used to communicate with the terminal to transmit the analysis and evaluation results; A fault warning module is used to issue a fault warning prompt according to the analysis and evaluation results.
2. The intelligent instrument detection system according to claim 1, characterized in that: The environment creation module is connected to the image acquisition module and the working parameter monitoring module, the image acquisition module is connected to the visual detection and analysis module, the working parameter monitoring module is connected to the visual detection and analysis module, the working parameter monitoring module and the visual detection and analysis module are connected to the comprehensive safety factor analysis and evaluation module, and the comprehensive safety factor analysis and evaluation module is connected to the communication module and the fault warning module.
3. The intelligent instrument detection system according to claim 1, characterized in that: The hardware part of the environment creation module includes a chassis, a radiator, an auxiliary controller, an ARINC429 module, a 1553B module and an RS232 / 422 / 485 serial port module.
4. The intelligent instrument detection system according to claim 3 is characterized in that: The image acquisition module includes a visual inspection box, a multi-degree-of-freedom positioning fixture and an industrial camera.
5. The intelligent instrument detection system according to claim 4 is characterized in that: The environment creation module and the image acquisition module are both connected to a computer terminal, and the visual detection and analysis module, the working parameter monitoring module, the comprehensive safety factor analysis and evaluation module, the communication module and the fault warning module are all placed inside the computer terminal.
6. The intelligent instrument detection system according to claim 5, characterized in that: The visual inspection and analysis module detection steps specifically include: S1, the image acquisition module transmits the acquired instrument image to the visual inspection and analysis module, which analyzes it; S2, the visual inspection and analysis module has built-in pointer instrument reading recognition algorithms and digital instrument reading recognition algorithms, which perform pointer instrument reading recognition and digital instrument reading recognition on instrument images respectively; S3, passing the identified readings to the comprehensive safety factor analysis and evaluation module for analysis.
7. The intelligent instrument detection system according to claim 6, characterized in that: In S2, the pointer instrument reading recognition algorithm uses the yolov5s model to identify the pointer instrument from the captured image, then uses the yolov8x-pose model to detect the scale lines and key points of the pointer in the instrument, then uses the DBNetpp model to detect the value box and uses the SATRN model for text recognition, and finally performs post-processing to obtain the reading result.
8. The intelligent instrument detection system according to claim 7, characterized in that: When the yolov8x-pose model recognizes the key points of the pointer, the first and last points of the pointer are taken as the key points, and the rectangular frame composed of the key points is taken as the target detection frame. When recognizing the scale lines of the pointer instrument, each numerical frame is regarded as the same category for target detection, and each numerical frame carries the corresponding scale points as numerical key points. When post-processing, opencv is used to post-process the SATRN model recognition output, distinguish the inner and outer diameters of the pointer instrument, determine the order of the scale points, and correct the OCR recognition results. It is judged whether the recognition results meet the perspective transformation. If the perspective transformation is met, the points blocked by the pointer are supplemented and the readings are calculated. If the perspective transformation is not met, the perspective transformation is performed, and then the blocked points are supplemented and the readings are calculated.
9. The intelligent instrument detection system according to claim 6, characterized in that: In S2, the digital meter reading recognition algorithm is to identify the position of the dial from the acquired image through target detection, segment the dial position, further identify the numbers in the dial through target detection, and intelligently read the numbers in the dial based on the identified dial numbers.
10. An intelligent instrument detection method, applied to an intelligent instrument detection system according to any one of claims 1 to 9, characterized in that: The specific steps include: S4, creating an instrumentation simulation working environment, outputting a working environment simulation signal, and transmitting the working environment simulation signal to the instrumentation under test; S5, the instrument under test receives the working environment simulation signal and displays the instrument reading according to the simulation signal; S6, capturing instrument readings, and comparing the instrument readings with the working environment simulation signal to determine the accuracy of the instrument readings; S7, the comprehensive safety factor analysis and evaluation module outputs a fault analysis report based on the judgment result, reports a fault warning, gives an explanation and a solution report, and transmits the report to the communication terminal.
