Wind shield device operation fault monitoring system and judgment method based on image recognition

Abstract

The invention discloses a wind shield device operation fault monitoring system and judgment method based on image recognition, and belongs to the technical field of fault monitoring control systems. The fault monitoring control system is operated, a wind shield device controller receives an instruction of an operation station, a motor control cabinet is controlled by controlling on-off of a contactor, and finally forward rotation, reverse rotation and stopping of a motor are controlled. The operation fault judgment method comprises the following steps: S1, establishing a set of wind shield device operation fault monitoring control system; and S2, operating the monitoring algorithm for the operation fault of the wind shielding device based on image recognition. According to the invention, remote automatic control of the wind shielding device is realized, the operation fault of the wind shielding device can be automatically identified, and emergency processing and alarm prompting are carried out. According to the method, edge feature detection of the same frame of picture can be realized according to the Sobel operator, the efficiency of picture similarity calculation can be improved, meanwhile, the similarity of two groups of picture structures can be calculated through the SSIM index, and theoretical and data support is provided for judging whether the wind shielding device has an operation fault or not.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault monitoring and control systems, and in particular relates to a windshield device operation fault monitoring system and a determination method based on image recognition. Background Art

[0002] In winter, when the outdoor steam temperature is as low as -20℃, the direct air-cooled coal-fired generator sets in the north need to focus on the anti-freezing problem of the air-cooling island. One anti-freezing facility is a wind shield, such as Figure 1 As shown, the windshield is installed at the air outlet of the air cooling island fan. When antifreeze is needed, the motor control cabinet is used to turn on the forward rotation of the motor, and the windshield is opened to block the air outlet of the air cooling island fan, blocking air convection, weakening the cooling effect of the heat exchange surface here, and achieving the purpose of antifreeze; when antifreeze is not needed, the motor control cabinet is used to turn on the reverse rotation of the motor, and the windshield is closed, and the air outlet of the air cooling island fan flows normally, and the air cooling island operates normally. The current windshield can only be controlled locally under the air cooling island, and there are the following problems: 1) The air cooling island is far away from the control room, and can only be controlled locally under the air cooling island, and cannot be remotely controlled in the control room, which is inconvenient to adjust and has a large workload for personnel; 2) The working environment of the windshield is harsh, and its mechanical operation structure or motor may have operating problems and may eventually cause jamming and other faults, and cannot be turned on and off normally. When an operating fault occurs, the air cooling island is at a high place, and it is controlled locally under the air cooling island. The distance is far, and personnel cannot detect the movement status of the windshield in time. Summary of the invention

[0003] The purpose of the present invention is to propose a windshield device operation fault monitoring system and judgment method based on image recognition, which realizes remote automatic control of the windshield device, and can automatically identify the operation fault of the windshield device and perform emergency processing and alarm prompts.

[0004] The technical solution adopted by the present invention is:

[0005] A windshield device operation fault monitoring system based on image recognition, comprising a remote control system, the remote control system comprising

[0006] The contactor is connected to the motor control cabinet of the windshield device;

[0007] A windshield device controller, connected to the contactor;

[0008] An operating station connected to the windshield device controller;

[0009] The windshield device controller receives instructions from the operating station, controls the motor control cabinet by controlling the on and off of the contactor, and ultimately controls the forward, reverse, and stop of the motor.

[0010] A method for determining a windshield device operating fault based on image recognition comprises the following steps:

[0011] S1. Establish a windshield device operation fault monitoring and control system;

[0012] S2. Run a monitoring algorithm for windshield device operation failure based on image recognition.

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] 1. The present invention realizes remote automatic control of the windshield device, and can automatically identify operating failures of the windshield device and perform emergency processing and alarm prompts.

[0015] 2. The present invention can realize edge feature detection of the same-frame pictures according to the Sobel operator, which helps to improve the efficiency of picture similarity calculation. At the same time, the SSIM index can calculate the similarity of the structures of two groups of pictures, providing theoretical and data support for determining whether the windshield device has an operational failure.

[0016] 3. The method proposed in the present invention has great universality, is easy to implement in production practice, and has high promotion and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic diagram of the working of the windshield device;

[0018] Figure 2 Schematic diagram of a windshield device operation fault monitoring and control system of the present invention;

[0019] Figure 3 The present invention provides a method for determining a windshield device operating fault;

[0020] Among them: 1. Contactor; 2. Wind shield controller; 3. Operation station; 4. Operation fault monitoring system; 5. Online camera equipment; 6. Precision fluorescent ruler. DETAILED DESCRIPTION

[0021] In order to better understand the purpose, structure and function of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings.

[0022] The present invention includes the following contents:

[0023] 1) Establish a set of windshield device operation fault monitoring and control system;

[0024] 2) Monitoring algorithm for windshield device operation failure based on image recognition.

[0025] like Figure 2 As shown, the present invention provides a windshield device operation fault monitoring system based on image recognition, including a remote control system for the windshield device, the remote control system includes

[0026] Contactor 1 is connected to the motor control cabinet of the windshield device;

[0027] A windshield device controller 2 is connected to the contactor 1;

[0028] The operation station 3 is connected to the windshield device controller 2; the motor of the windshield device can be controlled to rotate forward, reverse, or stop at the operation station 3.

[0029] The windshield device controller 2 receives the instruction from the operating station 3, controls the motor control cabinet by controlling the on and off of the contactor 1, and finally controls the forward rotation, reverse rotation and stop of the motor.

[0030] A precision fluorescent ruler 6 is added in the direction of movement of the windshield device, and a set of high-resolution online camera equipment 5 is added to obtain image information of the windshield device and the precision fluorescent ruler 6 in real time, and transmit it to the operation fault monitoring system 4.

[0031] An operation fault monitoring system 4 is added to receive image information from a high-resolution online camera device 5 and the motor operation status of the operating station 3 in real time, to diagnose the operation faults of the windshield device, and transmit the operation fault status of the windshield device to the operating station 3. If an operation fault occurs, the motor movement is prohibited and an alarm is given. If no operation fault occurs, no operation is performed.

[0032] like Figure 3 As shown, the present invention provides a method for determining an operating fault of a windshield device based on image recognition, comprising the following steps:

[0033] S1. Establish a windshield device operation fault monitoring and control system;

[0034] S2. Run a monitoring algorithm for windshield device operation failure based on image recognition.

[0035] The monitoring algorithm is: based on the image information of the high-resolution online camera device 5 and the motor operating status of the operating station 3, Sobel edge detection and SSIM structural similarity calculation are performed to judge the operating fault status of the windshield device by comparing the actual operation and historical image information.

[0036] The following is a detailed explanation of the algorithm and calculation process:

[0037] S21. From the issuance of the windshield device controller 2 command to the end of a complete operation cycle, the windshield device is collected frame by frame during the normal operation of the history of the picture, cut redundant picture information, retain the edge of the device features, stored in the database;

[0038] S22. Feature extraction using Sobel: The Sobel operator is a classic edge detection method that uses convolution operations to perform local differential calculations on images, and then estimates the gradient direction and amplitude of the image, which can be used to extract image features.

[0039] The Sobel operator usually has two 3x3 convolution kernels, one for detecting horizontal edges and the other for detecting vertical edges.

[0040] Specifically:

[0041] Convolution kernel of horizontal gradient:

[0042]

[0043] Convolution kernel for vertical gradient:

[0044]

[0045] Through these two gradients, the gradient magnitude and gradient direction of each pixel in the image can be calculated by the following formula:

[0046] Gradient Magnitude:

[0047]

[0048] Gradient direction:

[0049]

[0050] Among them, the gradient magnitude represents the significance of the edge (i.e., the strength of the edge), and the gradient direction represents the direction of the edge.

[0051] Compared with simple differential operators such as Roberts, the Sobel operator is more robust to noise; at the same time, compared with the Canny operator, the Sobel operator is more efficient. Using the Sobel operator to extract edge features in the same frame number pictures in the history and actual operation process helps to improve the calculation efficiency of picture similarity;

[0052] S23. Use SSIM to calculate image similarity: The structural similarity index (SSIM) can be used to measure the similarity between two images. Different from traditional pixel-level metrics such as mean square error (MSE) or peak signal-to-noise ratio (PSNR), SSIM pays more attention to the structural information of the image, which is consistent with the human visual system's perception of image quality. For the identification of windshield failures, it is necessary to focus on the structural differences of the features of the same-frame images.

[0053] SSIM divides the similarity of two images x and y to be compared into three parts:

[0054] 1. Brightness similarity l(x,y):

[0055]

[0056] in,

[0057]

[0058] μ x ,μ y are the average values ​​of image x and y, respectively, used to represent brightness information, and C1 is a constant, usually defined as:

[0059] C1=(K1L) 2

[0060] Among them, K1<<1, L is the image grayscale coefficient.

[0061] 2. Contrast similarity c(x,y):

[0062]

[0063] in,

[0064]

[0065] are the variances of images x and y, respectively, used to represent contrast information, and C2 is a constant, usually defined as:

[0066] C2=(K2L) 2

[0067] Wherein, K2<<1, and L is defined as above.

[0068] 3. Structural similarity s(x,y):

[0069]

[0070] in,

[0071]

[0072] σ xy is the covariance of images x and y, which is used to describe structural information.

[0073] The three components of brightness, contrast, and structure work together to comprehensively measure the similarity between two images.

[0074] S24. For high-quality demand scenarios such as industrial inspection, SSIM usually sets a higher threshold. However, considering that noise and light in actual operation have a certain degree of influence on the calculation of brightness and contrast, the threshold setting can be appropriately reduced to the range of (0.6, 0.95) to highlight the advantages of structural contrast function calculation, thereby ensuring the rationality of the judgment. When the similarity of edge features of images with the same number of frames is less than k, there may be a potential operational failure. When the similarity of multiple consecutive frames of images is less than k, it can be determined that the windshield device has an operational failure. It should be noted that the threshold setting can be dynamically adjusted according to actual task requirements, and the threshold proposed in the present invention can be used as a reference.

[0075] S25. Transmit frame-by-frame images of the windshield device in operation to the monitoring system, calculate the image similarity in real time based on the Sobel operator and the SSIM algorithm, and when it is determined that an operational failure has occurred in the windshield device, the monitoring system sends a command to the windshield device controller 2 to terminate the operation of the windshield device.

[0076] It is to be understood that the present invention is described by some embodiments, and it is known to those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the scope of protection of the present invention.

Claims

1. A windshield device operation fault monitoring system based on image recognition, characterized in that: A remote control system is included, wherein the remote control system includes A contactor (1) is connected to a motor control cabinet of the windshield device; A windshield device controller (2) connected to the contactor (1); An operating station (3) connected to the windshield device controller (2); The windshield device controller (2) receives instructions from the operating station (3), controls the motor control cabinet by controlling the on and off of the contactor (1), and finally controls the forward rotation, reverse rotation, and stop of the motor.

2. The windshield device operation fault monitoring system based on image recognition according to claim 1 is characterized in that: A precision fluorescent ruler (6) is added in the moving direction of the windshield device, and a set of online camera equipment (5) is added to obtain image information of the windshield device and the precision fluorescent ruler (6) in real time.

3. The windshield device operation fault monitoring system based on image recognition according to claim 2 is characterized in that: The windshield device operation fault monitoring system based on image recognition also includes an operation fault monitoring system (4) for receiving image information from an online camera device (5) and the motor operation status of an operation station (3) in real time, performing operation fault diagnosis on the windshield device, and transmitting the operation fault status of the windshield device to the operation station (3).

4. A method for determining a windshield device operation fault monitoring system based on image recognition according to claims 1 to 3, characterized in that: The following steps are involved: S1. Establish a windshield device operation fault monitoring and control system; S2. Run a monitoring algorithm for windshield device operation failure based on image recognition.

5. The method for determining operating faults of a windshield device based on image recognition according to claim 4, characterized in that: The monitoring algorithm in S2 is: based on the image information of the online camera device (5) and the motor operating status of the operating station (3), Sobel edge detection and SSIM structural similarity calculation are performed to judge the operating fault status of the windshield device by comparing the actual operation and historical image information.

6. The method for determining operating faults of a windshield device based on image recognition according to claim 5, characterized in that: The specific steps of the monitoring algorithm in S2 are: S21. From the issuance of the windshield device controller (2) command to the end of a complete operation cycle, the windshield device is collected frame by frame during the normal operation of the historical process, redundant image information is cropped, the device edge features are retained, and stored in the database; S22. Feature extraction using Sobel: The Sobel operator performs local difference calculation on the image through convolution operation, and then estimates the gradient direction and amplitude of the image to extract the features of the image; S23. Use SSIM to calculate the image similarity; S24. Setting a threshold k for similarity calculation; when the similarity of edge features of the same number of frames is less than k, there may be a potential operational failure; when the similarity of multiple consecutive frames is less than k, it can be determined that the windshield device has an operational failure; S25. Transmitting frame-by-frame images of the windshield device in operation to the monitoring system, calculating the image similarity in real time based on the Sobel operator and the SSIM algorithm, and when determining that an operational failure has occurred in the windshield device, the monitoring system sends a command to the windshield device controller (2) to terminate the operation of the windshield device.

7. The method for determining operating faults of a windshield device based on image recognition according to claim 6, characterized in that: The specific calculation method of S23 is: SSIM divides the similarity of two images x and y to be compared into three parts: a. Brightness similarity l(x,y): in, μ x ,μ y are the average values ​​of image x and y, respectively, used to represent brightness information, and C1 is a constant, usually defined as: C1=(K1L) 2 Where, K1<<1, L is the image grayscale coefficient; b. Contrast similarity c(x,y): in, are the variances of images x and y, respectively, used to represent contrast information, and C2 is a constant, defined as: C2=(K2L) 2 Where, K2<<1, L is the image grayscale coefficient; c. Structural similarity s(x,y): in, σ xy is the covariance of images x and y, which is used to describe structural information.

Images