A fault identification and analysis method and system for sintering machine trolley railing

Abstract

The present application relates to the field of metallurgical sintering, and provides a method and system for identifying and analyzing the faults of a sintering machine trolley guardrail, by acquiring images captured by a visible light camera and a thermal infrared device; identifying the images captured by the visible light camera, performing image analysis on the thermal infrared images of the sintering machine guardrail, and storing the first analysis result; after the sintering machine trolley circulates once, acquiring the image captured by the thermal infrared device for the second time, performing image analysis, and storing the second analysis result; if the first analysis result is the same as the second analysis result, outputting the final identification and analysis result; if they are different, re-acquiring the image, and performing identification and analysis. The present application provides a method and system for identifying and analyzing the faults of a sintering machine trolley guardrail, which can timely alarm and feedback the fault type and the corresponding trolley steel stamp number, reduce the time the trolley runs with a fault, reduce the workload of manual inspections, and ensure the normal operation of the sintering machine.

Description

Technical Field

[0001] The present application relates to the field of metallurgical sintering, and in particular to a fault identification and analysis method and system for a sintering machine trolley guardrail. Background Art

[0002] Sintering is a crucial step in the steel production process. The general process involves mixing iron ore powder, lime powder, and coal powder in a specific proportion to form a mixture, which is then calcined to produce sintered ore with sufficient strength and particle size. The sintering machine is the core equipment in the steel production process, and its proper operation directly impacts the stability of the entire production process. The sintering machine trolley, as its core component, plays a crucial role in the sintering machine's operating efficiency.

[0003] like Figure 1 The figure shows a schematic structural diagram of a sintering machine trolley in the prior art, comprising: a frame 01, grate bars 02, a guardrail 03 and a roller 04; wherein the rotation of the roller 04 can drive the grate bars 02 to move, and the grate bars 02 are used to carry the mixture; the guardrail 03 provided on the frame 01 is used to block the mixture and play a sealing role to prevent the mixture from overflowing.

[0004] During the sintering process, as the sintering material burns from top to bottom, the temperature distribution in the vertical direction of the guardrail 03 is very uneven, and the temperature difference between the upper and lower parts of the sintering material is large. Since the mixture on the grate 02 is completely burned, the working temperature is generally 800-1000℃. The sintered mixture is then cooled to below 150℃, which will produce a corresponding temperature difference. Therefore, the guardrail 03 will work under this hot and cold working condition for a long time, which will produce complex alternating thermal stress, and then cause thermal fatigue. Therefore, fatigue cracking is prone to occur under the action of alternating thermal stress. As time goes by, the cracks expand and even perforation or breakage may occur.

[0005] In order to detect this type of fault in the guardrail 03, in actual production, the guardrail 03 is inspected through manual inspections and central control monitoring. If cracks are found in the guardrail 03, targeted treatment measures are taken. However, this method cannot immediately detect the presence of cracks in the guardrail 03, as well as the area and location of the cracks. Moreover, the cracks in the inner wall of the guardrail 03 cannot be detected when the trolley is in operation. As a result, the cracks in the guardrail 03 continue to grow, resulting in a deterioration in the sealing performance of the production process. In severe cases, this can cause leakage of the mixed material and air leakage, which can further reduce the quality of the products produced by the sintering machine and affect normal production. Summary of the Invention

[0006] In order to promptly detect such faults of the sintering machine trolley guardrail and obtain the type of fault, the location and area of ​​the perforation or fracture, and the shape, location and area of ​​the crack, the present application provides a fault identification and analysis method and system for the sintering machine trolley guardrail.

[0007] In a first aspect of the present application, a method for identifying and analyzing a fault of a sintering machine trolley guardrail is provided, wherein the method comprises:

[0008] Acquire images collected by visible light cameras and thermal infrared instruments.

[0009] Identify the image captured by the visible light camera. If the steel stamp number on the sintering machine frame is located in the middle of the captured image in the image captured by the visible light camera, store the image captured by the thermal infrared device corresponding to the image captured by the visible light camera as the thermal infrared image of the sintering machine railing, and store the steel stamp number as the first identified steel stamp number.

[0010] Perform image analysis on the thermal infrared image of the sintering machine guardrail, store the image analysis result, and mark the image analysis result as the first analysis result.

[0011] After the sintering machine trolley circulates once, the steel stamp number is identified for the second time, the image collected by the second thermal infrared device is obtained, the image collected by the second thermal infrared device is analyzed, the result of the image analysis is stored, and the result of the image analysis is marked as the second analysis result.

[0012] If the first analysis result is the same as the second analysis result, the final recognition analysis result is output; if the first analysis result is different from the second analysis result, the image is collected again and the recognition analysis is performed.

[0013] The steps of image analysis are as follows:

[0014] The fault temperature threshold of the thermal infrared image of the sintering machine trolley guardrail is set, and all pixel points in the thermal infrared image set of the sintering machine trolley guardrail and the pixel values ​​corresponding to the pixel points are obtained.

[0015] According to the pixel value, the pixel value is filtered to obtain a temperature value corresponding to the pixel point.

[0016] According to the temperature value, if the temperature value is greater than or equal to the fault temperature threshold, the temperature value is marked as a fault temperature value; if the temperature value is less than the fault temperature threshold, the temperature value is marked as a normal temperature value, the pixel point corresponding to the fault temperature value is a fault pixel point, and the pixel point corresponding to the normal temperature value is a normal pixel point.

[0017] If all are normal pixels, the sintering machine trolley guardrail is normal. If there are faulty pixels, then based on the shape of the arrangement of the faulty pixels, if the shape is a hole formed by multiple pixels gathered together, or a strip formed by multiple pixels gathered together, then the fault type is perforation or fracture. If the shape is a thread-like shape formed by multiple pixels connected together, then the fault type is a crack.

[0018] If the fault is a perforation or a fracture, the fault temperature value and the normal temperature value are binarized to obtain the area and position of the perforation or fracture region.

[0019] If the fault is a crack, the pixel points are gradient-divided using a Laplace model to obtain a gradient image.

[0020] The gradient image is converted into a binary image, morphological processing is performed on the binary image, pixel points corresponding to the crack are screened out, and data closure processing is performed on the pixel points corresponding to the crack to obtain the shape, position and area of ​​the crack.

[0021] Optionally, in the filtering step, the filtering is implemented using a Gaussian filtering model, and the Gaussian filtering model is specifically:

[0022]

[0023] A coordinate system is established on the thermal infrared image set, with the vertex of the upper corner of the thermal infrared image as the origin, x and y as the position of the pixel point in the coordinate system, and σ as the variance of the coordinate point.

[0024] Optionally, it is characterized in that, in the binarization step, the binarization step is specifically:

[0025] The faulty pixel is marked as 255 and marked in white, and the normal pixel is marked as 0 and marked in black. The conversion model is:

[0026]

[0027] T is the fault temperature threshold. The perforated or fractured area forms a color difference with the normal area. All perforated or fractured areas are displayed in white, and the normal area is displayed in black. The white part is located and the area is calculated.

[0028] Optionally, the fault temperature threshold is a pre-collected fault panel, and the temperature value corresponding to the pixel point of the fault area of ​​the fault panel is obtained. The fault temperature threshold is T, 50°C≤T≤200°C.

[0029] Optionally, the Laplace model is:

[0030]

[0031] satisfy The conditional points constitute the gradient image.

[0032] Optionally, the step of converting the gradient image into a binary image is specifically as follows:

[0033] is the second-order differential expression of the pixel point, Perform convolution, the convolution model is:

[0034]

[0035] According to the gradient algorithm, pixel points whose gradient values ​​corresponding to the pixel points are greater than the gradient threshold are marked as 1, and pixel points whose gradient values ​​are less than or equal to the gradient threshold are marked as 0, thereby obtaining a binary image.

[0036] In a second aspect of the present application, a fault identification and analysis system for a sintering machine trolley guardrail is provided, the system comprising an image acquisition module, an image recognition module, an image primary analysis module, an image secondary analysis module and a verification module.

[0037] The image acquisition module is used to obtain images captured by the visible light camera and the thermal infrared device.

[0038] The image recognition module is used to identify the image captured by the visible light camera. If the steel stamp number on the sintering machine frame is located in the middle of the captured image in the image captured by the visible light camera, the image captured by the thermal infrared instrument corresponding to the image captured by the visible light camera is stored as a thermal infrared image of the sintering machine railing, and the steel stamp number is stored as the first recognized steel stamp number.

[0039] The image initial analysis module is used to perform image analysis on the thermal infrared image of the sintering machine guardrail, store the image analysis result, and mark the image analysis result as the first analysis result.

[0040] The image secondary analysis module is used to identify the steel stamp number for the second time after the sintering machine trolley circulates once, obtain the image collected by the thermal infrared device for the second time, perform image analysis on the image collected by the thermal infrared device for the second time, store the result of the image analysis, and mark the result of the image analysis as the second analysis result.

[0041] The verification module is used to output the final recognition analysis result if the first analysis result is the same as the second analysis result, and to re-collect the image and perform recognition analysis if the first analysis result is different from the second analysis result.

[0042] In the image primary analysis module and the image secondary analysis module, the units performing the image analysis step include: an original data acquisition unit, a filtering unit, a temperature conversion unit, a fault type judgment unit, a binarization processing unit, a gradient image unit and a closure processing unit.

[0043] The raw data acquisition unit is used to set the fault temperature threshold of the thermal infrared image of the sintering machine trolley guardrail and acquire all pixel points in the thermal infrared image set of the sintering machine trolley guardrail and the pixel values ​​corresponding to the pixel points.

[0044] The filtering unit is used to perform filtering processing on the pixel value according to the pixel value to obtain the temperature value corresponding to the pixel point.

[0045] The temperature conversion unit is used to mark the temperature value as a fault temperature value based on the temperature value. If the temperature value is greater than or equal to the fault temperature threshold, the temperature value is marked as a normal temperature value if the temperature value is less than the fault temperature threshold. The pixel point corresponding to the fault temperature value is a fault pixel point, and the pixel point corresponding to the normal temperature value is a normal pixel point.

[0046] The fault type judgment unit is used to judge that if all the pixels are normal, the sintering machine trolley guardrail is normal; if there are faulty pixels, then based on the shape of the arrangement of the faulty pixels, if the shape is a hole formed by multiple pixels gathered together, or a strip formed by multiple pixels gathered together, the fault type is perforation or fracture; if the shape is a thread-like shape formed by multiple pixels connected together, the fault type is crack.

[0047] The binarization processing unit is used to perform binarization processing on the fault temperature value and the normal temperature value if the fault is a perforation or a fracture, so as to obtain the area and position of the perforation or fracture area.

[0048] The gradient image unit is configured to perform gradient division on the pixel points using a Laplace model to obtain a gradient image if the fault is a crack.

[0049] The closure processing unit is used to convert the gradient image into a binary image, perform morphological processing on the binary image, filter out pixel points corresponding to the crack, and perform data closure processing on the pixel points corresponding to the crack to obtain the shape, position and area of ​​the crack.

[0050] Optionally, in the filtering step, the filtering is implemented using a Gaussian filtering model, and the Gaussian filtering model is specifically:

[0051]

[0052] A coordinate system is established on the thermal infrared image set, with the vertex of the upper corner of the thermal infrared image as the origin, x and y as the position of the pixel point in the coordinate system, and σ as the variance of the coordinate point.

[0053] Optionally, in the binarization step, the binarization step is specifically:

[0054] The faulty pixel is marked as 255 and marked in white, and the normal pixel is marked as 0 and marked in black. The conversion model is:

[0055]

[0056] T is the fault temperature threshold. The perforated or fractured area forms a color difference with the normal area. All perforated or fractured areas are displayed in white, and the normal area is displayed in black. The white part is located and the area is calculated.

[0057] Optionally, the fault temperature threshold is a pre-collected fault panel, and the temperature value corresponding to the pixel point of the fault area of ​​the fault panel is obtained. The fault temperature threshold is T, 50°C≤T≤200°C.

[0058] Optionally, the Laplace model is:

[0059]

[0060] satisfy The conditional points constitute the gradient image.

[0061] Optionally, the step of converting the gradient image into a binary image is specifically as follows:

[0062] is the second-order differential expression of the pixel point, Perform convolution, the convolution model is:

[0063]

[0064] According to the gradient algorithm, pixel points whose gradient values ​​corresponding to the pixel points are greater than the gradient threshold are marked as 1, and pixel points whose gradient values ​​are less than or equal to the gradient threshold are marked as 0, thereby obtaining a binary image.

[0065] It can be seen from the above technical solution that the present application provides a fault identification and analysis method and system for a sintering machine trolley guardrail, by acquiring images captured by a visible light camera and a thermal infrared device; identifying the image captured by the visible light camera, and if the steel stamp number on the sintering machine frame is located in the middle of the captured image in the image captured by the visible light camera, then the image captured by the thermal infrared device corresponding to the image captured by the visible light camera is stored as a thermal infrared image of the sintering machine guardrail, and the steel stamp number is stored as the first identification steel stamp number; performing image analysis on the thermal infrared image of the sintering machine guardrail, storing the result of the image analysis, and marking the result of the image analysis as the first analysis result; after the sintering machine trolley circulates once, identifying the steel stamp number for the second time, acquiring the image captured by the thermal infrared device for the second time, performing image analysis on the image captured by the thermal infrared device for the second time, storing the result of the image analysis, and marking the result of the image analysis as the second analysis result; if the first analysis result is the same as the second analysis result, outputting the final identification and analysis result, and if the first analysis result is different from the second analysis result, re-capturing the image for identification and analysis. The embodiment of the present application provides a fault identification and analysis method for a sintering machine trolley guardrail, which can promptly alarm and feedback the fault type and the corresponding trolley steel stamp number, reduce the time the trolley runs with a fault, reduce the manual inspection workload of these workers, and ensure the normal operation of the sintering machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0067] Figure 1 It is a structural diagram of an existing sintering machine trolley;

[0068] Figure 2 Schematic diagram of the installation of the visible light camera and thermal infrared imager provided in the embodiment of the present application;

[0069] Figure 3 A side view of a sintering machine trolley provided in an embodiment of the present application;

[0070] Figure 4 A schematic flow chart of a method for identifying and analyzing a fault of a sintering machine trolley railing provided in an embodiment of the present application;

[0071] Figure 5 A flowchart of image analysis provided in an embodiment of the present application;

[0072] Figure 6 A framework diagram of a fault identification and analysis system for a sintering machine trolley railing provided in an embodiment of the present application;

[0073] Figure 7 A unit framework diagram for performing image analysis steps provided in an embodiment of the present application.

[0074] in:

[0075] 01-frame, 011-steel stamp number, 02-grate, 03-railway, 04-roller, 05-trolley wheel, 06-visible light camera, 07-thermal infrared imager. DETAILED DESCRIPTION

[0076] The following embodiments are described in detail, with examples illustrated in the accompanying drawings. When the following description refers to the drawings, identical numbers in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following embodiments are not intended to represent all possible implementations consistent with the present application. They are merely examples of systems and methods consistent with certain aspects of the present application, as detailed in the claims.

[0077] To illustrate the scenario of the embodiment of the present application, Figure 2 As shown, it is a schematic diagram of the installation of the visible light camera and the thermal infrared imager provided in an embodiment of the present application, wherein the thermal infrared imager 07 is arranged on both sides of the sintering machine trolley, the visible light camera 06 is arranged on one side of the sintering machine trolley, the thermal infrared imager 07 is arranged at the waistline of the railing 03, and the distance between the thermal infrared imager 07 and the railing 03 is 1-3 meters. The visible light camera 06 is arranged directly below the thermal infrared imager 07, and the thermal infrared imager 07 is used to collect thermal infrared images of the railing 03, and the visible light camera 06 is used to collect images of the railing 03.

[0078] In addition, if Figure 3 As shown, it is a side view of the sintering machine trolley provided in an embodiment of the present application, wherein a frame 01 is provided on the wheel 05, and the steel stamp number 011 is located in the middle of the frame 01. When the steel stamp number 011 is in the middle of the image in the picture captured by the visible light camera 06, it means that the captured image includes a whole railing 03.

[0079] See also Figure 4 , which is a flow chart of a fault identification and analysis method for a sintering machine trolley railing provided in an embodiment of the present application.

[0080] In a first aspect, the present application provides a method for identifying and analyzing a fault of a sintering machine trolley guardrail, the method comprising:

[0081] S101, acquiring images captured by a visible light camera and a thermal infrared device.

[0082] During the image acquisition process, the sintering machine trolley is in motion, and the visible light camera and the thermal infrared instrument continuously take pictures of the sintering machine railing.

[0083] S102, identifying the image captured by the visible light camera. If the steel stamp number on the sintering machine frame is located in the middle of the captured image in the image captured by the visible light camera, the image captured by the thermal infrared device corresponding to the image captured by the visible light camera is stored as a thermal infrared image of the sintering machine railing, and the steel stamp number is stored as the first recognized steel stamp number.

[0084] The steel stamp number recognition in the embodiment of the present application uses a visible light camera for recognition, and the steel stamp number is a combination of numbers and English letters.

[0085] S103, performing image analysis on the thermal infrared image of the sintering machine guardrail, storing the image analysis result, and marking the image analysis result as the first analysis result.

[0086] S104, after the sintering machine trolley circulates once, the steel stamp number is identified for the second time, the image collected by the thermal infrared device is obtained for the second time, image analysis is performed on the image collected by the thermal infrared device for the second time, the result of the image analysis is stored, and the result of the image analysis is marked as the second analysis result.

[0087] S105: If the first analysis result is the same as the second analysis result, the final recognition analysis result is output; if the first analysis result is different from the second analysis result, the image is collected again and the recognition analysis is performed.

[0088] The two images collected from the same guardrail are compared to ensure that the detection results are correct. If the two detection results are different, re-identification and analysis are required.

[0089] See also Figure 5 A flowchart of image analysis provided in an embodiment of the present application.

[0090] After a penetrating crack appears on the guardrail, the trolley is filled with mixed ore, which is burned in the ignition furnace and gradually ignites the ore on the trolley. The temperature of the ignited red ore is high, reaching thousands of degrees. The guardrail acts as a baffle for the hot ore. If cracks or perforations appear, the temperature of the cracks and perforations will be higher than that of the surrounding guardrails. This can be used to determine whether the guardrail has cracks or perforations, and to locate and calculate the area of ​​the abnormal parts.

[0091] The steps of image analysis are as follows:

[0092] S1041, setting a fault temperature threshold of the thermal infrared image of the sintering machine trolley guardrail, and obtaining all pixel points in the thermal infrared image set of the sintering machine trolley guardrail and the pixel values ​​corresponding to the pixel points.

[0093] S1042: Perform filtering processing on the pixel value according to the pixel value to obtain a temperature value corresponding to the pixel point.

[0094] Gaussian filtering is a linear smoothing filter suitable for eliminating Gaussian noise and can be used for noise reduction in image processing. Gaussian filtering is a process of performing a weighted average of the entire image. The value of each pixel is a weighted average of its own value and the values ​​of other pixels in its neighborhood. The specific operation of Gaussian filtering is to scan each pixel in the image using a template (also called a convolution or mask) and replace the value of the central pixel of the template with the weighted average grayscale value of the pixels in the neighborhood determined by the template.

[0095] In the filtering step, the filtering is implemented using a Gaussian filtering model, which is specifically:

[0096]

[0097] A coordinate system is established on the thermal infrared image set, with the vertex of the upper corner of the thermal infrared image as the origin, x and y as the position of the pixel point in the coordinate system, and σ as the variance of the coordinate point.

[0098] S1043. Based on the temperature value, if the temperature value is greater than or equal to the fault temperature threshold, the temperature value is marked as a fault temperature value; if the temperature value is less than the fault temperature threshold, the temperature value is marked as a normal temperature value, and the pixel point corresponding to the fault temperature value is a fault pixel point, and the pixel point corresponding to the normal temperature value is a normal pixel point.

[0099] S1044, if all are normal pixels, the sintering machine trolley guardrail is normal. If there are faulty pixels, then according to the shape of the arrangement of the faulty pixels, if the shape is a hole formed by multiple pixels gathered together, or a strip formed by multiple pixels gathered together, then the fault type is perforation or fracture. If the shape is a thread-like shape formed by multiple pixels connected together, then the fault type is crack.

[0100] When the sintering machine guardrail is normal, the sintering process of the sintering machine proceeds normally. Once a fault pixel is found, the entire sintering site immediately enters an alert state, and then further corresponding measures are taken according to the type of fault. If perforation or breakage occurs, the guardrail needs to be replaced. If the fault is a crack, it will be prompted that the guardrail needs to be replaced in the near future.

[0101] S1045: If the fault is a perforation or a fracture, binarization is performed on the fault temperature value and the normal temperature value to obtain the area and position of the perforation or fracture region.

[0102] Since the perforation of the guardrail causes a high local temperature, the fracture of the guardrail will also cause temperature anomalies at the fracture site. Therefore, a threshold value can be set for sampling and binary image processing to select the area within the high temperature range. According to the current image temperature value, a threshold value T (50℃≤T≤200℃) is set. The pixel temperature value greater than the fault temperature threshold T is set to 200℃, and the pixel temperature value less than the fault temperature threshold is set to 50℃. The conversion process is as follows:

[0103]

[0104] T is the fault temperature threshold. The perforated or fractured areas have a color difference from the normal areas. All perforated or fractured areas are displayed in white, while the normal areas are displayed in black. The white parts are located and their areas are calculated.

[0105] S1046: If the fault is a crack, perform gradient division on the pixel points using a Laplace model to obtain a gradient image.

[0106] When cracks appear on the railing, the high temperature of the ore on the trolley will cause a large difference in temperature between the cracked railing and the surrounding temperature. This difference can be found using the gradient algorithm. Here, the Laplace model is used to segment the edge. The Laplace model is an operator that uses second-order differentials, which is actually the divergence of the gradient. The second-order differential expression is:

[0107]

[0108]

[0109]

[0110] satisfy The conditional points constitute the gradient image, and f(x,y) is the pixel point on the image coordinate (x,y).

[0111] S1047, converting the gradient image into a binary image, performing morphological processing on the binary image to filter out pixel points corresponding to the crack, and performing data closure processing on the pixel points corresponding to the crack to obtain the shape, position and area of ​​the crack.

[0112] The steps of converting the gradient image into a binary image are specifically as follows:

[0113] is the second-order differential expression of the pixel point, Perform convolution, the convolution model is:

[0114]

[0115] Using a gradient algorithm, pixels with gradient values ​​greater than a threshold are marked as 1, while pixels with gradient values ​​less than or equal to the threshold are marked as 0. Using morphological dilation and erosion, an opening operation removes discrete noise points, and a closing operation closes the crack data, resulting in a binary image. This processed image can then identify the crack's state, including its location and area.

[0116] It can be seen from the above technical solution that the embodiment of the present application provides a fault identification and analysis method for a sintering machine trolley guardrail, by acquiring images captured by a visible light camera and a thermal infrared device; identifying the image captured by the visible light camera, if the steel stamp number on the sintering machine frame is located in the middle position of the captured image in the image captured by the visible light camera, then the image captured by the thermal infrared device corresponding to the image captured by the visible light camera is stored as the thermal infrared image of the sintering machine guardrail, and the steel stamp number is stored as the first identification steel stamp number; performing image analysis on the thermal infrared image of the sintering machine guardrail, storing the result of the image analysis, and marking the result of the image analysis as the first analysis result; after the sintering machine trolley circulates once, identifying the steel stamp number for the second time, acquiring the image captured by the thermal infrared device for the second time, performing image analysis on the image captured by the thermal infrared device for the second time, storing the result of the image analysis, and marking the result of the image analysis as the second analysis result; if the first analysis result is the same as the second analysis result, outputting the final identification and analysis result, if the first analysis result is different from the second analysis result, re-capturing the image for identification and analysis. The embodiment of the present application provides a fault identification and analysis method for a sintering machine trolley guardrail, which can promptly alarm and feedback the fault type and the corresponding trolley steel stamp number, reduce the time the trolley runs with a fault, reduce the manual inspection workload of these workers, and ensure the normal operation of the sintering machine.

[0117] See also Figure 6 , which is a framework diagram of a fault identification and analysis system for a sintering machine trolley railing provided in an embodiment of the present application.

[0118] Corresponding to the aforementioned embodiment of the method for identifying and analyzing a sintering machine trolley guardrail fault, the second aspect of the embodiments of the present application provides a system for identifying and analyzing a sintering machine trolley guardrail fault. For details not disclosed in the fault identification and analysis method provided in the embodiments of the present application, please refer to the fault identification and analysis system provided in the embodiments of the present application. The system for identifying and analyzing a sintering machine trolley guardrail fault includes an image acquisition module, an image recognition module, a primary image analysis module, a secondary image analysis module, and a verification module.

[0119] The image acquisition module is used to obtain images captured by the visible light camera and the thermal infrared device.

[0120] The image recognition module is used to identify the image captured by the visible light camera. If the steel stamp number on the sintering machine frame is located in the middle of the captured image in the image captured by the visible light camera, the image captured by the thermal infrared instrument corresponding to the image captured by the visible light camera is stored as a thermal infrared image of the sintering machine railing, and the steel stamp number is stored as the first recognized steel stamp number.

[0121] The image initial analysis module is used to perform image analysis on the thermal infrared image of the sintering machine guardrail, store the image analysis result, and mark the image analysis result as the first analysis result.

[0122] The image secondary analysis module is used to identify the steel stamp number for the second time after the sintering machine trolley circulates once, obtain the image collected by the thermal infrared device for the second time, perform image analysis on the image collected by the thermal infrared device for the second time, store the result of the image analysis, and mark the result of the image analysis as the second analysis result.

[0123] The verification module is used to output the final recognition analysis result if the first analysis result is the same as the second analysis result, and to re-collect the image and perform recognition analysis if the first analysis result is different from the second analysis result.

[0124] See also Figure 7 , which is a unit framework diagram for performing image analysis steps provided in an embodiment of the present application.

[0125] In the image primary analysis module and the image secondary analysis module, the units performing the image analysis step include: an original data acquisition unit, a filtering unit, a temperature conversion unit, a fault type judgment unit, a binarization processing unit, a gradient image unit and a closure processing unit.

[0126] The raw data acquisition unit is used to set the fault temperature threshold of the thermal infrared image of the sintering machine trolley guardrail and acquire all pixel points in the thermal infrared image set of the sintering machine trolley guardrail and the pixel values ​​corresponding to the pixel points.

[0127] The filtering unit is used to perform filtering processing on the pixel value according to the pixel value to obtain the temperature value corresponding to the pixel point.

[0128] The temperature conversion unit is used to mark the temperature value as a fault temperature value based on the temperature value. If the temperature value is greater than or equal to the fault temperature threshold, the temperature value is marked as a normal temperature value if the temperature value is less than the fault temperature threshold. The pixel point corresponding to the fault temperature value is a fault pixel point, and the pixel point corresponding to the normal temperature value is a normal pixel point.

[0129] The fault type judgment unit is used to judge that if all the pixels are normal, the sintering machine trolley guardrail is normal; if there are faulty pixels, then based on the shape of the arrangement of the faulty pixels, if the shape is a hole formed by multiple pixels gathered together, or a strip formed by multiple pixels gathered together, the fault type is perforation or fracture; if the shape is a thread-like shape formed by multiple pixels connected together, the fault type is crack.

[0130] The binarization processing unit is used to perform binarization processing on the fault temperature value and the normal temperature value if the fault is a perforation or a fracture, so as to obtain the area and position of the perforation or fracture area.

[0131] The gradient image unit is configured to perform gradient division on the pixel points using a Laplace model to obtain a gradient image if the fault is a crack.

[0132] The closure processing unit is used to convert the gradient image into a binary image, perform morphological processing on the binary image, filter out pixel points corresponding to the crack, and perform data closure processing on the pixel points corresponding to the crack to obtain the shape, position and area of ​​the crack.

[0133] The images captured by the visible light camera and the thermal infrared instrument are input into the image acquisition module, and the image acquisition module transfers the captured images to the image recognition module and identifies the image captured by the visible light camera. If the steel stamp number on the sintering machine frame is located in the middle of the captured image in the image captured by the visible light camera, the image captured by the thermal infrared instrument corresponding to the image captured by the visible light camera is stored as the thermal infrared image of the sintering machine railing, and the steel stamp number is stored as the first recognized steel stamp number; the image recognition module transfers the thermal infrared image of the sintering machine railing to the primary image analysis module, performs image analysis on the thermal infrared image of the sintering machine railing, stores the result of the image analysis, and marks the result of the image analysis as the first analysis result; the image recognition module will recognize the steel stamp number for the second time after the sintering machine trolley circulates once, obtains the image captured by the thermal infrared instrument for the second time, performs image analysis on the image captured by the thermal infrared instrument for the second time, stores the result of the image analysis, and marks the result of the image analysis as the second analysis result. The primary image analysis module transmits the first analysis result to the verification module, and the secondary image analysis module transmits the second analysis result to the verification module. If the first analysis result is the same as the second analysis result, the final recognition analysis result is output; if the first analysis result is different from the second analysis result, the image is re-captured and the recognition analysis is performed.

[0134] In the unit that performs the image analysis step, the set fault temperature threshold of the thermal infrared image of the sintering machine trolley guardrail and all pixel points in the thermal infrared image set of the sintering machine trolley guardrail, as well as the pixel values ​​corresponding to the pixel points, are input into the original data acquisition unit; the original data acquisition unit transfers the pixel value corresponding to the pixel point to the filtering unit, and filters the pixel value according to the pixel value to obtain the temperature value corresponding to the pixel point; the filtering unit transfers the temperature value corresponding to the pixel point to the temperature conversion unit, and the original data acquisition unit transfers the fault temperature threshold to the temperature conversion unit, and according to the temperature value, if the temperature value is greater than or equal to the fault temperature threshold, the temperature value is marked as a fault temperature value, and if the temperature value is less than the fault temperature threshold, the temperature value is marked as a normal temperature value, the pixel point corresponding to the fault temperature value is a fault pixel point, and the pixel point corresponding to the normal temperature value is a normal pixel point. The temperature conversion unit transmits the fault pixel points and the normal pixel points to the fault type judgment unit. If all the pixels are normal, the sintering machine trolley railing is normal. If there are faulty pixels, according to the shape of the arrangement of the faulty pixels, if the shape is a plurality of pixels gathered to form a hole, or a plurality of pixels gathered to form a strip, the fault type is perforation or fracture. If the shape is a plurality of pixels connected to form a thread, the fault type is crack. The fault type judgment unit transmits the fault type perforation or fracture to the binarization processing unit. The fault temperature value and the normal temperature value are binarized to obtain the area and position of the perforation or fracture area; the fault type judgment unit transmits the fault type crack to the gradient image unit, and performs gradient division on the pixel points to obtain a gradient image; the gradient image unit transmits the gradient image to the closure processing unit, and converts the gradient image into a binary image, performs morphological processing on the binary image, filters out the pixel points corresponding to the crack, and performs data closure processing on the pixel points corresponding to the crack to obtain the shape, position and area of ​​the crack.

[0135] It can be seen from the above technical solution that the embodiment of the present application provides a fault identification and analysis method and system for a sintering machine trolley guardrail, by acquiring images captured by a visible light camera and a thermal infrared device; identifying the image captured by the visible light camera, if the steel stamp number on the sintering machine frame is located in the middle position of the captured image in the image captured by the visible light camera, then the image captured by the thermal infrared device corresponding to the image captured by the visible light camera is stored as the thermal infrared image of the sintering machine guardrail, and the steel stamp number is stored as the first identification steel stamp number; performing image analysis on the thermal infrared image of the sintering machine guardrail, storing the result of the image analysis, and marking the result of the image analysis as the first analysis result; after the sintering machine trolley circulates once, identifying the second identification steel stamp number, acquiring the image captured by the thermal infrared device for the second time, performing image analysis on the image captured by the thermal infrared device for the second time, storing the result of the image analysis, and marking the result of the image analysis as the second analysis result; if the first analysis result is the same as the second analysis result, outputting the final identification and analysis result, if the first analysis result is different from the second analysis result, re-capturing the image for identification and analysis. The embodiment of the present application provides a fault identification and analysis method for a sintering machine trolley guardrail, which can promptly alarm and feedback the fault type and the corresponding trolley steel stamp number, reduce the time the trolley runs with a fault, reduce the manual inspection workload of these workers, and ensure the normal operation of the sintering machine.

[0136] Similar parts between the embodiments provided in this application can be referenced to each other. The specific implementation methods provided above are only a few examples under the overall concept of this application and do not constitute a limitation on the scope of protection of this application. For those skilled in the art, any other implementation methods expanded based on the scheme of this application without expending creative work shall fall within the scope of protection of this application.

Claims

1. A method for identifying and analyzing faults of a sintering machine trolley guardrail, characterized in that: include: Acquire images collected by visible light cameras and thermal infrared instruments; Identify the image captured by the visible light camera. If the stamp number on the sintering machine frame is located in the middle of the captured image, store the image captured by the thermal infrared device corresponding to the image captured by the visible light camera as the thermal infrared image of the sintering machine railing, and store the stamp number as the first identified stamp number. Performing image analysis on the thermal infrared image of the sintering machine guardrail, storing the image analysis result, and marking the image analysis result as the first analysis result; After the trolley has passed through the sintering machine once, the steel stamp number is recognized for the second time, an image captured by the second thermal infrared instrument is obtained, image analysis is performed on the image captured by the second thermal infrared instrument, the result of the image analysis is stored, and the result of the image analysis is marked as the second analysis result; If the first analysis result is the same as the second analysis result, the final recognition analysis result is output; if the first analysis result is different from the second analysis result, the image is collected again and the recognition analysis is performed; The steps of image analysis are as follows: Set the fault temperature threshold of the thermal infrared image of the sintering machine trolley railing, and obtain all the pixel points in the thermal infrared image set of the sintering machine trolley railing and the pixel values ​​corresponding to the pixel points; According to the pixel value, filtering is performed on the pixel value to obtain a temperature value corresponding to the pixel point; According to the temperature value, if the temperature value is greater than or equal to the fault temperature threshold, the temperature value is marked as a fault temperature value; if the temperature value is less than the fault temperature threshold, the temperature value is marked as a normal temperature value, and the pixel corresponding to the fault temperature value is a fault pixel, and the pixel corresponding to the normal temperature value is a normal pixel; If all pixels are normal, the sintering machine trolley railing is normal. If there are faulty pixels, Based on the shape of the arrangement of the faulty pixels, if the shape is a hole formed by multiple pixels clustered together, or a strip formed by multiple pixels clustered together, the fault type is a perforation or a break; if the shape is a thread formed by multiple pixels connected together, the fault type is a crack. If the fault is a perforation or a fracture, the fault temperature value and the normal temperature value are binarized to obtain the area and position of the perforation or fracture area; If the fault is a crack, performing gradient division on the pixel points using a Laplace model to obtain a gradient image; The gradient image is converted into a binary image, morphological processing is performed on the binary image, pixel points corresponding to the crack are screened out, and data closure processing is performed on the pixel points corresponding to the crack to obtain the shape, position and area of ​​the crack.

2. A method for identifying and analyzing faults of a sintering machine trolley guardrail according to claim 1, characterized in that: In the filtering step, the filtering is implemented using a Gaussian filtering model, which is specifically: A coordinate system is established on the thermal infrared image set, with the vertex of the upper corner of the thermal infrared image as the origin, x and y as the position of the pixel point in the coordinate system, and σ as the variance of the coordinate point.

3. The method for fault identification and analysis of a sintering machine trolley guardrail according to claim 1, characterized in that: In the binarization step, the binarization step is specifically as follows: The faulty pixel is marked as 255 and marked in white, and the normal pixel is marked as 0 and marked in black. The conversion model is: T is the fault temperature threshold. The perforated or fractured area forms a color difference with the normal area. All perforated or fractured areas are displayed in white, and the normal area is displayed in black. The white part is located and the area is calculated.

4. A method for identifying and analyzing faults of a sintering machine trolley guardrail according to claim 3, characterized in that: The fault temperature threshold is a pre-collected fault panel, and the temperature value corresponding to the pixel point of the fault area of ​​the fault panel is obtained. The fault temperature threshold is T, 50°C≤T≤200°C.

5. The method for fault identification and analysis of a sintering machine trolley guardrail according to claim 1, characterized in that: The Laplace model is: satisfy The conditional points constitute the gradient image.

6. The method for fault identification and analysis of a sintering machine trolley guardrail according to claim 1, characterized in that: The steps of converting the gradient image into a binary image are specifically as follows: is the second-order differential expression of the pixel point, Perform convolution, the convolution model is: According to the gradient algorithm, pixel points whose gradient values ​​corresponding to the pixel points are greater than the gradient threshold are marked as 1, and pixel points whose gradient values ​​are less than or equal to the gradient threshold are marked as 0, thereby obtaining a binary image.

7. A fault identification and analysis system for a sintering machine trolley guardrail, characterized in that: The fault identification and analysis system for a sintering machine trolley guardrail is used to execute the fault identification and analysis method for a sintering machine trolley guardrail according to any one of claims 1 to 6, comprising an image acquisition module, an image recognition module, an image primary analysis module, an image secondary analysis module, and a verification module; The image acquisition module is used to obtain images captured by the visible light camera and the thermal infrared device; The image recognition module is used to recognize the image captured by the visible light camera. If the steel stamp number on the sintering machine frame is located in the middle of the captured image, the image captured by the thermal infrared device corresponding to the image captured by the visible light camera is stored as the thermal infrared image of the sintering machine guardrail, and the steel stamp number is stored as the first recognized steel stamp number; The image primary analysis module is used to perform image analysis on the thermal infrared image of the sintering machine guardrail, store the image analysis result, and mark the image analysis result as the first analysis result; The image secondary analysis module is used to identify the steel stamp number for the second time after the trolley passes through the sintering machine once, obtain the image collected by the thermal infrared device for the second time, perform image analysis on the image collected by the thermal infrared device for the second time, store the result of the image analysis, and mark the result of the image analysis as the second analysis result; The verification module is configured to output a final recognition analysis result if the first analysis result is the same as the second analysis result, and to recapture the image and perform recognition analysis if the first analysis result is different from the second analysis result; In the image primary analysis module and the image secondary analysis module, the units performing the image analysis step include: a raw data acquisition unit, a filtering unit, a temperature conversion unit, a fault type judgment unit, a binarization processing unit, a gradient image unit, and a closure processing unit; The raw data acquisition unit is used to set a fault temperature threshold of the thermal infrared image of the sintering machine trolley guardrail, and to acquire all pixel points in the thermal infrared image set of the sintering machine trolley guardrail and the pixel values ​​corresponding to the pixel points; The filtering unit is configured to perform filtering processing on the pixel value according to the pixel value to obtain a temperature value corresponding to the pixel point; The temperature conversion unit is configured to, based on the temperature value, mark the temperature value as a fault temperature value if the temperature value is greater than or equal to the fault temperature threshold; and mark the temperature value as a normal temperature value if the temperature value is less than the fault temperature threshold, and mark the pixel corresponding to the fault temperature value as a fault pixel, and mark the pixel corresponding to the normal temperature value as a normal pixel; The fault type judgment unit is used to determine that the sintering machine trolley guardrail is normal if all pixels are normal; If there are faulty pixels, based on the shape of the arrangement of the faulty pixels, if the shape is a hole formed by multiple pixels clustered together, or a strip formed by multiple pixels clustered together, the fault type is a perforation or a break; if the shape is a thread formed by multiple pixels connected together, the fault type is a crack; The binarization processing unit is configured to perform binarization processing on the fault temperature value and the normal temperature value if the fault is a perforation or a fracture, so as to obtain the area and position of the perforation or fracture area; The gradient image unit is configured to perform gradient division on the pixel points using a Laplace model to obtain a gradient image if the fault is a crack; The closure processing unit is used to convert the gradient image into a binary image, perform morphological processing on the binary image, filter out pixel points corresponding to the crack, and perform data closure processing on the pixel points corresponding to the crack to obtain the shape, position and area of ​​the crack.

8. A fault identification and analysis system for a sintering machine trolley guardrail according to claim 7, characterized in that: In the filtering step, the filtering is implemented using a Gaussian filtering model, which is specifically: A coordinate system is established on the thermal infrared image set, with the vertex of the upper corner of the thermal infrared image as the origin, x and y as the position of the pixel point in the coordinate system, and σ as the variance of the coordinate point.

9. The fault identification and analysis system for a sintering machine trolley guardrail according to claim 7, characterized in that: In the binarization step, the binarization step is specifically as follows: The faulty pixel is marked as 255 and marked in white, and the normal pixel is marked as 0 and marked in black. The conversion model is: T is the fault temperature threshold. The perforated or fractured area forms a color difference with the normal area. All perforated or fractured areas are displayed in white, and the normal area is displayed in black. The white part is located and the area is calculated.

10. A fault identification and analysis system for a sintering machine trolley guardrail according to claim 9, characterized in that: The fault temperature threshold is a pre-collected fault panel, and the temperature value corresponding to the pixel point of the fault area of ​​the fault panel is obtained. The fault temperature threshold is T, 50°C≤T≤200°C.

11. The fault identification and analysis system for a sintering machine trolley guardrail according to claim 7, characterized in that: The Laplace model is: satisfy The conditional points constitute the gradient image.

12. A fault identification and analysis system for a sintering machine trolley guardrail according to claim 7, characterized in that: The steps of converting the gradient image into a binary image are specifically as follows: is the second-order differential expression of the pixel point, Perform convolution, the convolution model is: According to the gradient algorithm, pixel points whose gradient values ​​corresponding to the pixel points are greater than the gradient threshold are marked as 1, and pixel points whose gradient values ​​are less than or equal to the gradient threshold are marked as 0, thereby obtaining a binary image.

Images