Method and system for judging furnace coking state

Abstract

The present invention provides a method and system for judging the coking state of a furnace, and belongs to the technical field of boilers. The method comprises: performing image segmentation on the target images corresponding to the maximum viewing angle range of each combustion layer of the furnace in each preset orientation, and obtaining a plurality of image blocks of each target image; performing texture feature extraction on each image block, and obtaining a texture feature vector of each image block; determining the image block proportion data corresponding to each preset label in each target image based on the texture feature vector of each image block of each target image and the texture feature vector corresponding to each preset label determined in advance; obtaining the coking state information of each combustion layer based on the image block proportion data corresponding to each preset label in each target image of each combustion layer; obtaining the coking state information of the furnace based on the coking state information of each combustion layer. Thus, the overall furnace can be intuitively and quantitatively grasped as a whole and in part, and the effect of accurately judging the coking state of the overall furnace can be achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of boilers, and in particular to a method for judging a furnace coking state, a system for judging a furnace coking state, a machine-readable storage medium and an electronic device. Background Art

[0002] Boiler coking is a common problem in boiler operation. Usually, the ash particles moving with the flue gas are cooled together with the flue gas due to the heat absorption of the heating surface of the furnace water-cooled wall. If the liquid slag particles have solidified due to the temperature drop before approaching the water-cooled wall or the furnace wall, when they adhere to the heating surface tube wall, a loose ash layer will be formed, which can be easily removed by blowing soot during operation. When the temperature in the furnace is high, some ash particles have reached a molten or semi-molten state. If these ash particles are not sufficiently cooled to a solidified state before reaching the heating surface, they have a high bonding ability and are easy to adhere to the flue gas scouring heating surface or the furnace wall, and even reach a molten state. The adhesion of molten or semi-molten ash particles and unburned coke causes coking to continue to develop.

[0003] During the operation of the combustion furnace, the temperature is too high and the furnace is closed, which makes it impossible to observe the furnace visually and comprehensively. The coking in the furnace will adhere to the water-cooled wall, causing the water-cooled wall to have a temperature difference, which can easily lead to cracking of the water-cooled wall and incomplete combustion. In addition, the temperature and combustion products of each combustion layer in the furnace are different, and the degree of coking in each layer is also different.

[0004] The prior art trains a convolutional neural network by acquiring temperature data of several measuring points of coking or non-coking in the same time period, and after inputting the sampled measuring point temperature data into the convolutional neural network, obtains the image features of the temperature data to determine whether it is coking or not. Its image features are feature images predicted by temperature measuring points, not intuitive images. In addition, the temperature measuring points are a small sample size data set relative to the entire combustion boiler. Therefore, this prediction method is not accurate enough and cannot subjectively display the coking state of the boiler.

[0005] The existing technology also uses the image through the flame high temperature filter to perform deep learning-based coking image recognition, and calculates the coking score of the result, and provides two different feedbacks: early warning and intervention according to the score result. However, this solution only recognizes the image results obtained by the filter. The recognition of this method is only a small window image recognition, and the pixel points of the coke block in the image are quantitatively analyzed, and the coking state of the entire furnace cannot be judged.

[0006] Therefore, how to accurately judge the coking state of the entire furnace is a problem that needs to be solved urgently. Summary of the invention

[0007] The purpose of the embodiments of the present invention is to provide a method and system for determining the coking state of a furnace, so as to at least solve the above-mentioned problem of how to accurately determine the coking state of the entire furnace.

[0008] In order to achieve the above object, the first aspect of the present invention provides a method for judging the coking state of a furnace, comprising:

[0009] Performing image segmentation on the target images corresponding to the maximum viewing angle range of each combustion layer in each preset orientation, and obtaining multiple image blocks of each target image corresponding to each combustion layer;

[0010] Extracting texture features of each image block to obtain a texture feature vector of each image block;

[0011] Based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the predetermined texture feature vectors corresponding to each preset label, determine the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer;

[0012] Based on the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer, the coking state information of each combustion layer is obtained;

[0013] Based on the coking state information of each combustion layer, the coking state information of the furnace is obtained.

[0014] Optionally, the acquisition rules of the target images corresponding to the maximum viewing angle range of each combustion layer in the furnace in each preset orientation include:

[0015] For each combustion layer of the furnace, the furnace appearance video data corresponding to each preset position is collected based on the inserted endoscope camera set at each preset position of the combustion layer; wherein the furnace appearance video data includes the video image of the corresponding inserted endoscope camera from the beginning of insertion into the furnace to the preset maximum depth;

[0016] The furnace appearance video data corresponding to each preset position of each combustion layer of the furnace is frame-cut and captured, and the captured images are used for image stitching to obtain the target image of the maximum viewing angle range corresponding to each preset position of each combustion layer of the furnace.

[0017] Optionally, the above-mentioned extraction of texture features from each image block to obtain a texture feature vector of each image block includes:

[0018] Divide each image block into multiple blocks of a 3x3 neighborhood, compare the pixel value of the central pixel of each block with the pixel values ​​of each adjacent pixel of the corresponding central pixel, and assign values ​​to each adjacent pixel of each central pixel based on the comparison result and a preset pixel assignment rule;

[0019] Based on the value assignment results of each adjacent pixel point of each central pixel point, a local binary pattern operator corresponding to each central pixel point of each image block is obtained;

[0020] Based on the local binary pattern operator corresponding to each central pixel point of each image block, a local binary pattern operator histogram corresponding to each image block is obtained;

[0021] After normalizing the local binary pattern operator histogram corresponding to each image block, feature vector connection is performed based on the local binary pattern operator histogram of each image block after the normalization process to obtain the texture feature vector corresponding to each image block.

[0022] Optionally, the above preset pixel assignment rule includes:

[0023] The pixel points in the 3x3 neighborhood block whose pixel value is greater than the pixel value of the corresponding central pixel are assigned a value of 1, otherwise they are assigned a value of 0.

[0024] Optionally, the above-mentioned local binary pattern operator corresponding to each central pixel of each image block is obtained based on the value assignment result of each adjacent pixel of each central pixel, including:

[0025] The assignment results of the adjacent pixels of each central pixel are accumulated to obtain the local binary pattern operator corresponding to each central pixel.

[0026] Optionally, the determination rules of the texture feature vectors corresponding to the preset labels include:

[0027] Collecting historical furnace appearance images corresponding to each preset label as feature sample images corresponding to each preset label;

[0028] The texture features of each feature sample image are extracted to obtain the texture feature vector corresponding to each preset label.

[0029] Optionally, the above-mentioned extraction of texture features from each feature sample image includes:

[0030] Dividing each characteristic sample image to obtain a plurality of image sub-blocks corresponding to each characteristic sample image;

[0031] For each pixel of each image sub-block, the pixel value of the pixel is compared with the pixel values ​​of each neighboring pixel of the pixel, and based on the comparison result and a preset pixel value assignment rule, each neighboring pixel of the pixel is assigned a value;

[0032] Based on the assignment results of each adjacent pixel point of each pixel point, a local binary pattern operator corresponding to each pixel point is obtained;

[0033] Based on the local binary pattern operator corresponding to each pixel point of each image sub-block, a local binary pattern operator histogram corresponding to each image sub-block is obtained;

[0034] After normalizing the local binary pattern operator histogram corresponding to each image sub-block, feature vector connection is performed based on the local binary pattern operator histogram of each image sub-block of each feature sample image after the normalization process to obtain the texture feature vector corresponding to each feature sample image.

[0035] Optionally, the above-mentioned determination of the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer based on the texture feature vector of each image block of each target image corresponding to each combustion layer and the predetermined texture feature vector corresponding to each preset label includes:

[0036] Based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the texture feature vectors corresponding to each predetermined preset label, the image blocks of each target image corresponding to each combustion layer are classified according to the preset label by using the K nearest neighbor classification algorithm to obtain the label of each image block of each target image corresponding to each combustion layer;

[0037] Based on the labels of the image blocks of the target images corresponding to the combustion layers, the number of image blocks corresponding to the preset labels in the target images corresponding to the combustion layers is counted;

[0038] Based on the statistical results of the number of image blocks corresponding to each preset label in each target image corresponding to each combustion layer, the proportion data of the image blocks corresponding to each preset label in each target image corresponding to each combustion layer is determined.

[0039] Optionally, the above-mentioned data on the proportion of image blocks corresponding to each preset label in each target image corresponding to each combustion layer is used to obtain the coking state information of each combustion layer, including:

[0040] The number of image blocks corresponding to the same preset label of each combustion layer is accumulated to obtain the total number of image blocks corresponding to each preset label of each combustion layer;

[0041] Based on the total amount of image blocks corresponding to each preset label corresponding to each combustion layer, the coking state information of each combustion layer is determined.

[0042] Optionally, the coking state information of the furnace is obtained based on the coking state information of each combustion layer, including:

[0043] The number of image blocks corresponding to the same preset label of all combustion layers is accumulated to obtain the total number of image blocks corresponding to each preset label corresponding to the furnace;

[0044] Based on the total amount of image blocks corresponding to each preset label corresponding to the furnace, the coking state information of the furnace is determined.

[0045] Optionally, the above method for judging the furnace coking state further includes:

[0046] Based on the total amount of image blocks corresponding to each preset label corresponding to each combustion layer, a statistical chart corresponding to each combustion layer is formed;

[0047] Based on the total amount of image blocks corresponding to each preset label corresponding to the furnace, a statistical chart corresponding to the furnace is formed.

[0048] A second aspect of the present invention provides a furnace coking state judgment system, comprising:

[0049] An image segmentation module is used to perform image segmentation on the target images corresponding to the maximum viewing angle range of each combustion layer in each preset orientation, and obtain multiple image blocks of each target image corresponding to each combustion layer;

[0050] A texture feature extraction module is used to extract texture features from each image block to obtain a texture feature vector of each image block;

[0051] An image block ratio determination module is used to determine image block ratio data corresponding to each preset label in each target image corresponding to each combustion layer based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the predetermined texture feature vectors corresponding to each preset label;

[0052] A combustion layer coking state determination module is used to obtain the coking state information of each combustion layer based on the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer;

[0053] The furnace coking state determination module is used to obtain the furnace coking state information based on the coking state information of each combustion layer.

[0054] In a third aspect of the present invention, a machine-readable storage medium is provided, on which instructions are stored, and when the instructions are executed by a processor, the processor is configured to execute the above-mentioned method for determining the coking state of a furnace.

[0055] In a fourth aspect of the present invention, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the furnace coking state judgment method when executing the computer program.

[0056] Through the above technical scheme, a method and system for judging the coking state of a furnace is provided, which performs image segmentation on the target images of the maximum viewing angle range of each combustion layer of the furnace collected at each preset position, and obtains multiple image blocks corresponding to each target image. The target image with the maximum viewing angle range can ensure that the appearance image of each combustion layer of the furnace is more comprehensive, which is conducive to ensuring the accuracy of the subsequent judgment of the internal coking state of the furnace. Texture feature extraction is performed on each image block to obtain the texture feature vector of each image block. According to the texture feature vector of each image block of each target image, the texture feature comparison is performed in the texture feature vector corresponding to each preset label in advance to obtain the image block proportion data corresponding to each preset label in each target image. According to the image block proportion data corresponding to each preset label in each target image of each combustion layer, the coking state information of each combustion layer is obtained. According to the coking state information of each combustion layer, the coking state information of the furnace is obtained. The method and system use texture feature extraction and comparison to retrieve the labels of each acquired image block, thereby improving the accuracy. This method and system does not directly judge whether the target image is coked or not, but realizes the quantitative evaluation of the coking state inside the furnace by counting the non-coking label, fly ash attachment label, and coking label in the target image. In this method and system, the quantitative evaluation of each combustion layer can also be accumulated into the quantitative evaluation of the entire furnace, and the coking state of the water-cooled wall of each combustion layer can be grasped separately, and the coking state of all water-cooled walls inside the furnace can also be grasped, so as to have an intuitive and quantitative grasp of the entire furnace and its parts, and thus achieve the effect of accurately judging the coking state of the entire furnace.

[0057] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0059] Figure 1 It is a flow chart of a method for judging the coking state of a furnace provided by one embodiment of the present invention;

[0060] Figure 2 It is a front view of an installation position of an in-depth endoscope camera head provided by one embodiment of the present invention;

[0061] Figure 3 It is a top view of an installation position of an in-depth endoscope camera head provided by one embodiment of the present invention;

[0062] Figure 4It is a block diagram of a furnace coking state judgment system provided by one embodiment of the present invention;

[0063] Figure 5 It is a schematic diagram of the structure of an electronic device provided by a preferred embodiment of the present invention.

[0064] Description of Reference Numerals

[0065] 10 - electronic device, 100 - processor, 101 - memory, 102 - computer program. DETAILED DESCRIPTION

[0066] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0067] Figure 1 FIG. 1 is a flow chart of a method for determining the coking state of a furnace provided by an embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides a method for judging the coking state of a furnace, comprising:

[0068] S110: performing image segmentation on the collected target images corresponding to the maximum viewing angle range of each combustion layer in each preset orientation of the furnace to obtain multiple image blocks of each target image corresponding to each combustion layer;

[0069] Specifically, each target image is segmented into different sub-blocks (for example, N*M sub-blocks), thereby obtaining a plurality of image blocks corresponding to each target image.

[0070] In some implementations of the present embodiment, the collection rules of the target images corresponding to the maximum viewing angle range of each combustion layer in the above-mentioned furnace in each preset orientation include: for each combustion layer in the furnace, based on the inserted endoscope camera arranged at each preset orientation of the combustion layer, the furnace appearance video data corresponding to each preset orientation is collected; wherein the furnace appearance video data includes the video image of the corresponding inserted endoscope camera from the beginning of inserting into the furnace to the preset maximum depth; the furnace appearance video data corresponding to each preset orientation of each combustion layer in the furnace is frame-cut, and the cut images are used for image stitching to obtain the target images of the maximum viewing angle range corresponding to each preset orientation of each combustion layer in the furnace.

[0071] Among them, the invasive endoscope camera is a visible light camera.

[0072] Please refer to Figure 2 and Figure 3 , Figure 2 is a front view of an installation position of an endoscopic camera head provided by an embodiment of the present invention, Figure 3It is a top view of the installation position of an endoscopic camera provided by an embodiment of the present invention. Specifically, observation holes are opened at the four corners of each combustion layer of the furnace of the boiler, and an endoscopic camera is provided to extend into the furnace through the observation holes to shoot the appearance image of the inside of the furnace, so as to collect the furnace appearance video data inside the furnace. The video image acquisition process of the endoscopic camera located at different positions and different combustion layers of the furnace is as follows: the endoscopic camera starts to shoot and record the video from the moment it is inserted into the furnace, and the video images until it reaches the preset maximum depth (wherein the preset maximum depth can be the maximum depth position that the endoscopic camera can extend into the corresponding combustion layer) are saved, that is, the video images of the endoscopic camera under all the insertion paths are recorded. And by extracting and intercepting the video data of each furnace appearance, the intercepted images are spliced ​​by image stitching technology to form a panoramic picture, that is, a picture of the maximum viewing angle range that can be observed by the endoscopic camera under the full path, which is the target image. Observation holes are set at the four corners of each combustion layer, and each observation hole can obtain a full-path furnace internal water-cooled wall appearance picture with the maximum viewing angle as the target image. If there are i combustion layers, there are 4*i target images. This method obtains the internal image of the furnace with the maximum viewing angle, which provides image data with the maximum viewing angle for grasping the overall coking state of the furnace. The real and intuitive state of the furnace can be captured by the inserted endoscope camera, and the full-view images under the insertion path of the inserted endoscope camera can be displayed through the image stitching technology, so that the internal image of the furnace can be obtained with the maximum area. The more comprehensive the appearance image of the internal furnace is, the more accurate the judgment of the internal coking state of the furnace will be.

[0073] In the above implementation process, the images acquired by the inserted endoscope cameras located at different positions and different combustion layers of the furnace are spliced ​​in the manner of image frame extraction and splicing. The insertion paths of the inserted endoscope cameras are consistent, the viewing angles of the inserted endoscope cameras are consistent, and the number of frames to be spliced ​​is consistent, so the size of the final spliced ​​picture of each inserted endoscope camera is consistent, that is, the size of the target image of the maximum viewing angle range of the entire path of the inserted endoscope camera is consistent. And each combustion layer has four target images, for example: the images of the A combustion layer are A1, A2, A3, A4; the images of the B combustion layer are B1, B2, B3, B4, and the C, D, E and other layers are like this.

[0074] S120: extracting texture features from each image block to obtain a texture feature vector of each image block;

[0075] In some implementations of the present embodiment, the above-mentioned texture feature extraction of each image block to obtain the texture feature vector of each image block includes: dividing each image block into multiple blocks of a 3x3 neighborhood, comparing the pixel value of the central pixel of each block with the pixel value of each adjacent pixel corresponding to the central pixel, and assigning a value to each adjacent pixel of each central pixel based on the comparison result and a preset pixel assignment rule; obtaining a local binary pattern operator corresponding to each central pixel of each image block based on the assignment result of each adjacent pixel of each central pixel; obtaining a local binary pattern operator histogram corresponding to each image block based on the local binary pattern operator corresponding to each central pixel of each image block; after normalizing the local binary pattern operator histogram corresponding to each image block, feature vector connection is performed based on the local binary pattern operator histogram of each image block after the normalization to obtain the texture feature vector corresponding to each image block.

[0076] Furthermore, the above-mentioned preset pixel assignment rule includes: assigning a pixel value of 1 to a pixel in a 3x3 neighborhood block whose pixel value is greater than the pixel value of a corresponding central pixel, and assigning a value of 0 otherwise.

[0077] Furthermore, the above-mentioned local binary pattern operator corresponding to each central pixel of each image block is obtained based on the assignment results of each adjacent pixel point of each central pixel point, including: accumulating the assignment results of each adjacent pixel point corresponding to each central pixel point to obtain the local binary pattern operator corresponding to each central pixel point.

[0078] Specifically, for each image block, each image block is regarded as a plurality of square blocks in a 3x3 neighborhood, and the pixel value of the central pixel of each block is compared with the pixel value of each corresponding adjacent pixel, and the pixel point in the square block of the 3x3 neighborhood whose pixel value is greater than the pixel value of the corresponding central pixel is assigned 1, otherwise it is assigned 0, and then the 8 points in the 3x3 neighborhood can generate an 8-bit binary number after comparison, and the assignment results of each adjacent pixel point corresponding to each central pixel are accumulated to obtain the local binary pattern operator of the central pixel point of the window. Then, the local binary pattern operator histogram of each image block is calculated, that is, the frequency of each operator in each image block, and finally the local binary pattern operator histogram is normalized, and the local binary pattern operator histogram of each image block is connected into a feature vector, that is, the texture feature vector of the corresponding image block.

[0079] S130: determining the image block proportion data corresponding to each preset label in each target image corresponding to each burning layer based on the texture feature vector of each image block of each target image corresponding to each burning layer and the predetermined texture feature vector corresponding to each preset label;

[0080] In some implementations of the present embodiment, the above-mentioned determination of the proportion data of image blocks corresponding to each preset label in each target image corresponding to each combustion layer based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the texture feature vectors corresponding to each preset label determined in advance includes: based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the texture feature vectors corresponding to each preset label determined in advance, using the K nearest neighbor classification algorithm to perform preset label classification on each image block of each target image corresponding to each combustion layer to obtain the label of each image block of each target image corresponding to each combustion layer; based on the label of each image block of each target image corresponding to each combustion layer, counting the number of image blocks corresponding to each preset label in each target image corresponding to each combustion layer; based on the statistical result of the number of image blocks corresponding to each preset label in each target image corresponding to each combustion layer, determining the proportion data of image blocks corresponding to each preset label in each target image corresponding to each combustion layer.

[0081] Among them, the preset labels include no-coking labels, fly ash attachment labels and coking labels.

[0082] Specifically, according to the K nearest neighbor classification algorithm, the texture feature vector of each image block is compared with the texture feature vector corresponding to each predetermined preset label to classify each image block and determine the label of each image block. Since the number of image blocks corresponding to each target image is N*M, the labels of N*M image blocks can be obtained. Then the number of image blocks corresponding to each preset label of the entire target image is counted to obtain the total number of three types of labels in the complete target image in the total number of N*M image blocks in the target image: the number of non-coking labels, the number of fly ash attached labels, and the number of coking labels. The cumulative total number of the three types of labels is N*M. Therefore, the proportion of each label in the target image can be calculated, which can quantitatively describe the focusing state of the target image, that is, the percentage of non-coking, the percentage of fly ash attached, and the percentage of coking in the target image under the corresponding extended endoscope camera angle.

[0083] It should be noted that this method uses texture feature extraction and comparison to retrieve the labels of each image block. Since the water-cooled wall of the furnace is normally in the form of parallel water pipes, which is a striped texture; fly ash is generally attached to the gaps of the texture, which is also reflected in the texture style; coke blocks also have special textures, so this method uses texture retrieval to improve the accuracy.

[0084] S140: Obtaining coking state information of each combustion layer based on image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer;

[0085] In some implementations of the present embodiment, the above-mentioned data on the proportion of image blocks corresponding to each preset label in each target image corresponding to each combustion layer is used to obtain the coking status information of each combustion layer, including: adding up the number of image blocks corresponding to the same preset label of each combustion layer to obtain the total number of image blocks corresponding to each preset label corresponding to each combustion layer; and determining the coking status information of each combustion layer based on the total number of image blocks corresponding to each preset label corresponding to each combustion layer.

[0086] Specifically, each combustion layer has four endoscopic perspectives. For each combustion layer, the number of different labels of the corresponding four target images is accumulated to obtain the coking-free ratio, fly ash adhesion ratio and coking ratio of the furnace in different combustion layers. Statistical charts, such as histograms and pie charts, can also be formed to intuitively and quantitatively grasp the coking status of each combustion layer.

[0087] S150: Based on the coking state information of each combustion layer, the coking state information of the furnace is obtained.

[0088] In some implementations of the present embodiment, the coking status information of the furnace is obtained based on the coking status information of each combustion layer, including: accumulating the number of image blocks corresponding to the same preset label of all combustion layers to obtain the total number of image blocks corresponding to each preset label corresponding to the furnace; determining the coking status information of the furnace based on the total number of image blocks corresponding to each preset label corresponding to the furnace.

[0089] Specifically, by adding up the number of different labels of all combustion layers, we can get the percentage of non-coking, percentage of fly ash adhesion and percentage of coking of the whole furnace. Similarly, we can also form statistical charts, such as histograms, pie charts, etc., so as to intuitively and quantitatively grasp the coking status of the whole furnace.

[0090] In the above implementation process, the method performs image segmentation on the target images of the maximum viewing angle range of each combustion layer of the furnace collected at each preset orientation, and obtains multiple image blocks corresponding to each target image. The target image with the maximum viewing angle range can ensure that the appearance image of each combustion layer of the furnace is more comprehensive, which is conducive to ensuring the accuracy of the subsequent judgment of the internal coking state of the furnace. Texture feature extraction is performed on each image block to obtain the texture feature vector of each image block. According to the texture feature vector of each image block of each target image, texture feature comparison is performed in the texture feature vector corresponding to each predetermined preset label to obtain the image block proportion data corresponding to each preset label in each target image. According to the image block proportion data corresponding to each preset label in each target image of each combustion layer, the coking state information of each combustion layer is obtained. According to the coking state information of each combustion layer, the coking state information of the furnace is obtained. The method uses texture feature extraction and comparison to retrieve the labels of each acquired image block, thereby improving the accuracy. This method does not directly classify the target image to determine whether it is coked, but rather achieves a quantitative evaluation of the coking state inside the furnace by counting the non-coking labels, fly ash attachment labels, and coking labels in the target image. In this method, the quantitative evaluation of each combustion layer can also be accumulated into a quantitative evaluation of the entire furnace, and the coking state of the water-cooled wall of each combustion layer can be grasped separately, and the coking state of all water-cooled walls inside the furnace can also be grasped, so that the overall and partial intuitive and quantitative grasp of the entire furnace is achieved, and the effect of accurately judging the coking state of the entire furnace is achieved.

[0091] In some implementations of the present embodiment, the rules for determining the texture feature vector corresponding to each preset label include: collecting historical furnace appearance images corresponding to each preset label as feature sample images corresponding to each preset label; extracting texture features from each feature sample image to obtain the texture feature vector corresponding to each preset label.

[0092] Specifically, from the collected historical furnace appearance images, the textured parts are labeled according to the coking state. Among them, from normal to coking, they can be divided into three labels: "no coking", "fly ash adhesion", and "coking" as feature sample images. Texture features are extracted from each feature sample image corresponding to each preset label, so as to obtain the texture feature vector corresponding to each preset label.

[0093] In some implementations of the present embodiment, the above-mentioned texture feature extraction of each feature sample image includes: performing image division on each feature sample image to obtain multiple image sub-blocks corresponding to each feature sample image; for each pixel point of each image sub-block, comparing the pixel value of the pixel point with the pixel values ​​of each adjacent pixel point of the pixel point, and assigning values ​​to each adjacent pixel point of the pixel point based on the comparison result and a preset pixel assignment rule; obtaining a local binary pattern operator corresponding to each pixel point based on the assignment result of each adjacent pixel point of each pixel point; obtaining a local binary pattern operator histogram corresponding to each image sub-block based on the local binary pattern operator corresponding to each pixel point of each image sub-block; after normalizing the local binary pattern operator histogram corresponding to each image sub-block, feature vector connection is performed based on the local binary pattern operator histogram of each image sub-block of each feature sample image after the normalization process to obtain a texture feature vector corresponding to each feature sample image.

[0094] In detail, the feature sample image is locally binarized. The specific process is as follows: the feature sample image is divided into several 16×16 image sub-blocks. For each image sub-block, the grayscale values ​​of the adjacent 8 pixels are compared with its central pixel value. If the surrounding pixel value is greater than the central pixel value, the position of the pixel is marked as 1, otherwise it is 0. In this way, the 8 points in the 3*3 neighborhood can generate an 8-bit binary number after comparison, and the local binary pattern operator of the central pixel of the window is obtained by accumulating it. Then the histogram of the local binary pattern operator of each image sub-block is calculated, that is, the frequency of each operator. Finally, the histogram is normalized, and the histogram of each image sub-block is connected into a feature vector, which is the texture feature vector of the feature sample image.

[0095] In some implementations of the present embodiment, the above-mentioned method for judging the coking state of the furnace also includes: forming a statistical chart corresponding to each combustion layer based on the total amount of image blocks corresponding to each preset label corresponding to each combustion layer; forming a statistical chart corresponding to the furnace based on the total amount of image blocks corresponding to each preset label corresponding to the furnace.

[0096] Figure 4 FIG. 1 is a block diagram of a furnace coking state judgment system provided by an embodiment of the present invention. Figure 4 As shown, an embodiment of the present invention provides a furnace coking state judgment system, comprising:

[0097] An image segmentation module is used to perform image segmentation on the target images corresponding to the maximum viewing angle range of each combustion layer in each preset orientation, and obtain multiple image blocks of each target image corresponding to each combustion layer;

[0098] A texture feature extraction module is used to extract texture features from each image block to obtain a texture feature vector of each image block;

[0099] An image block ratio determination module is used to determine image block ratio data corresponding to each preset label in each target image corresponding to each combustion layer based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the predetermined texture feature vectors corresponding to each preset label;

[0100] A combustion layer coking state determination module is used to obtain the coking state information of each combustion layer based on the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer;

[0101] The furnace coking state determination module is used to obtain the furnace coking state information based on the coking state information of each combustion layer.

[0102] Specifically, the system performs image segmentation on the target images of the maximum viewing angle range of each combustion layer of the furnace collected at each preset orientation, and obtains multiple image blocks corresponding to each target image. The target images with the maximum viewing angle range can ensure that the appearance images of each combustion layer of the furnace are more comprehensive, which is conducive to ensuring the accuracy of the subsequent judgment of the internal coking state of the furnace. Texture feature extraction is performed on each image block to obtain the texture feature vector of each image block. According to the texture feature vectors of each image block of each target image, texture feature comparison is performed in the texture feature vectors corresponding to each predetermined preset label to obtain the image block proportion data corresponding to each preset label in each target image. According to the image block proportion data corresponding to each preset label in each target image of each combustion layer, the coking state information of each combustion layer is obtained. According to the coking state information of each combustion layer, the coking state information of the furnace is obtained. The system uses texture feature extraction and comparison to retrieve the labels of each acquired image block, thereby improving the accuracy. The system does not directly judge whether the target image is coked or not, but realizes the quantitative evaluation of the coking state inside the furnace by counting the non-coking label, fly ash attachment label, and coking label in the target image. In this system, the quantitative evaluation of each combustion layer can also be accumulated into the quantitative evaluation of the entire furnace, and the coking state of the water-cooled wall of each combustion layer can be grasped separately, and the coking state of all water-cooled walls inside the furnace can also be grasped, so as to have an intuitive and quantitative grasp of the entire furnace and its parts, and achieve the effect of accurately judging the coking state of the entire furnace.

[0103] The embodiment of the present invention further provides a machine-readable storage medium, on which instructions are stored. When the instructions are executed by the processor 100, the processor 100 is configured to execute the above-mentioned method for determining the coking state of the furnace.

[0104] Machine-readable storage media include permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0105] An embodiment of the present invention further provides an electronic device 10, which includes a memory 101, a processor 100, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, the above-mentioned method for determining the coking state of the furnace is implemented.

[0106] like Figure 5 FIG. 1 is a schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 5 As shown, the electronic device 10 of this embodiment includes: a processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, the steps in the above method embodiment are implemented. Alternatively, when the processor 100 executes the computer program 102, the functions of each module / unit in the above device embodiment are implemented.

[0107] Exemplarily, the computer program 102 may be divided into one or more modules / units, one or more modules / units are stored in the memory 101, and are executed by the processor 100 to complete the present invention. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 102 in the electronic device 10. For example, the computer program 102 may be divided into an image segmentation module, a texture feature extraction module, an image block ratio determination module, a combustion layer coking state determination module, and a furnace coking state determination module.

[0108] The electronic device 10 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The electronic device 10 may include, but is not limited to, a processor 100 and a memory 101. Those skilled in the art will appreciate that Figure 5 It is only an example of the electronic device 10 and does not constitute a limitation of the electronic device 10. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0109] The processor 100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0110] The memory 101 may be an internal storage unit of the electronic device 10, such as a hard disk or memory of the electronic device 10. The memory 101 may also be an external storage device of the electronic device 10, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card (FlashCard), etc. equipped on the electronic device 10. Further, the memory 101 may also include both an internal storage unit of the electronic device 10 and an external storage device. The memory 101 is used to store computer programs and other programs and data required by the electronic device 10. The memory 101 may also be used to temporarily store data that has been output or is to be output.

[0111] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0112] It should be understood by those skilled in the art that the embodiments of the present application may be provided as methods, systems, or computer program 102 products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program 102 product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0113] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program 102 products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by the computer program 102 instructions. These computer program 102 instructions can be provided to a processor 100 of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor 100 of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program 102 instructions may also be stored in a computer readable memory 101 that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory 101 produce an article of manufacture including an instruction device that implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] These computer program 102 instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0116] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0117] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for judging the coking state of a furnace, characterized in that: include: Performing image segmentation on the target images corresponding to the maximum viewing angle range of each combustion layer in each preset orientation, and obtaining multiple image blocks of each target image corresponding to each combustion layer; Extracting texture features of each image block to obtain a texture feature vector of each image block; Based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the predetermined texture feature vectors corresponding to each preset label, determine the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer; Based on the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer, the coking state information of each combustion layer is obtained; Based on the coking state information of each combustion layer, the coking state information of the furnace is obtained.

2. The method for judging the coking state of a furnace according to claim 1, characterized in that: The acquisition rules of the target images corresponding to the maximum viewing angle range of each combustion layer in the furnace in each preset orientation include: For each combustion layer of the furnace, the furnace appearance video data corresponding to each preset position is collected based on the inserted endoscope camera set at each preset position of the combustion layer; wherein the furnace appearance video data includes the video image of the corresponding inserted endoscope camera from the beginning of insertion into the furnace to the preset maximum depth; The furnace appearance video data corresponding to each preset position of each combustion layer of the furnace is frame-cut and captured, and the captured images are used for image stitching to obtain the target image of the maximum viewing angle range corresponding to each preset position of each combustion layer of the furnace.

3. The method for judging the coking state of a furnace according to claim 1, characterized in that: The step of extracting texture features from each image block to obtain a texture feature vector of each image block includes: Divide each image block into multiple blocks of a 3x3 neighborhood, compare the pixel value of the central pixel of each block with the pixel values ​​of each adjacent pixel of the corresponding central pixel, and assign values ​​to each adjacent pixel of each central pixel based on the comparison result and a preset pixel assignment rule; Based on the value assignment results of each adjacent pixel point of each central pixel point, a local binary pattern operator corresponding to each central pixel point of each image block is obtained; Based on the local binary pattern operators corresponding to the central pixels of each image block, a local binary pattern operator histogram corresponding to each image block is obtained; After normalizing the local binary pattern operator histogram corresponding to each image block, feature vector connection is performed based on the normalized local binary pattern operator histogram of each image block to obtain a texture feature vector corresponding to each image block.

4. The method for judging the coking state of a furnace according to claim 3, characterized in that: The preset pixel assignment rules include: The pixel points in the 3x3 neighborhood block whose pixel value is greater than the pixel value of the corresponding central pixel are assigned a value of 1, otherwise they are assigned a value of 0.

5. The method for judging the coking state of a furnace according to claim 3, characterized in that: The method of obtaining the local binary pattern operator corresponding to each central pixel of each image block based on the value assignment results of each adjacent pixel of each central pixel includes: The assignment results of the adjacent pixels of each central pixel are accumulated to obtain the local binary pattern operator corresponding to each central pixel.

6. The method for judging the coking state of a furnace according to claim 1, characterized in that: The rules for determining the texture feature vector corresponding to each preset label include: Collecting historical furnace appearance images corresponding to each preset label as feature sample images corresponding to each preset label; The texture features of each feature sample image are extracted to obtain the texture feature vector corresponding to each preset label.

7. The method for judging the coking state of a furnace according to claim 6, characterized in that: The extracting of texture features from each feature sample image comprises: Dividing each characteristic sample image to obtain a plurality of image sub-blocks corresponding to each characteristic sample image; For each pixel of each image sub-block, the pixel value of the pixel is compared with the pixel values ​​of each neighboring pixel of the pixel, and based on the comparison result and a preset pixel value assignment rule, each neighboring pixel of the pixel is assigned a value; Based on the assignment results of each adjacent pixel point of each pixel point, a local binary pattern operator corresponding to each pixel point is obtained; Based on the local binary pattern operator corresponding to each pixel point of each image sub-block, a local binary pattern operator histogram corresponding to each image sub-block is obtained; After normalizing the local binary pattern operator histogram corresponding to each image sub-block, feature vector connection is performed based on the normalized local binary pattern operator histogram of each image sub-block of each feature sample image to obtain the texture feature vector corresponding to each feature sample image.

8. The method for judging the coking state of a furnace according to claim 1, characterized in that: The method of determining the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer based on the texture feature vector of each image block of each target image corresponding to each combustion layer and the predetermined texture feature vector corresponding to each preset label comprises: Based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the texture feature vectors corresponding to each predetermined preset label, the image blocks of each target image corresponding to each combustion layer are classified according to the preset label by using the K nearest neighbor classification algorithm to obtain the label of each image block of each target image corresponding to each combustion layer; Based on the labels of the image blocks of the target images corresponding to the combustion layers, the number of image blocks corresponding to the preset labels in the target images corresponding to the combustion layers is counted; Based on the statistical results of the number of image blocks corresponding to each preset label in each target image corresponding to each combustion layer, the proportion data of the image blocks corresponding to each preset label in each target image corresponding to each combustion layer is determined.

9. The method for judging the coking state of a furnace according to claim 8, characterized in that: The coking state information of each combustion layer is obtained based on the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer, including: The number of image blocks corresponding to the same preset label of each combustion layer is accumulated to obtain the total number of image blocks corresponding to each preset label of each combustion layer; Based on the total amount of image blocks corresponding to each preset label corresponding to each combustion layer, the coking state information of each combustion layer is determined.

10. The method for judging the coking state of a furnace according to claim 9, characterized in that: The coking state information of the furnace is obtained based on the coking state information of each combustion layer, including: The number of image blocks corresponding to the same preset label of all combustion layers is accumulated to obtain the total number of image blocks corresponding to each preset label corresponding to the furnace; Based on the total amount of image blocks corresponding to each preset label corresponding to the furnace, the coking state information of the furnace is determined.

11. The method for judging the coking state of a furnace according to claim 10, characterized in that: Also includes: Based on the total amount of image blocks corresponding to each preset label corresponding to each combustion layer, a statistical chart corresponding to each combustion layer is formed; Based on the total amount of image blocks corresponding to each preset label corresponding to the furnace, a statistical chart corresponding to the furnace is formed.

12. A furnace coking state judgment system, characterized in that: include: An image segmentation module is used to perform image segmentation on the target images corresponding to the maximum viewing angle range of each combustion layer in each preset orientation, and obtain multiple image blocks of each target image corresponding to each combustion layer; A texture feature extraction module is used to extract texture features from each image block to obtain a texture feature vector of each image block; An image block ratio determination module is used to determine image block ratio data corresponding to each preset label in each target image corresponding to each combustion layer based on the texture feature vectors of each image block of each target image corresponding to each combustion layer and the predetermined texture feature vectors corresponding to each preset label; A combustion layer coking state determination module is used to obtain the coking state information of each combustion layer based on the image block proportion data corresponding to each preset label in each target image corresponding to each combustion layer; The furnace coking state determination module is used to obtain the furnace coking state information based on the coking state information of each combustion layer.

13. A machine-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by a processor, the processor is configured to execute the method for determining the furnace coking state as described in any one of claims 1 to 11.

14. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for determining the furnace coking state as described in any one of claims 1 to 11 is implemented.

Images