A method and device for visually detecting material particle size of an online blast furnace feeding belt

Abstract

The present invention discloses an online blast furnace feeding belt material particle size visual detection method and device, including obtaining the original image of the production raw material on the blast furnace feeding conveyor belt through a camera; performing guided filtering processing on the original image to obtain a filtered image; performing binarization processing on the filtered image to obtain a binary image; optimizing the binary image using a morphological processing method and an area filtering method to obtain an optimized image; performing distance transformation on the optimized image to obtain a distance image; obtaining points with distance values ​​greater than a preset value from the distance image, performing watershed transformation on the filtered image with the points as starting seed points to obtain a watershed segmentation result; marking and area statistics on the watershed segmentation result to obtain the particle size data of the production raw material. The present invention can automatically and accurately detect the particle size of the production raw material, and improves the timeliness and accuracy of the production raw material particle size detection.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence detection, and specifically to an online blast furnace feeding belt material particle size visual detection method and device. Background Art

[0002] Sintered ore and coke are the main raw materials for blast furnace smelting, and the size distribution of their ore blocks directly affects the blast furnace condition and molten iron quality. At present, the raw material particle size (also called particle size) of the sintering plant mainly relies on manual sampling for analysis and monitoring, so there are problems such as small sampling volume, incomplete data, and serious lag in data analysis. Due to the long process flow, many influencing factors and complex mechanism of the sintering process, the manual sampling method cannot be used to guide the sintering batching operation in a timely and effective manner, so the batching operation still relies on the accumulated inherent experience. Therefore, how to efficiently and accurately analyze the particle size of sintered ore and coke in order to guide the sintering batching operation in a timely manner has become a technical problem that needs to be solved in this field. Summary of the invention

[0003] In view of the above problems, the present application provides an online blast furnace feeding belt material particle size visual detection method, which is used to solve the technical problems of low efficiency and delayed analysis in that the raw materials for blast furnace smelting rely on manual sampling for particle size analysis.

[0004] To achieve the above purpose, the present application provides: an online blast furnace feeding belt material particle size visual detection method, comprising the following steps:

[0005] The original image of the production material on the blast furnace feed conveyor belt is obtained by the camera;

[0006] Performing guided filtering on the original image to remove noise in the original image to obtain a filtered image;

[0007] Binarizing the filtered image to obtain a binary image;

[0008] The binary image is optimized by using a morphological processing method and an area filtering method to remove holes between the production materials and obtain an optimized image;

[0009] Performing distance transformation on the optimized image to obtain a distance image;

[0010] Acquire points whose distance values ​​are greater than a preset value from the distance image, and perform watershed transformation on the filtered image with the points as starting seed points to obtain a watershed segmentation result;

[0011] The watershed segmentation result is marked and area statistics are performed to obtain the particle size data of the production raw material.

[0012] Furthermore, the watershed segmentation result is marked and area statistics are performed to obtain the particle size data of the production raw material, including:

[0013] Obtaining each closed region in the watershed segmentation result, and calculating the area of ​​each closed region;

[0014] Each of the closed regions is marked according to the area, and the area is converted into a particle size of the production material.

[0015] Furthermore, the step of obtaining each closed area in the watershed segmentation result includes:

[0016] Removing the first unclosed area located at the boundary in the watershed segmentation result, and obtaining each first closed area whose contour is completely closed and each second unclosed area which is not completely closed in the watershed segmentation result;

[0017] Determine whether the length ratio of the unclosed portion in each of the second unclosed regions is less than a preset value, and if so, fit the second unclosed regions into completely closed second closed regions;

[0018] The first closed area and the second closed area are obtained.

[0019] Furthermore, the method of optimizing the binary image by using a morphological processing method and an area filtering method to remove holes between the production materials includes:

[0020] Firstly, the binary image is optimized by using a morphological processing method to remove first holes between the production materials that are smaller than the first hole diameter;

[0021] Then, area filtering is performed to remove second holes between the production materials that are larger than the first aperture and smaller than the second aperture, wherein the second aperture is larger than the first aperture.

[0022] Furthermore, the binarization process is an adaptive threshold binarization process.

[0023] Furthermore, the method further comprises the steps of:

[0024] According to the result of the area statistics, a plurality of particle size intervals are divided, and the proportion of the production raw material in each of the particle size intervals is calculated.

[0025] Furthermore, before dividing the plurality of particle size intervals according to the area statistics result, the method further includes the following steps:

[0026] Identify the type of the production raw material according to the original image, where the type of the production raw material includes sintered ore or coke;

[0027] The step of dividing a plurality of particle size intervals according to the result of the area statistics and calculating the proportion of the production raw material in each particle size interval comprises:

[0028] Different particle size intervals are set according to different types of production raw materials; and the proportion of different types of production raw materials in each corresponding particle size interval is calculated.

[0029] Furthermore, dividing the area statistics into multiple particle size intervals and calculating the proportion of the production raw materials in each particle size interval also includes:

[0030] A mask is generated according to the result of the area statistics, and the mask is covered on the surface of the original image or the surface of the filtered image to obtain a mask image, wherein at least two colors are used in the mask to mark the production materials of different particle sizes.

[0031] In order to solve the above technical problems, this application also provides another technical solution:

[0032] An online blast furnace feeding belt material particle size visual detection device, comprising:

[0033] An image acquisition module is arranged above the blast furnace feed conveyor belt and is used to obtain the original image of the production raw materials on the blast furnace feed conveyor belt;

[0034] The visual algorithm module is connected to the image acquisition module and is used to perform guided filtering processing on the original image to remove noise from the image and obtain a filtered image; binarize the filtered image to obtain a binary image including a foreground image and a background image, wherein the production raw materials are located in the foreground image; optimize the binary image using a morphological processing method and an area filtering method to remove holes between the production raw materials and obtain an optimized image; perform distance transformation on the optimized image to obtain a distance image; obtain points with distance values ​​greater than a preset value from the distance image, and perform watershed transformation on the filtered image with the points as starting seed points to obtain a watershed segmentation result; mark and perform area statistics on the watershed segmentation result to obtain the particle size data of the production raw materials.

[0035] Furthermore, the image acquisition module includes a camera and a fill light, and the camera and the fill light are installed on a gantry bracket, and the gantry bracket spans above the blast furnace feed conveyor belt.

[0036] Different from the prior art, the above technical solution can analyze and detect the particle size of the raw materials on the blast furnace feed conveyor belt in real time. The detection process does not require manual intervention, which effectively solves the problems of lag, small detection volume and inaccurate data in manual sampling detection. In addition, the above technical solution can remove the noise in the original image by performing guided filtering on the original image, and can avoid the detection error caused by the smoothing of the edge of the raw material; and the above solution can remove the holes between the raw materials by binarizing the filtered image, and optimize the processing through morphological processing methods and area filtering, so as to obtain a clear outline of the raw material, thereby improving the accuracy of subsequent watershed segmentation, thereby improving the accuracy of raw material particle size detection.

[0037] The above-mentioned records related to the invention content are only an overview of the technical solution of the present application. In order to enable ordinary technicians in the field to more clearly understand the technical solution of the present application, and then implement it according to the text of the specification and the contents recorded in the drawings, and to make the above-mentioned purpose and other purposes, features and advantages of the present application easier to understand, the following is an explanation in combination with the specific implementation mode and drawings of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings are only used to illustrate the principles, implementation methods, applications, characteristics and effects of the specific embodiments of the present invention and other related contents, and shall not be considered as limiting the present application.

[0039] In the drawings of the specification:

[0040] Figure 1 A positional relationship diagram of a blast furnace feed conveyor belt and a camera according to a specific implementation method;

[0041] Figure 2 It is a flow chart of the on-line blast furnace feeding belt material particle size visual detection method described in the specific implementation method;

[0042] Figure 3 An original image of the production material described in the specific implementation method;

[0043] Figure 4 A filtered image obtained by the guided filtering process described in the specific implementation manner;

[0044] Figure 5 It is a binary image obtained after the binarization process described in the specific implementation method;

[0045] Figure 6 The optimized image obtained after the optimization process described in the specific implementation manner;

[0046] Figure 7 A distance image obtained after distance transformation as described in the specific implementation manner;

[0047] Figure 8 The image obtained after watershed transformation as described in the specific implementation manner;

[0048] Fig. 9 The mask image obtained after mask covering in a specific implementation manner;

[0049] Fig.10 It is a module block diagram of the online blast furnace feeding belt material particle size visual detection device described in the specific implementation method;

[0050] The reference numerals in the above drawings are described as follows:

[0051] 1. Feed conveyor belt; 2. Camera; 3. Fill light; 4. Gantry bracket; 21. Adjustment bracket; 100. Online blast furnace feeding belt material particle size visual detection device; 101. Image acquisition module; 102. Visual algorithm module. DETAILED DESCRIPTION

[0052] In order to explain in detail the possible application scenarios, technical principles, specific schemes that can be implemented, and the purposes and effects that can be achieved, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0053] Reference to "embodiment" herein means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The term "embodiment" appearing in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or association with other embodiments. In principle, in the present application, as long as there is no technical contradiction or conflict, the various technical features mentioned in the embodiments can be combined in any way to form a corresponding implementable technical solution.

[0054] Unless otherwise defined, the technical terms used in this document have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms in this document is only for describing specific embodiments and is not intended to limit this application.

[0055] In the description of this application, the term "and / or" is an expression used to describe the logical relationship between objects, indicating that three relationships may exist, for example, A and / or B, which means: A exists, B exists, and A and B exist at the same time. In addition, the character " / " in this article generally indicates that the objects before and after are in an "or" logical relationship.

[0056] In the present application, terms such as “first” and “second” are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship of quantity, priority or sequence between these entities or operations.

[0057] Without further limitations, in this application, the words "include", "comprises", "has" or other similar open-ended expressions used in sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product including the elements, so that the process, method or product including a series of elements may include not only those limited elements, but also other elements not explicitly listed, or also include elements inherent to such process, method or product.

[0058] Similar to the understanding in the Examination Guidelines, in this application, expressions such as "greater than", "less than", "exceed" and the like are understood to exclude the number itself; expressions such as "above", "below", "within" and the like are understood to include the number itself. In addition, in the description of the embodiments of this application, "multiple" means more than two (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups", "multiple times", etc., unless otherwise clearly and specifically limited.

[0059] In the description of the embodiments of the present application, space-related expressions used, such as "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "vertical", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicate the orientation or position relationship based on the orientation or position relationship shown in the specific embodiments or drawings, and are only for the convenience of describing the specific embodiments of the present application or facilitating the reader's understanding, and do not indicate or imply that the referred device or component must have a specific position, a specific orientation, or be constructed or operated in a specific orientation, and therefore cannot be understood as a limitation on the embodiments of the present application.

[0060] Unless otherwise expressly specified or limited, in the description of the embodiments of the present application, the terms such as "install", "connect", "connect", "fix", "set", etc. used should be understood in a broad sense. For example, the "connection" can be a fixed connection, a detachable connection, or an integrated setting; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. For technicians in the technical field to which the present application belongs, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0061] See also Figures 1 to 9 This embodiment provides an online blast furnace feeding belt material particle size visual detection method. The online blast furnace feeding belt material particle size visual detection method can be applied to the real-time detection of raw material particle size in blast furnace production of steel plants. Figure 1 As shown, the raw materials for the blast furnace of the steel plant are conveyed into the blast furnace by the blast furnace feed conveyor belt 1. The raw materials are melted at high temperature in the blast furnace and a series of chemical reactions occur, thereby obtaining the molten iron required for steel production. A gantry support 4 is arranged above the blast furnace feed conveyor belt 1, and a camera 2 is arranged on the gantry support 4. The camera 2 is installed on the crossbeam of the gantry support 4 by adjusting the support 21. The elevation angle of the camera 2 can be adjusted by adjusting the support 21. The image (i.e., the original image) of the raw materials on the blast furnace feed conveyor belt 1 can be obtained by the camera 2, and then the image is analyzed to obtain the particle size of the production material. In some embodiments, a fill light 3 can also be arranged on the gantry support 4 to increase the light intensity within the shooting range of the camera 2. Among them, the raw materials include sintered ore and coke; according to the production process requirements, the blast furnace feed conveyor belt 1 delivers sintered ore or coke into the blast furnace in different time periods. That is, in a certain time period, the blast furnace feed conveyor belt 1 only conveys sintered ore, and in another time period, the blast furnace feed conveyor belt 1 only conveys coke. Therefore, the online blast furnace feeding belt material particle size visual detection method in this embodiment can be used to detect the particle size of sintered ore and the particle size of coke. The online blast furnace feeding belt material particle size visual detection method in this embodiment can detect the particle size of blast furnace production raw materials such as sintered ore and coke in real time, and can overcome the interference of ambient light changes and holes between production raw materials, thereby improving the accuracy of particle size detection.

[0062] like Figure 2 As shown, in this embodiment, the online blast furnace feeding belt material particle size visual detection method includes the following steps:

[0063] S201, obtaining an original image of the production raw material on the blast furnace feed conveyor belt through a camera;

[0064] S202, performing guided filtering on the original image to remove noise in the original image to obtain a filtered image;

[0065] S203, performing binarization processing on the filtered image to obtain a binary image;

[0066] S204, optimizing the binary image by using a morphological processing method and an area filtering method to remove holes between the production materials to obtain an optimized image;

[0067] S205, performing distance transformation on the optimized image to obtain a distance image;

[0068] S206, obtaining points whose distance values ​​are greater than a preset value from the distance image, and performing a watershed transform on the filtered image with the points as starting seed points to obtain a watershed segmentation result;

[0069] S207, marking and area statistics are performed on the watershed segmentation result to obtain particle size data of the production raw material.

[0070] like Figure 2 As shown, in step S201, the camera 2 can shoot the production materials on the feed conveyor belt 1 from top to bottom. The camera 2 can be a high-definition industrial camera or a thermal imager. A thermal imager is a device that uses infrared thermal imaging technology to detect the infrared radiation of the target object and convert the image of the temperature distribution of the target object into a visible image by means of signal processing, photoelectric conversion, etc. Figure 3 As shown, an image of the production material on the feed conveyor belt 1 obtained by the camera 2, that is, an original image.

[0071] In step S202, the noise in the original image can be removed by guided filtering, and the edges of the raw materials are kept from being smoothed to the maximum extent while removing the noise, thereby ensuring the accuracy of subsequent particle size calculation. Figure 4 As shown, this is the filtered image obtained after guided filtering.

[0072] In some embodiments, before step S202, the region of interest of the visual algorithm can be set first, and then the original image obtained in step S201 is cut to leave only the image in the region of interest, wherein the region of interest of the visual algorithm can be a rectangle or a square. The guided filtering process of the original image in step S202, that is, the guided filtering process of the image in the region of interest, can prevent the conveyor belt or shadow at the edge of the original image from interfering with the subsequent analysis.

[0073] In step S203, the gaps between the raw materials in the filter image can be clearly distinguished by binarization processing, thereby highlighting the outline of the raw materials such as sintered ore or coke. Figure 5 As shown, it is a binary image obtained after binarization processing, in which the black part is the production raw material, and the white part is the gap between the production raw materials.

[0074] In step S204, the binary image is optimized by a morphological processing method and an area filtering method, wherein the morphological processing method includes an image morphological erosion operation and an image morphological dilation operation, which can be optimized by a morphological opening operation and a morphological closing operation of the image. In the morphological opening operation, the binary image is first eroded to eliminate isolated points whose median values ​​are higher than those of neighboring points in the binary image, so as to achieve the purpose of removing noise in the image, and to disconnect the narrow neck, and separate the two production materials at the thin connection point (i.e., the position where the two production materials are attached together), so as to ensure the independence of the production material contour; and then the image after the opening operation is expanded to compensate for the area reduction caused by erosion in the larger connected domain. Therefore, in this embodiment, the noise in the binary image can be removed by the morphological processing method without reducing the area of ​​the main area in the image; therefore, the voids between the production materials are filled without increasing the area occupied by the noise.

[0075] The area filter can filter out images outside the selected threshold range, that is, filter out connected areas outside the selected threshold range according to the area of ​​the connected domain in the binary image, thereby suppressing and preventing interference from connected areas outside the selected threshold range.

[0076] Preferably, when the binary image is optimized by using the morphological processing method and the area filtering method, the binary image is first optimized by using the morphological processing method to remove the first holes between the production materials that are smaller than the first aperture; then the area filtering is performed to remove the second holes between the production materials that are larger than the first aperture and smaller than the second aperture, wherein the second aperture is larger than the first aperture. In this embodiment, the larger holes between the production materials can be filtered out by the area filtering. The smaller holes between the production materials can be filtered out by the morphological processing method. Figure 6 As shown, it is an optimized image obtained after the optimization process in step S204.

[0077] In step S205, the distance between each pixel and the pixel in the image that meets a certain condition can be calculated by distance transformation, and then the grayscale value is transformed using the calculated distance. The distance metrics used include: Euclidean distance, chessboard distance, block distance (Manhattan distance), etc. In this embodiment, the block distance is preferably used. In this embodiment, the distance transformation can be used to calculate the distance value from each pixel in the image to the background pixel closest to it, and the distance value is applied to the watershed transformation in step S206. Figure 7 As shown, this is the distance image obtained after distance transformation.

[0078] In step S206, during the watershed transformation, the point with the larger distance value in step S205 is used as the starting seed point of the segmentation algorithm in the watershed transformation, and the image after the guide filter (i.e. Figure 4 Filtered image) to perform watershed transform to obtain the result of watershed segmentation. Figure 8 As shown, this is the image obtained after watershed transformation.

[0079] In step S207, the area of ​​each closed region in the watershed segmentation result is counted, and the area of ​​the closed region is converted into the granularity of the production material, wherein the larger the area of ​​the closed region is, the larger the granularity of the corresponding production material is.

[0080] Taking into account the uneven surface illumination of sintered ore and coke (i.e., the surface illumination of a portion of the raw materials in the original image is higher, while the surface illumination of another portion of the raw materials is lower), in order to reduce the impact of uneven surface illumination, in one embodiment, the binarization processing of the above step S203 adopts adaptive threshold binarization processing. The traditional binarization threshold is fixed, and is usually determined based on the average brightness of the entire image or the brightness of the center position. Therefore, its processing effect is not ideal for images with uneven illumination and complex background. In this embodiment, an adaptive threshold binarization algorithm is used to adaptively determine the binarization threshold based on the local features of the image. It takes into account the illumination changes in the image and has stronger robustness to the illumination changes in the image. At the same time, with the help of the integral map, the calculation of the adaptive threshold is accelerated. Among them, the adaptive threshold is obtained by the following calculation formula:

[0081] ;

[0082] Where T is the adaptive threshold, For is the side length of the visual algorithm area of ​​interest of the square with the center point, t is the set threshold parameter, is the horizontal coordinate of the point, is the ordinate of the point, The pixel of the point. For each point and point in the area The horizontal coordinate difference of / 2, / 2]; For each point and point in the area The vertical coordinate difference of / 2, / 2].

[0083] In the above embodiment, the particle size of the raw materials on the blast furnace feed conveyor belt is analyzed and detected in real time, and the detection process does not require manual intervention, which effectively solves the problems of lag, small detection volume and inaccurate data in manual sampling detection. In addition, the above technical solution can remove the noise in the original image by performing guided filtering on the original image, and can avoid the edge of the raw material from being smoothed; and the above solution can remove the holes between the raw materials by binarizing the filtered image, and optimizing the processing through morphological processing methods and area filtering, so as to obtain a clear outline of the raw material, thereby improving the accuracy of subsequent watershed segmentation, thereby improving the accuracy of raw material particle size detection.

[0084] In one embodiment, step S207 marks and counts the area of ​​the watershed segmentation result to obtain the particle size data of the production raw material, including:

[0085] Obtaining each closed region in the watershed segmentation result, and calculating the area of ​​each closed region;

[0086] Each of the closed regions is marked according to the area, and the area is converted into a particle size of the production material.

[0087] Among them, each closed area in the watershed segmentation result corresponds to a production material, so the particle size of the corresponding production material can be obtained by calculating the area of ​​each closed area. When marking the closed area, several area intervals or particle size intervals can be set in advance. For example, the preset particle size intervals include: particle size interval 1: 1-2cm, particle size interval 2: 2-3cm, particle size interval 3: 3-4cm, and then the corresponding particle size interval is marked according to the calculated particle size of the production material. The particle size distribution of the current production material can be intuitively obtained through the above marking. In this embodiment, the proportion of the production material in each of the particle size intervals can be further calculated.

[0088] In the above step S206, the watershed segmentation result includes not only closed areas but also some unclosed areas. After analysis, it is found that some of these unclosed areas are also production raw materials. Therefore, in order to improve the accuracy of particle size analysis, in one embodiment, the acquisition of each closed area in the watershed segmentation result includes:

[0089] Removing the first unclosed area located at the boundary in the watershed segmentation result, and obtaining each first closed area whose contour is completely closed and each second unclosed area which is not completely closed in the watershed segmentation result;

[0090] Determine whether the length ratio of the unclosed portion in each of the second unclosed regions is less than a preset value, and if so, fit the second unclosed regions into completely closed second closed regions;

[0091] The first closed area and the second closed area are obtained.

[0092] After analysis, it was found that some of these unclosed areas are located in the area of ​​interest of the visual algorithm or the edge of the original image, so the production materials that are not fully identified or photographed, and some of them are due to the special shape of the surface of the production material, which causes it to fail to obtain a closed area after the watershed transformation. Therefore, in this embodiment, the first unclosed area located at the boundary in the watershed segmentation result is first removed, and then by judging whether the length ratio of the unclosed part in the unclosed area is less than the preset value, the above-mentioned production materials that cannot form a closed area can be identified, and then this part of the second unclosed area that fails to form a closed area is fitted into a closed second closed area, so as to be added to the granularity statistics. In this embodiment, the preset value can be set according to the type, shape and approximate granularity of the production material. Therefore, in this embodiment, the shortcomings of the existing watershed transformation can be overcome, and the production materials of the watershed segmentation result can be more accurately identified, thereby improving the accuracy of granularity statistics.

[0093] In one embodiment, before dividing the multiple particle size intervals according to the result of the area statistics, the method further includes the step of identifying the type of the production raw material according to the original image, and the type of the production raw material includes sintered ore or coke.

[0094] The method of dividing the area statistics into a plurality of particle size intervals and calculating the proportion of the production raw materials in each particle size interval comprises: setting different particle size intervals according to different types of production raw materials; and calculating the proportion of different types of production raw materials in each corresponding particle size interval.

[0095] In this embodiment, image features are further extracted from the original image provided in step S201, and then an artificial intelligence algorithm is used to identify whether the currently transmitted production raw material is sintered ore or coke. The system sets different particle size intervals according to different production raw material types, and outputs the particle size interval ratios of different raw material types in the adaptation step S207.

[0096] In one embodiment, a mask is generated according to the result of the area statistics, and the mask is covered on the surface of the original image or the surface of the filtered image to obtain a mask image, in which at least two colors are used to mark the production materials of different particle sizes. Fig. 9As shown in the figure, the mask image is obtained after the mask is covered. The mask is a colored layer with a certain degree of perspective. When the mask is covered on the surface of the original image or the surface of the filtered image, not only can the surface of the original image or the production materials in the filtered image be seen, but also a certain color will be covered on the production materials, such as Fig. 9 As shown, the colors of different positions in the mask are determined according to the particle sizes of the production materials corresponding to the positions, so that production materials of different particle sizes are covered with different colors.

[0097] like Fig.10 As shown, in another embodiment, an online blast furnace feeding belt material particle size visual detection device 100 is provided. The online blast furnace feeding belt material particle size visual detection device 100 includes: an image acquisition module 101 and a visual algorithm module 102.

[0098] The image acquisition module 101 is disposed above the blast furnace feed conveyor belt 1 and is used to obtain the original image of the production material on the blast furnace feed conveyor belt.

[0099] The visual algorithm module 102 is connected to the image acquisition module 101, and is used to perform guided filtering processing on the original image to remove noise from the image and obtain a filtered image; binarize the filtered image to obtain a binary image including a foreground image and a background image, wherein the production raw materials are located in the foreground image; optimize the binary image using a morphological processing method and an area filtering method to remove holes between the production raw materials and obtain an optimized image; perform distance transformation on the optimized image to obtain a distance image; obtain points with distance values ​​greater than a preset value from the distance image, and perform watershed transformation on the filtered image with the points as starting seed points to obtain a watershed segmentation result; mark and perform area statistics on the watershed segmentation result to obtain the particle size data of the production raw materials.

[0100] like Figure 1 As shown, the image acquisition module includes a camera and a fill light 3, and the camera and the fill light are installed on a gantry bracket 4, and the gantry bracket 4 spans above the blast furnace feed conveyor belt 1. In this embodiment, the camera can be a high-definition industrial camera, and the gantry bracket 4 spans the belt to avoid the vibration caused by the blast furnace feed conveyor belt 1; and the fill light is set on both sides of the gantry bracket 4, which can improve the quality of the original image obtained and highlight the surface characteristics of the production raw materials.

[0101] The online blast furnace feeding belt material particle size visual detection device 100 of this embodiment can analyze and detect the particle size of the production materials on the blast furnace feeding conveyor belt in real time. The detection process does not require manual intervention, which effectively solves the problems of lag and inaccurate data in manual sampling detection. In addition, the above technical solution removes noise in the original image by guiding filtering the original image, and can avoid smoothing the edges of the production materials; and the above solution removes holes between the production materials by binarizing the filtered image, and optimizes the processing through morphological processing methods and area filtering, so as to obtain a clear outline of the production materials, thereby improving the accuracy of subsequent watershed segmentation, thereby improving the accuracy of the production raw material particle size detection.

[0102] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of this application, this does not limit the scope of patent protection of this application. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concept of this application using the contents recorded in the specification and drawings of this application, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the scope of patent protection of this application.

Claims

1. An online blast furnace feeding belt material particle size visual detection method, characterized in that: The following steps are involved: The original image of the production material on the blast furnace feed conveyor belt is obtained by the camera; Performing guided filtering on the original image to remove noise in the original image to obtain a filtered image; Binarizing the filtered image to obtain a binary image; The binary image is optimized by using a morphological processing method and an area filtering method to remove holes between the production materials and obtain an optimized image; Performing distance transformation on the optimized image to obtain a distance image; Acquire points whose distance values ​​are greater than a preset value from the distance image, and perform watershed transformation on the filtered image with the points as starting seed points to obtain a watershed segmentation result; Marking and area counting the watershed segmentation results to obtain particle size data of the production raw material; The watershed segmentation result is marked and the area is counted to obtain the particle size data of the production raw material, including: Obtaining each closed region in the watershed segmentation result, and calculating the area of ​​each closed region; Marking each of the closed areas according to the area, and converting the area into the particle size of the production material; The obtaining of each closed area in the watershed segmentation result comprises: Removing the first unclosed area located at the boundary in the watershed segmentation result, and obtaining each first closed area whose contour is completely closed and each second unclosed area which is not completely closed in the watershed segmentation result; a part of the unclosed area is production raw material; Determine whether the length ratio of the unclosed portion in each of the second unclosed regions is less than a preset value, and if so, fit the second unclosed regions into completely closed second closed regions; The first closed area and the second closed area are obtained.

2. The online blast furnace feeding belt material particle size visual detection method according to claim 1 is characterized in that: The method of optimizing the binary image by using a morphological processing method and an area filtering method to remove holes between the production materials includes: Firstly, the binary image is optimized by using a morphological processing method to remove first holes between the production materials that are smaller than the first hole diameter; Then, area filtering is performed to remove second holes between the production materials that are larger than the first aperture and smaller than the second aperture, wherein the second aperture is larger than the first aperture.

3. The online blast furnace feeding belt material particle size visual detection method according to claim 1 is characterized in that: The binarization process is an adaptive threshold binarization process.

4. The online blast furnace feeding belt material particle size visual detection method according to claim 1 is characterized in that: Also includes the steps: According to the result of the area statistics, a plurality of particle size intervals are divided, and the proportion of the production raw material in each of the particle size intervals is calculated.

5. The online blast furnace feeding belt material particle size visual detection method according to claim 4 is characterized in that: Before dividing the plurality of particle size intervals according to the area statistics result, the method further includes the following steps: Identify the type of the production raw material according to the original image, where the type of the production raw material includes sintered ore or coke; The step of dividing a plurality of particle size intervals according to the result of the area statistics and calculating the proportion of the production raw material in each particle size interval comprises: Different particle size intervals are set according to different types of production raw materials; and the proportion of different types of production raw materials in each corresponding particle size interval is calculated.

6. The online blast furnace feeding belt material particle size visual detection method according to claim 5 is characterized in that: The step of dividing a plurality of particle size intervals according to the result of the area statistics and calculating the proportion of the production raw material in each particle size interval further includes: A mask is generated according to the result of the area statistics, and the mask is covered on the surface of the original image or the surface of the filtered image to obtain a mask image, wherein at least two colors are used in the mask to mark the production materials of different particle sizes.

7. An online blast furnace feeding belt material particle size visual detection device, characterized in that: include: An image acquisition module is arranged above the blast furnace feed conveyor belt and is used to obtain the original image of the production raw materials on the blast furnace feed conveyor belt; A visual algorithm module, connected to the image acquisition module, for performing guided filtering on the original image to remove noise from the original image and obtain a filtered image; Binarizing the filtered image to obtain a binary image including a foreground image and a background image, wherein the production raw material is located in the foreground image; optimizing the binary image by using a morphological processing method and an area filtering method to remove holes between the production raw materials to obtain an optimized image; Performing distance transformation on the optimized image to obtain a distance image; obtaining points whose distance values ​​are greater than a preset value from the distance image, and performing watershed transformation on the filtered image with the points as starting seed points to obtain a watershed segmentation result; Marking and area counting the watershed segmentation results to obtain particle size data of the production raw material; The watershed segmentation result is marked and the area is counted to obtain the particle size data of the production raw material, including: Obtaining each closed region in the watershed segmentation result, and calculating the area of ​​each closed region; Marking each of the closed areas according to the area, and converting the area into the particle size of the production material; The obtaining of each closed area in the watershed segmentation result comprises: Removing the first unclosed area located at the boundary in the watershed segmentation result, and obtaining each first closed area whose contour is completely closed and each second unclosed area which is not completely closed in the watershed segmentation result; a part of the unclosed area is production raw material; Determine whether the length ratio of the unclosed portion in each of the second unclosed regions is less than a preset value, and if so, fit the second unclosed regions into completely closed second closed regions; The first closed area and the second closed area are obtained.

8. The online blast furnace feeding belt material particle size visual detection device according to claim 7 is characterized in that: The image acquisition module comprises a camera and a fill light, wherein the camera and the fill light are mounted on a gantry support, and the gantry support spans above the blast furnace feed conveyor belt.

Images