A sintering machine tail state monitoring method based on machine vision

Abstract

The application discloses a sintering machine tail state monitoring method based on machine vision. The method comprises the following steps: S1, preprocessing of the input sintering machine tail image; S2, color gamut conversion of the original image to extract the combustion area in the image; S3, image processing after extracting the combustion area and extraction of the machine tail area; S4, feature value calculation of the extracted machine tail area; and S5, sintering state judgment and monitoring according to the feature value. The application constructs a sintering machine tail state monitoring system through machine vision and image processing related technologies, realizes real-time monitoring of the sintering machine, gives a prompt and a warning to the abnormal part, further realizes intelligent production of the sintering process section, improves the production efficiency of the steel enterprise, reduces the labor cost, and further promotes the intelligent manufacturing transformation and upgrading and intelligent development of the steel metallurgy field in China.

Description

Technical Field

[0001] This invention relates to the fields of intelligent manufacturing in metallurgy, machine vision, and image processing; more specifically, it relates to a method for monitoring the condition of the tail end of a sintering machine based on machine vision. Background Technology

[0002] The sintering process typically involves a complex environment, often characterized by high temperatures and high dust levels. The ignition temperature can reach over 1000 degrees Celsius, with the final temperature around 400 degrees Celsius. Therefore, effective and scientific control of the sintering process is a major research focus for sintering professionals. Currently, most sintering workers in China rely primarily on human experience and skills to control the process. This involves the fire watcher observing the cross-section of the sintered cake at the tail of the sintering machine and making judgments based on their experience. However, the environment at the tail of the sintering machine is generally harsh, and the experience of different fire watchers varies. Therefore, the harsh environment and differences in human experience contribute to the difficulty in controlling the sintering process. Anomalies can significantly impact the quality of sintered ore and the efficiency of blast furnace smelting. Therefore, it is necessary to develop a machine vision-based method for monitoring the condition of the sintering machine tail. Through image processing and combined with machine learning and other related algorithms, images of the sintering machine tail can be acquired and analyzed in real time. The thickness of the red-hot layer, FeO content, and area of ​​the black ore layer in the images can be detected in real time. The detection results and early warnings can be fed back to the operators on-site or in the central control room in real time, thereby realizing real-time monitoring of the sintering machine tail. This further enables intelligent production in the sintering process, improves the production efficiency of steel enterprises, and reduces labor costs and production or safety problems caused by human factors. Summary of the Invention

[0003] In view of the defects and shortcomings of the existing technology, the present invention provides a method for steel profile attitude discrimination based on Hough transform.

[0004] To achieve the above objectives, the technical solution provided by this invention is as follows:

[0005] The present invention provides a machine vision-based method for monitoring the tail condition of a sintering machine, comprising the following steps:

[0006] S1: Preprocess the input sintering machine tail image;

[0007] S2: Perform color gamut conversion on the original image to extract the burning area in the image;

[0008] S3: Process the image after extracting the combustion area and extract the tail area;

[0009] S4: Calculate the feature values ​​of the extracted tail region;

[0010] S5: Judge and monitor the sintering state based on characteristic values.

[0011] Furthermore, step S1 specifically includes:

[0012] S11: Preprocess the input sintering machine tail image. First, perform image denoising by performing median filtering on the image containing noise points to suppress noise while preserving the original details of the image.

[0013] S12: Perform histogram equalization on the image after median filtering. This process enhances the contrast between the target and the background in images with poor contrast and blurred key areas.

[0014] S13: After completing the median filtering and histogram equalization of the image in steps S11 and S12, the image is Gaussian blurred to facilitate color extraction in subsequent color gamut conversion.

[0015] Furthermore, step S3 specifically includes:

[0016] S31: Binarize the image after extracting the combustion area to obtain a binarized image.

[0017] S32: Fill the holes in the combustion area in the image to obtain the cross-sectional image of the sintering machine tail section after filling.

[0018] S33: Extract the contours of the filled combustion areas and calculate the corresponding center coordinates and area.

[0019] S34: Sort the regions according to their area, and take the center coordinates of the uppermost region of the regions with the largest area as the center coordinates of the tail region of the sintering machine. Extract the tail region from the original image according to these coordinates for subsequent feature value calculation.

[0020] Furthermore, the median filtering operation uses a suitable two-dimensional sliding template to traverse each pixel of the image from left to right and from top to bottom, selects the pixel value of each pixel and its surrounding neighboring pixels, sorts these pixel values, and takes the pixel value located in the middle position as the pixel value of the current pixel, thereby eliminating isolated noise points.

[0021] Furthermore, the histogram equalization process first scans each pixel in the original image and calculates the histogram of pixel values ​​in different channels; second, it calculates the cumulative distribution function of the histogram for each channel; then, based on the cumulative distribution function and the histogram equalization principle, it obtains the mapping relationship between the input and output; finally, it performs image transformation based on the mapping relationship to enhance the contrast between the target object and the background in the image. The mapping method is as follows:

[0022]

[0023] Where S k The value of the current channel pixel is the result of the cumulative distribution function mapping, where n is the sum of the pixels in the image. j L represents the number of pixel levels in the current channel, where L is the total number of pixel levels in a single channel of the image.

[0024] Furthermore, the Gaussian blur, also called Gaussian smoothing, essentially replaces the value of the center pixel with the weighted average of the surrounding pixels. The weights are distributed according to a normal distribution; the closer to the center, the greater the weight, and the further away from the center, the smaller the weight. The density function of the two-dimensional normal distribution of the image is:

[0025]

[0026] Assuming the center pixel coordinates of the image are (0,0), when the value of σ is 1.5, the weight assignment values ​​of this normal distribution to its surrounding 8 pixels are as follows. In order to make the sum of its weights equal to 1, the weight matrix needs to be normalized, resulting in the Gaussian blur template as follows:

[0027]

[0028] After normalizing the weight matrix to obtain the Gaussian blur template, it is applied to the sintering machine tail image. Each pixel is traversed, and each pixel is multiplied by its corresponding weight value and summed with the weight value of its eight neighboring pixels. The resulting value replaces the original pixel value. After performing the above operation on all pixels, the Gaussian blur of the image is completed.

[0029] Furthermore, the aforementioned color gamut conversion and burning region extraction involves converting the original RGB (red, green, blue) image into an HSV (chroma, saturation, luminance) image, and then using the HSV luminance channel values ​​to distinguish burning and non-burning regions in the image. The RGB to HSV conversion formula is as follows:

[0030]

[0031] Furthermore, the hole filling process involves filling small, missing holes within a region using a suitable algorithm to create a complete area. Essentially, a hole is a background region enclosed by a boundary connecting foreground pixels, which can typically be filled using a constrained dilation method. This involves constructing a structuring element, finding a point within the hole, dilating it using this structuring element, and then constraining it using the complement of the original image (intersection operation). This process is iterated and repeated until the algorithm converges, resulting in the hole-filled image.

[0032] Furthermore, the calculation of the region's center coordinates and area generally uses the zeroth and first moments of the image to calculate the region's area and its center coordinates. The formula for calculating the zeroth moment of the image is as follows:

[0033]

[0034] Where V(i,j) represents the pixel value at point (i,j) in the image. When the image is a binary image, this value is 1, i.e., M. 00 It is the sum of white pixels in the image, so it can be used to calculate the area of ​​a region.

[0035] The formula for calculating the first moment of an image is as follows:

[0036]

[0037] When the image is a binary image, the value of V(i,j) is 1, and M... 10 It involves summing the x-axis coordinates of the white regions in the image; therefore, the first moment can be used to find the center coordinates of a binary image.

[0038]

[0039] Furthermore, the feature value calculation of the tail area includes the thickness of the red-fire layer, temperature statistics, black frame layer area, and pore area. For the calculation of the red-fire layer thickness, the sintering tail area is divided into six parts from left to right along the horizontal axis. The brightness values ​​of pixels in the HSV three channels of each column within each part are counted within the range [220, 250]. The average value of all columns in that part is then taken to obtain the average thickness of the red-fire layer for that part. For the temperature statistics, the calculation is performed in two directions: horizontally and vertically. Horizontally, the average brightness value of all pixels in each column is used as the temperature index for that column. Vertically, the average brightness value of all pixels in each row is used as the temperature value for that row. The indicators are as follows: For the black frame area, it is calculated in both horizontal and vertical directions. In the horizontal direction, all pixels with a brightness value between [10, 100] in each column are counted as the black frame area of ​​that column, reflecting the uniformity of the fabric in the horizontal direction. In the vertical direction, all pixels with a brightness value between [10, 100] in each row are counted as the black frame area of ​​that row, reflecting the uniformity of the fabric in the vertical direction. Finally, for the pore area, all pixels with a brightness value greater than 250 in the tail area are counted. The area of ​​these pixels that meet the requirements is divided by the number of regions formed by the pixels to obtain the unit pore area of ​​the tail area, which reflects the content of ferrous oxide (FeO) during the sintering process.

[0040] This invention provides a machine vision-based method for monitoring the condition of the sintering machine tail section. Through image processing and machine learning, it filters and denoises the cross-sectional image of the sintering machine tail section, performs histogram equalization to enhance the contrast between the target and background, then applies Gaussian blur and color gamut conversion to transform the original image into an HSV image. Next, it extracts the tail section region through brightness differences, regional coordinates, and area calculations. Finally, it calculates feature values ​​such as the thickness of the red-hot layer, the area of ​​pores, and the area of ​​the black frame layer in the tail section region. These calculated feature values ​​are then output in text and chart form within a set parameter range and fed back to the operator. The establishment of this sintering machine tail section condition monitoring system enables intelligent identification and real-time monitoring of the sintering cake cross-section, effectively guiding improvements in process operation, reducing ironmaking costs, labor costs, and safety or production problems caused by human factors, thereby increasing the production efficiency of steel enterprises. This project, through visual recognition methods, further realizes intelligent control and scheduling of the sintering process, promoting the intelligent manufacturing transformation and upgrading and intelligent development of my country's steel metallurgy industry. Attached Figure Description

[0041] Figure 1 The flowchart of this invention

[0042] Figure 2 This is an extraction diagram of the tail area of ​​the sintering machine in this invention.

[0043] Figure 3 This is an extraction diagram of the red flame layer in the tail area of ​​the present invention.

[0044] Figure 4 This is an extraction diagram of the exhaust gas vent area of ​​the sintering machine in this invention.

[0045] Figure 5 This is a chart showing the condition detection of the sintering machine tail section in this invention.

[0046] Figure 6 This is a front-end display of the sintering machine tail condition monitoring software in this invention. Detailed Implementation

[0047] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0048] In the description of this invention, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0049] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0050] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0051] Example 1

[0052] This embodiment presents a machine vision-based method for monitoring the state of the sintering machine tail section, including median filtering and noise reduction of the sintering machine tail section image, histogram equalization, Gaussian blurring, HSV color gamut conversion, calculation of region center coordinates and area, calculation of feature values ​​of the tail section region, and detection of the sintering state. The steps are as follows:

[0053] Step 1: For the input sintering machine tail section image, first perform median filtering. Use a 3*3 template to traverse each pixel from left to right and from top to bottom, select the pixel value of each pixel and its surrounding neighboring pixels, sort these pixel values, and replace the pixel value of the current pixel with the pixel value in the middle position. After performing the above operation on all pixels, the median filtering of the image is completed.

[0054] Step 2: Perform histogram equalization on the median-filtered image. For the input image's R (red), G (green), and B (blue) channels, count the pixel values ​​of all pixels in each channel and classify them. Count the number of pixels at each pixel level to form a histogram. Then, sum the histograms from each channel to calculate the cumulative distribution function for each channel. Interpolate the cumulative distribution function to obtain new pixel values. Replace the original pixel values ​​with the new pixel values ​​for each channel to obtain the result of histogram equalization of the color image. Histogram equalization compresses the small percentage of pixels in the image, making the histogram distribution more uniform and expanding the dynamic range of pixel value differences, thereby enhancing the overall contrast of the image.

[0055] Step 3: Apply Gaussian blur to the histogram-equalized image to facilitate color extraction in the subsequent HSV color gamut conversion. Specifically, a Gaussian smoothing template is used to iterate through each pixel, replacing the value of the center pixel with the weighted average of the surrounding pixels. The Gaussian smoothing template used in this example is as follows:

[0056]

[0057] Gaussian smoothing template

[0058] Step 4: After Gaussian blurring the image of the sintering machine tail, convert it to the HSV color gamut space. That is, use the color gamut transformation formula to convert the R, G, B three-channel image to the HSV three-channel image. After the conversion, extract all pixels in the image whose brightness channel value is greater than 200 to obtain the combustion area of ​​the sintering machine tail section.

[0059] Step 5: After obtaining the combustion area of ​​the sintering machine tail section from Step 4, the sintering machine tail area used for feature value calculation in this example is above the sintering section. Therefore, the combustion area is first filled with holes. In this example, the structural element selected is a 3*3 cross structure. The hole area is expanded using this structural element, and then constrained (intersection operation) using the complement of the original image. The image after hole filling is obtained through continuous iteration.

[0060] Step Six: After filling the holes, the contours of the filled combustion area are extracted. The area and center coordinates of all contours are calculated. Then, the areas are sorted, and the contour with the highest center ordinate among the top three is selected. This contour is the contour of the sintering machine tail area. The coordinate position information of this contour is then extracted from the HSV image obtained in Step Four to obtain the sintering machine tail area used for feature value calculation, as shown below. Figure 2 As shown.

[0061] Step 7: Next, feature value calculation is performed on the extracted tail region. First, the pixels with brightness channel values ​​in the range [220, 250] in the tail region image are extracted to obtain the red flame layer image of the tail region, as shown below. Figure 3 As shown, the image is then divided into 6 parts from left to right along the horizontal axis. The number of red flame layer pixels in each part is counted, and the average value of all columns in that part is taken to obtain the average thickness of the red flame layer in each part. The thickness distribution of the red flame layer in this example is shown in the image below. Figure 5 As shown.

[0062] Step 8: Next, calculate the average brightness value of all pixels in each column horizontally, using this as the temperature value index for that column. Similarly, calculate the average brightness value of all pixels in each row vertically, using this as the temperature value index for that row. In this example, the temperature distribution in the horizontal and vertical directions is as follows: Figure 5 As shown.

[0063] Step Nine: Similarly, in both the horizontal and vertical directions, for the horizontal direction, count all pixels in each column with a brightness value between [10, 100], and use this count as the area of ​​the black border layer for that column. This reflects the fabric uniformity in the horizontal direction. For the vertical direction, count all pixels in each row with a brightness value between [10, 100], and use this count as the area of ​​the black border layer for that row. This reflects the fabric uniformity in the vertical direction. In this example, the fabric uniformity in the horizontal and vertical directions is as follows: Figure 5 As shown.

[0064] Step 10: Finally, count all pixels in the tail area with a brightness value greater than 250 to obtain the image of the tail area vent region, as shown below. Figure 4 As shown, the area of ​​these qualified pixels is divided by the number of regions formed by the pixels to obtain the unit pore area of ​​the tail region, which is used to reflect the content of ferrous oxide (FeO) during sintering. In this example, the ferrous oxide (FeO) content is as follows: Figure 5 As shown in the example, the text and chart content of the feature values ​​are displayed together on the front-end page of the sintering machine tail status software, such as... Figure 6 As shown.

[0065] This invention provides a machine vision-based method for monitoring the condition of the sintering machine tail section. Through image processing and machine learning, it filters and denoises the cross-sectional image of the sintering machine tail section, performs histogram equalization to enhance the contrast between the target and background, then applies Gaussian blur and color gamut conversion to transform the original image into an HSV image. Next, it extracts the tail section region through brightness differences, regional coordinates, and area calculations. Finally, it calculates feature values ​​such as the thickness of the red-hot layer, the area of ​​pores, and the area of ​​the black frame layer in the tail section region. These calculated feature values ​​are then output in text and chart form within a set parameter range and fed back to the operator. The establishment of this sintering machine tail section condition monitoring system enables intelligent identification and real-time monitoring of the sintering cake cross-section, effectively guiding improvements in process operation, reducing ironmaking costs, labor costs, and safety or production problems caused by human factors, thereby increasing the production efficiency of steel enterprises. This project, through visual recognition methods, further realizes intelligent control and scheduling of the sintering process, promoting the intelligent manufacturing transformation and upgrading and intelligent development of my country's steel metallurgy industry.

[0066] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0067] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for monitoring the condition of the sintering machine tail section based on machine vision, characterized in that, The method includes the following steps: S1: Preprocess the input sintering machine tail image; S2: Perform color gamut conversion on the original image to extract the burning area in the image; S3: Process the image after extracting the combustion area and extract the tail region: Calculate the area and center coordinates of all contours, then sort the areas, select the contour with the highest center ordinate among the top three, which is the area contour of the sintering tail. Then extract the coordinate position information of the contour from the HSV image obtained in step four to obtain the sintering tail region for feature value calculation. S4: Calculate the feature values ​​of the extracted tail region; the feature value calculation of the tail region includes calculating the thickness of the red flame layer, temperature statistics, area of ​​the black frame layer, and / or pore area of ​​the calculated region, wherein, To calculate the thickness of the red-fired layer, the tail area of ​​the sintering machine is divided into 6 parts from left to right along the horizontal axis. The number of pixels with brightness values ​​between [220, 250] in the HSV three channels in all columns of each part is counted. Then, the average value of all columns in that part is taken to obtain the average thickness of the red-fired layer in that part. For temperature value statistics, the calculation is mainly based on two directions: horizontal and vertical. In the horizontal direction, the average brightness value of all pixels in each column is calculated as the temperature value index of that column. In the vertical direction, the average brightness value of all pixels in each row is calculated as the temperature value index of that row. For the black frame area, it is also calculated in both horizontal and vertical directions. In the horizontal direction, all pixels with a brightness value of [10, 100] in each column are counted as the black frame area of ​​that column, which is used to reflect the uniformity of the fabric in the horizontal direction. In the vertical direction, all pixels with a brightness value of [10, 100] in each row are counted as the black frame area of ​​that row, which is used to reflect the uniformity of the fabric in the vertical direction. For the pore area, count all pixels with a brightness value greater than 250 in the tail area, divide the area of ​​these pixels that meet the requirements by the number of regions formed by the pixels, and obtain the unit pore area of ​​the tail area, which is used to reflect the content index of ferrous oxide (FeO) during the sintering process. S5: Judge and monitor the sintering state based on characteristic values.

2. The sintering machine tail condition monitoring method based on machine vision as described in claim 1, characterized in that, The specific steps of S1 are as follows: S11: Denoise the input sintering machine tail image, perform median filtering on the image containing noise points to suppress noise while preserving the original details of the image; S12: Perform histogram equalization on the image after median filtering to enhance the contrast between the target and the background; S13: After completing the median filtering and histogram equalization of the image in steps S11 and S12, the image is subjected to Gaussian blur.

3. The sintering machine tail condition monitoring method based on machine vision as described in claim 1, characterized in that, The specific steps of step S3 are as follows: S31: Binarize the image after extracting the combustion area to obtain a binary image; S32: Fill the holes in the combustion area in the image to obtain the cross-sectional image of the sintering machine tail section after filling; S33: Extract the contours of the filled combustion areas and calculate the corresponding center coordinates and area; S34: Sort the regions according to their area, and take the center coordinates of the uppermost region of the regions with the largest area as the center coordinates of the tail region of the sintering machine. Extract the tail region from the original image according to these coordinates for subsequent feature value calculation.

4. The sintering machine tail condition monitoring method based on machine vision as described in claim 2, characterized in that: The median filtering operation involves selecting a suitable two-dimensional sliding template, traversing each pixel of the image from left to right and from top to bottom, selecting the pixel value of each pixel and its surrounding neighboring pixels, sorting these pixel values, and taking the pixel value located in the middle position as the pixel value of the current pixel to eliminate isolated noise points.

5. The sintering machine tail condition monitoring method based on machine vision as described in claim 2, characterized in that: The histogram equalization steps include: Scan each pixel in the original image and calculate a histogram of the image's pixel values ​​in different channels; Calculate the cumulative distribution function of the histogram for each channel; The mapping relationship between input and output is obtained based on the cumulative distribution function and histogram equalization principle; finally, image transformation is performed based on the mapping relationship to enhance the contrast between the target object and the background in the image.

Images