Method for detecting tilting angle of blast furnace distribution chute on line

Abstract

The invention discloses a method for detecting a tilting angle of a blast furnace distribution chute on line. The method comprises the following steps: S1, mounting an infrared thermal imager and collecting a calibration sequence; s2, real-time image acquisition; s3, image preprocessing: denoising the acquired real-time thermal image data, and enhancing preprocessing operation; s4, a control signal is accessed, wherein control information in a PLC of a control system is integrated, the rotating direction of the chute is judged, and the detection time is set; s5, chute edge feature extraction: extracting key features including chute edges for a single-frame image, and judging whether a chute edge contour is detected at the current time or not by combining a time sequence and a sliding window algorithm contrast calibration sequence; s6, angle calculation: calculating a chute tilting angle based on chute edge contour sequence data judged by feature extraction, and feeding back the chute tilting angle to a control center; and S7, real-time monitoring and feedback: repeating the steps S2-S6 to complete real-time monitoring of the angle change and feed the angle change back to the control system PLC.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence detection, specifically to a method for online detection of the tilt angle of a blast furnace charging chute. Background Technology

[0002] In the blast furnace ironmaking process, the tilting angle of the charging chute directly affects the material distribution and gas flow within the furnace, and is one of the key parameters for blast furnace operation. Traditional methods for detecting the tilting angle of the charging chute have the following main drawbacks:

[0003] Contact measurement method: The mechanical sensor directly measures the data. However, the equipment is susceptible to harsh environments such as high temperature, dust, and vibration, resulting in a short service life and high maintenance costs.

[0004] Traditional visible light visual measurement: Measurement is performed using a visible light camera. However, in the blast furnace environment, the lighting conditions are poor and there is a lot of dust. The image quality is poor when the blast furnace is in operation, so measurements can only be taken when the blast furnace is not in operation, and the accuracy is difficult to guarantee.

[0005] In summary, most current measurement methods require shutdown and cannot achieve real-time online monitoring, thus affecting blast furnace production efficiency. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method for online detection of the tilt angle of the blast furnace charging chute.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for online detection of the tilt angle of a blast furnace charging chute, comprising the following steps:

[0008] S1. Infrared thermal imager installation and calibration sequence acquisition: Install an infrared thermal imager at a suitable position on the top of the blast furnace to ensure complete imaging of the charging chute area; after installing the infrared thermal imager at the target position, rotate the chute in 5-degree increments to obtain a set of calibration sequences, which are then stored in a calibration sequence document for later use.

[0009] S2. Real-time image acquisition: Real-time thermal image data of the blast furnace charging chute area is acquired using an infrared thermal imager;

[0010] S3. Image preprocessing: Denoising and enhancement preprocessing operations are performed on the acquired real-time thermal image data;

[0011] S4. Access control signals: Integrate control information from the PLC of the integrated control system, determine the rotation direction of the chute, and set the detection timing;

[0012] S5. Chute edge feature extraction: Extract key features, including chute edges, from a single frame image. Combine time series and sliding window algorithm with the calibration sequence to determine whether the chute edge contour is detected at the current time.

[0013] S6. Angle Calculation: Calculate the tilting angle of the chute based on the chute edge contour sequence data determined by feature extraction, and feed it back to the control center;

[0014] S7. Real-time monitoring and feedback: Repeat steps S2-S6 to complete real-time monitoring of angle changes and feedback to the control system PLC.

[0015] As a preferred embodiment, the determination of the chute rotation direction and the setting of the detection timing in step S4 are specifically as follows:

[0016] Data acquisition: Collect the current position information provided by the PLC at set time intervals, add the current timestamp, and store it in the rotation angle buffer for later use;

[0017] Rotation direction determination: Set a sliding window and statistically analyze the changing trend of position information in the current and previous window periods. If the rotation angle value increases continuously over time, it is determined that the chute is rotating clockwise; otherwise, it is determined that the chute is rotating counterclockwise.

[0018] Detection timing setting: Based on the setting position of the infrared thermal imager, set the corresponding chute rotation angles at the start and end of the detection process during chute rotation, and record them as the start angle value cfg_rotateAngleB and the end angle value cfg_rotateAngleE, respectively, to ensure that the chute is in the area between the blast furnace flame and the thermal imager; the values ​​of cfg_rotateAngleB and cfg_rotateAngleE are both in the range of 0 to 360 degrees, and are not equal;

[0019] When the set start angle value cfg_rotateAngleB is less than the end angle value cfg_rotateAngleE, if the current chute rotation angle β obtained from the control system PLC is within the range of [cfg_rotateAngleB, cfg_rotateAngleE], the subsequent steps are executed; if it exceeds the range and the image cannot clearly show the chute outline, the subsequent detection is not executed.

[0020] When the set start angle value cfg_rotateAngleB is greater than the end angle value cfg_rotateAngleE, if the current chute rotation angle β obtained from the control system PLC is within the range of [cfg_rotateAngleB, 360] or [0, cfg_rotateAngleE], the subsequent steps are executed; if it exceeds the range and the image cannot clearly show the chute outline, the subsequent detection is not executed.

[0021] As a preferred embodiment, step S5, chute edge feature extraction, includes edge detection, straight line filtering, and intersection segment removal to obtain a candidate set of edge contour lines; specifically as follows:

[0022] Edge detection: For each frame of temperature data acquired from the infrared thermal imager, the temperature to bitmap algorithm is first used to preprocess the data into an infrared image in grayscale format. The Canny operator is used to detect edges, and the probabilistic Hough transform is used to detect straight lines, resulting in a set of line segments in the image. These line segments include the projection lines of the chute contour edge in the two-dimensional plane of the infrared image.

[0023] Linear filtering: Filtering based on multi-dimensional features, including:

[0024] Boundary check: Ensure that the endpoints of the projection lines are within the image area, set the x-coordinate value range to [0, width), and set the y-coordinate value range to [0, height);

[0025] Projection line center point Y coordinate range filtering: restricts the y coordinate of the line segment center point to be within the valid range; if the y value is too large or too small, it is determined that it is not the edge of the chute.

[0026] Projection line and X-axis intersection filtering: Remove straight lines whose intersection with the X-axis is not within the specified range (cfg_cpXMin, cfg_cpXMax); In three-dimensional space, the chute rotates around the central axis, so that the projection line of the chute edge in the picture only moves back and forth within a certain range. At this time, the x-coordinate of the intersection point of the projection line and the X-axis (y=0) only moves back and forth within a specific range.

[0027] Filtering the projection position of the center point of the projection line on the central axis: Eliminate line segments whose projection point on the central axis is in the upper half of the central axis;

[0028] Angle difference filtering: Remove projection lines whose angle with the central axis exceeds the preset threshold cfg_angleMax;

[0029] Crossing segment removal: Detect and remove intersecting line segments to avoid interfering with angle calculation;

[0030] After completing the above steps, the final set of projection lines is considered as the candidate edge contours of the chute. If the set is empty, it means that no edge contours that meet the conditions have been detected and no further operations are performed. If the set of projection lines is not empty at this time, the maximum, minimum and average values ​​of each projection angle in the current set are counted, and the timestamp, projection line data and statistical data at this time are stored in the effective frame buffer for later use.

[0031] As a preferred embodiment, in the process of extracting the chute edge features in step S5, distance filtering is performed after the angle difference filtering is completed: projection lines that are more than the preset threshold cfg_distanceMax at a distance from the center line are removed.

[0032] As a preferred embodiment, in the process of extracting the chute edge features in step S5, straight line merging is performed before the removal of intersecting line segments: line segments with the same direction and a distance less than a set value are merged, and the coordinates of the merged line segments are calculated using a weighted average method, with the weight being the line segment length.

[0033] As a preferred embodiment, the determination of the chute edge contour in step S5, which combines time series analysis and the sliding window algorithm, is as follows:

[0034] When performing chute edge feature extraction on the infrared image data acquired at the current time, if the obtained set of projection lines is not empty, then chute edge determination and projection angle calculation are performed; in the effective frame buffer, effective frame data of the sliding window number is taken from the current frame backward, and the trend_slope of the maximum, minimum and average projection angles is analyzed with time using the linear regression method; the calculated trend_slope is stored in the effective frame buffer for later use.

[0035] Based on the aforementioned determination of the chute rotation direction and the rotation angle during detection, it can be seen that the determination rule for the chute edge is: within the set sliding window, if the trend_slope value within the limited range [ts_min, ts_max] exceeds the set percentage threshold of the total value, then the projection line detected in the current frame is considered to be the chute edge.

[0036] If the set of projection lines in the current frame is determined to be the outline of the chute edge, then the average value of the projection angles in the current set is taken as the projection angle of the chute edge at this time.

[0037] As a preferred embodiment, the angle calculation in step S6 is performed as follows:

[0038] Extract the current frame and the data frames identified as chute edge contours within the previous second from the effective frame buffer based on the current frame's timestamp. Calculate the rate of change of the projection angle between each frame and the previous frame. Get a v i A sequence (i = 1, 2, ..., n) is generated; the sequence v is... i By comparing (i = 1, 2, ..., n) with the calibration sequence document stored in the system, the current actual tilt angle can be obtained.

[0039] As a preferred embodiment, the rectangular coordinate system of the infrared image is set as follows: the upper left corner of the two-dimensional infrared image data is regarded as the origin of the coordinate system, the horizontal direction to the right through the origin of the coordinate system is the positive direction of the X-axis of the rectangular coordinate system, and the vertical direction downward through the origin of the coordinate system is the positive direction of the Y-axis of the rectangular coordinate system.

[0040] As a preferred embodiment, the central axis is the projection line of the central axis of the chute rotation in three-dimensional space onto the two-dimensional plane of the infrared image, denoted as cl, with the upper endpoint denoted as cl_p0(cl_x0, cl_y0) and the lower endpoint denoted as cl_p1(cl_x1, cl_y1).

[0041] As a preferred embodiment, the projection line is: a line segment in three-dimensional space projected onto the two-dimensional plane of the infrared image, denoted as line(x0, y0; x1, y1); the projection angle of this line segment is: the angle θ between the projection line and the positive direction of the X-axis of the coordinate system, with a value range of [0, π); the center point of the projection line is denoted as centerPoint(centerX, centerY), where...

[0042] The beneficial effects of this invention are:

[0043] High-precision measurement: This method improves accuracy by at least 30% compared to traditional methods through multi-dimensional linear filtering and multi-source data fusion correction.

[0044] High reliability: This method avoids direct contact between the equipment and harsh environments such as high temperature, dust, and vibration through non-contact measurement, which extends the service life of the equipment by 2-3 times.

[0045] High robustness: This method uses optimized algorithms such as straight line merging and cross-elimination, which can accurately identify the edge of the chute under complex working conditions and has strong anti-interference ability.

[0046] Real-time response: This method has a short algorithm execution time, with a response time of less than 100ms, and supports 24-hour real-time monitoring.

[0047] Closed-loop control: This method forms a closed loop with the PLC system to achieve real-time feedback and control of measurement results, thereby improving control accuracy.

[0048] Easy maintenance: This method reduces the frequency and cost of equipment maintenance, with annual maintenance costs reduced by approximately 40%.

[0049] Production optimization: The precise angle control of this method helps to optimize the material distribution and improve the blast furnace utilization coefficient by 1-2%.

[0050] Safety Enhancement: This method avoids close-range manual measurement, thus improving operational safety. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the structure of the blast furnace and the charging chute.

[0052] Figure 2 This is a schematic diagram of the rectangular coordinate system for infrared images. Detailed Implementation

[0053] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0054] like Figure 1-2 As shown, a method for online detection of the tilt angle of a blast furnace charging chute includes the following steps:

[0055] S1. Infrared thermal imager installation and calibration sequence acquisition: Install an infrared thermal imager at a suitable position on the top of the blast furnace to ensure complete imaging of the charging chute area; after installing the infrared thermal imager at the target position, rotate the chute in 5-degree increments to obtain a set of calibration sequences, which are then stored in a calibration sequence document for later use.

[0056] S2. Real-time image acquisition: Real-time thermal image data of the blast furnace charging chute area is acquired using an infrared thermal imager;

[0057] S3. Image preprocessing: Denoising and enhancement preprocessing operations are performed on the acquired real-time thermal image data;

[0058] S4. Access control signals: Integrate control information from the PLC of the integrated control system, determine the rotation direction of the chute, and set the detection timing;

[0059] S5. Chute edge feature extraction: Extract key features, including chute edges, from a single frame image. Combine time series and sliding window algorithm with the calibration sequence to determine whether the chute edge contour is detected at the current time.

[0060] S6. Angle Calculation: Calculate the tilting angle of the chute based on the chute edge contour sequence data determined by feature extraction, and feed it back to the control center;

[0061] S7. Real-time monitoring and feedback: Repeat steps S2-S6 to complete real-time monitoring of angle changes and feedback to the control system PLC.

[0062] The rectangular coordinate system of the infrared image is set as follows: the upper left corner of the two-dimensional infrared image data is regarded as the origin of the coordinate system, the horizontal direction to the right through the origin of the coordinate system is the positive direction of the X-axis of the rectangular coordinate system, and the vertical direction downward through the origin of the coordinate system is the positive direction of the Y-axis of the rectangular coordinate system.

[0063] The central axis is the projection line of the central axis of the chute rotation in three-dimensional space onto the two-dimensional plane of the infrared image, denoted as cl, with the upper endpoint denoted as cl_p0(cl_x0,cl_y0) and the lower endpoint denoted as cl_p1(cl_x1,cl_y1).

[0064] The projection line is a line segment in three-dimensional space projected onto the two-dimensional plane of the infrared image, denoted as line(x0, y0, x1, y1); the projection angle of this line segment is the angle θ between the projection line and the positive direction of the X-axis of the coordinate system, with a value range of [0, π); the center point of the projection line is denoted as centerPoint(centerX, centerY), where...

[0065] The specific process for determining the chute rotation direction and setting the detection timing in step S4 is as follows:

[0066] Data acquisition: Collect the current position information provided by the PLC at set time intervals, add the current timestamp, and store it in the rotation angle buffer for later use;

[0067] Rotation direction determination: Set a sliding window and statistically analyze the changing trend of position information in the current and previous window periods. If the rotation angle value increases continuously over time, it is determined that the chute is rotating clockwise; otherwise, it is determined that the chute is rotating counterclockwise.

[0068] Detection timing setting: Based on the setting position of the infrared thermal imager, set the corresponding chute rotation angles at the start and end of the detection process during chute rotation, and record them as the start angle value cfg_rotateAngleB and the end angle value cfg_rotateAngleE, respectively, to ensure that the chute is in the area between the blast furnace flame and the thermal imager; the values ​​of cfg_rotateAngleB and cfg_rotateAngleE are both in the range of 0 to 360 degrees, and are not equal;

[0069] When the set start angle value cfg_rotateAngleB is less than the end angle value cfg_rotateAngleE, if the current chute rotation angle β obtained from the control system PLC is within the range of [cfg_rotateAngleB, cfg_rotateAngleE], the subsequent steps are executed; if it exceeds the range and the image cannot clearly show the chute outline, the subsequent detection is not executed.

[0070] When the set start angle value cfg_rotateAngleB is greater than the end angle value cfg_rotateAngleE, if the current chute rotation angle β obtained from the control system PLC is within the range of [cfg_rotateAngleB, 360 or [0, cfg_rotateAngleE], the subsequent steps are executed; if it exceeds the range and the image cannot clearly show the chute outline, the subsequent detection is not executed.

[0071] Step S5, chute edge feature extraction, includes edge detection, straight line filtering, and intersection segment removal to obtain a candidate set of edge contour lines; specifically as follows:

[0072] Edge detection: For each frame of temperature data acquired from the infrared thermal imager, the data is first preprocessed into a grayscale infrared image using a temperature-to-bitmap algorithm. The Canny operator is then used to detect edges, and the probabilistic Hough transform is used to detect straight lines, resulting in a set of line segments in the image. These line segments include the projection lines of the chute contour edges onto the two-dimensional plane of the infrared image. (In a special case, when a line segment in three-dimensional space is perpendicular to the two-dimensional plane of the infrared image, the projection of that line segment is a point, meaning the two endpoints of the line segment coincide.)

[0073] Linear filtering: Filtering based on multi-dimensional features, including:

[0074] Boundary check: Ensure that the endpoints of the projection lines are within the image area, set the x-coordinate value range to [0, width), and set the y-coordinate value range to [0, height);

[0075] Projection line center point Y coordinate range filtering: restricts the y coordinate of the line segment center point to be within the valid range; if the y value is too large or too small, it is determined that it is not the edge of the chute.

[0076] Projection line and X-axis intersection filtering: Remove straight lines whose intersection with the X-axis is not within the specified range (cfg_cpXMin, cfg_cpXMax); In three-dimensional space, the chute rotates around the central axis, so that the projection line of the chute edge in the picture only moves back and forth within a certain range. At this time, the x-coordinate of the intersection point of the projection line and the X-axis (y=0) only moves back and forth within a specific range.

[0077] Filtering the projection position of the center point of the projection line on the central axis: Eliminate line segments whose projection point on the central axis is in the upper half of the central axis;

[0078] Angle difference filtering: Remove projection lines whose angle with the central axis exceeds the preset threshold cfg_angleMax;

[0079] Distance filtering: Remove projection lines whose distance from the center line exceeds the preset threshold cfg_distanceMax.

[0080] Line merging: Merge line segments that are in the same direction and whose distance is less than a set value. Calculate the coordinates of the merged line segment using a weighted average method, with the weight being the line segment length.

[0081] Crossing segment removal: Detect and remove intersecting line segments to avoid interfering with angle calculation;

[0082] After completing the above steps, the final set of projection lines is considered as the candidate edge contours of the chute. If the set is empty, it means that no edge contours that meet the conditions have been detected and no further operations are performed. If the set of projection lines is not empty at this time, the maximum, minimum and average values ​​of each projection angle in the current set are counted, and the timestamp, projection line data and statistical data at this time are stored in the effective frame buffer for later use.

[0083] The specific process for determining the chute edge contour in step S5, which combines time series analysis and the sliding window algorithm, is as follows:

[0084] When performing chute edge feature extraction on the infrared image data acquired at the current time, if the obtained set of projection lines is not empty, then chute edge determination and projection angle calculation are performed; in the effective frame buffer, effective frame data of the sliding window number is taken from the current frame backward, and the trend_slope of the maximum, minimum and average projection angles is analyzed with time using the linear regression method; the calculated trend_slope is stored in the effective frame buffer for later use.

[0085] Based on the aforementioned determination of the chute rotation direction and the rotation angle during detection, it can be seen that the determination rule for the chute edge is: within the set sliding window, if the trend_slope value within the limited range [ts_min, ts_max] exceeds the set percentage threshold of the total value, then the projection line detected in the current frame is considered to be the chute edge.

[0086] If the set of projection lines in the current frame is determined to be the outline of the chute edge, then the average value of each projection angle in the current set is taken as the projection angle of the chute edge at this time.

[0087] The angle calculation in step S6 is performed as follows:

[0088] Extract the current frame and the data frames identified as chute edge contours within the previous second from the effective frame buffer based on the current frame's timestamp. Calculate the rate of change of the projection angle between each frame and the previous frame. Get a v i A sequence (i = 1, 2, ..., n) is generated; the sequence v is... iBy comparing (i = 1, 2, ..., n) with the calibration sequence document stored in the system, the current actual tilt angle can be obtained.

[0089] The above embodiments are merely illustrative of the principles and effects of the present invention, as well as some examples of its application, and are not intended to limit the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept of the present invention, and these modifications and improvements are all within the scope of protection of the present invention.

Claims

1. A method for online detecting the tilt angle of a blast furnace distribution chute, comprising the following steps: S1. Infrared thermal imager installation and calibration sequence collection: install the infrared thermal imager at a suitable position on the blast furnace top to ensure that the distribution chute area can be completely photographed; after the infrared thermal imager is installed at the target position, rotate the chute at every 5 degrees as a step, and obtain a set of calibration sequence stored in a calibration sequence document for standby; S2. Real-time image collection: collect real-time thermal image data of the blast furnace distribution chute area using the infrared thermal imager; S3. Image preprocessing: perform denoising and enhancement preprocessing operations on the collected real-time thermal image data; S4. Access control signal: integrate the control information in the control system PLC, and determine the rotation direction of the chute and set the detection timing; S5. Chute edge feature extraction: extract key features including the chute edge for a single frame image, and determine whether the chute edge contour is detected at the current time by comparing the time sequence and the sliding window algorithm with the calibration sequence; S6. Angle calculation: calculate the chute tilt angle based on the chute edge contour sequence data determined by the feature extraction, and feed back to the control center; S7. Real-time monitoring and feedback: repeat steps S2-S6 to complete real-time monitoring of angle changes and feedback to the control system PLC.

2. The method for detecting the tilting angle of the blast furnace distributing chute in line according to claim 1, characterized in that: The specific process of determining the rotation direction of the chute and setting the detection timing in step S4 is as follows: Data collection: collect the current position information provided by the PLC at a set time period, and store it in the rotation angle buffer area with a current timestamp for standby; Rotation direction determination: set a sliding window to count the change trend of the position information in the current and previous window period. If the rotation angle value increases with time, it is determined that the chute is rotating clockwise, otherwise, it is determined that the chute is rotating counterclockwise; Detection timing setting: according to the setting position of the infrared thermal imager, set the detection start and end rotation angles of the chute during rotation, respectively recorded as the start angle value cfg_rotateAngleB and the end angle value cfg_rotateAngleE parameters, to ensure that the chute is in the area between the blast furnace flame and the thermal imager; The value range of cfg_rotateAngleB and cfg_rotateAngleE is 0 to 360 degrees, and they are not equal; When the set start angle value cfg_rotateAngleB is less than the end angle value cfg_rotateAngleE, if the current chute rotation angle β obtained from the control system PLC is within the range [cfg_rotateAngleB, cfg_rotateAngleE], the subsequent steps are executed; if it exceeds the range, the image cannot clearly present the chute contour, and the subsequent detection is not performed; When the set start angle value cfg_rotateAngleB is greater than the end angle value cfg_rotateAngleE, if the current chute rotation angle β obtained from the control system PLC is in the range of [cfg_rotateAngleB, 360] or [0, cfg_rotateAngleE], the subsequent steps are executed; if it exceeds the range, the image cannot clearly present the chute profile, and the subsequent detection is not performed.

3. The method for detecting the tilting angle of the blast furnace distributing chute in line according to claim 1, characterized in that: The step S5 of extracting the chute edge features includes edge detection, straight line filtering, and cross line segment elimination to obtain an edge contour line candidate set; specifically as follows: Edge detection: for each frame of temperature data obtained from the infrared thermal imager, first pre-process the infrared image data in the format of a grayscale image using a temperature conversion algorithm, detect the edge using a Canny operator, detect the straight line using a probabilistic Hough transform, and obtain the line segment set in the picture; these line segments include the projection lines of the chute profile edge in the two-dimensional plane of the infrared image; Straight line filtering: filtering based on multi-dimensional features, including: Boundary check: ensure that the projection line endpoints are within the image range, set the x-coordinate value range to [0, width), and set the y-coordinate value range to [0, height); Projection line center point y-coordinate range filtering: limit the y-coordinate of the line segment center point to the effective range; if the y value is too large or too small, it is determined that it is not the chute edge; Projection line and X-axis intersection point filtering: eliminate straight lines whose intersection points with the X-axis are not within the specified range (cfg_cpXMin, cfg_cpXMax); in three-dimensional space, the chute rotates around the center axis, causing the projection lines of the chute edge in the picture to only reciprocate within a certain range, so the x-coordinate of the intersection point of the projection line and the X-axis (y=0) only reciprocates within a certain range; Projection line center point projection position filtering on the center axis: eliminate line segments whose projection points on the center axis are on the upper half of the center axis; Angle difference filtering: eliminate projection lines with an angle greater than the preset threshold cfg_angleMax with the center axis; Cross line segment elimination: detect and eliminate intersecting line segments to avoid interference with angle calculation; After the above steps are executed, the remaining projection line set is considered as the chute edge contour candidate; if the set is empty, it indicates that no edge contour meeting the conditions is detected, and the subsequent operation is not performed; if the projection line set is not empty, the maximum value, minimum value, and average value of the projection angles in the current set are calculated, and the timestamp, projection line data, and statistical data are stored in the effective frame buffer for backup.

4. The method for on-line detecting the tilting angle of the blast furnace burden chute according to claim 3, characterized in that: Distance filtering is performed after the angle difference filtering in the chute edge feature extraction process of step S5: eliminate projection lines with a distance greater than the preset threshold cfg_distanceMax from the center line.

5. The method for detecting the tilting angle of the blast furnace distributing chute in line according to claim 3, characterized in that: Before cross line segment elimination in the chute edge feature extraction process of step S5, straight line merging is performed first: merge line segments with consistent directions and distances less than a set value, and calculate the coordinates of the merged line segments using a weighted average method, with the weight being the line segment length.

6. The method for on-line detecting the tilting angle of the blast furnace burden chute according to claim 1, characterized in that: The chute edge profile determination combined with time series and sliding window algorithm in step S5 is specifically as follows: When performing the chute edge feature extraction on the infrared image data obtained at the current time, if the obtained projection line set is not empty, the chute edge determination and projection angle calculation are performed; the effective frame data of the sliding window number is taken from the current frame in the effective frame buffer, and the linear regression method is used to analyze the trend trend_slope of the maximum value, minimum value and average value of the projection angle with time; the calculated trend_slope is stored in the effective frame buffer for standby; According to the chute rotating direction determined by the foregoing determination and the rotating angle at the time of detection, it is known that the determination rule of the chute edge is that, within the set sliding window, the trend_slope value within the limited range [ts_min, ts_max] exceeds the set percentage threshold of the total value, that is, it is determined that the projection line detected in the current frame is the chute edge; If it is determined that the projection line set of the current frame is the chute edge profile, the average value of each projection angle in the current set is regarded as the projection angle of the chute edge at this time.

7. The method for detecting the tilting angle of the blast furnace distributing chute in line according to claim 1, characterized in that: The angle calculation of step S6 is specifically as follows: The current frame and the data frame determined as the edge profile of the chute within 1 second before the current frame are extracted from the effective frame buffer according to the timestamp of the current frame, and the change speed of the projection angle of each frame and the previous frame is calculated obtain a v i (i = 1, 2,..., n) sequence; compare the sequence v i (i = 1, 2,..., n) with the calibration sequence document stored in the system, and the current actual tilt angle can be obtained.

8. The method for detecting the tilting angle of the blast furnace distributing chute in line according to any one of claims 1-7, characterized in that: The setting mode of the rectangular coordinate system of the infrared image is that the upper left corner of the two-dimensional infrared image data is regarded as the origin of the coordinate system, the horizontal right direction through the origin of the coordinate system is the positive direction of the X-axis of the rectangular coordinate system, and the vertical downward direction through the origin of the coordinate system is the positive direction of the Y-axis of the rectangular coordinate system.

9. The method for on-line detecting the tilting angle of the distribution chute of a blast furnace according to claim 8, characterized in that: The center axis is the projection line of the center axis of the chute rotating motion in the three-dimensional space in the two-dimensional plane of the infrared image, recorded as cl, the line segment of the upper end point cl_p0(cl_x0, cl_y0) and the lower end point cl_p1(cl_x1, cl_y1).

10. The method for on-line detecting the tilting angle of the blast furnace distributing chute according to claim 9, characterized in that: The projection line is a line segment in a three-dimensional space projected into a line segment in a two-dimensional plane of an infrared image, denoted as line(x0, y0, x1, y1); the projection angle of the line segment is an included angle θ between the projection line and a positive direction of an X-axis of a coordinate system, and the value range is [0, π); and the projection line center point of the line segment is denoted as a point centerPoint(centerX, centerY), wherein

Images