A method for online real-time measurement of steel plates based on line structured light

Abstract

The present invention relates to an online real-time measurement method for steel plates based on line structured light, wherein the camera tracks the laser line irradiated to the surface of the steel plate in real time and continuously collects the current state image, and transmits the collected real-time image back to the host computer for analysis and processing, and calculates the length of the steel plate according to the coordinate transformation model; the host computer transmits the steel plate length information to the roller control system to realize the start and stop control of the roller, and then cooperates with the shearing machine to realize segmented shearing. The present invention proposes a new detection method in calibration and sizing by designing a software operating system, which can improve the detection accuracy of steel plates and realize the precise sizing of long steel plates on the production line. At the same time, the length measurement system assists the shearing machine to realize accurate shearing of steel plates.

Description

Technical Field

[0001] The invention belongs to the field of irregular surface length measurement, and in particular relates to an online real-time measurement method for a steel plate based on line structured light. Background Art

[0002] At present, the domestic steel plate rolling size is mainly based on customer requirements and the size of the cooling bed to select different specifications of slabs for rolling, which cannot achieve continuous hot rolling of steel plates, so the production efficiency is low. Therefore, it is necessary to use segmented shears to cut the continuous rolled steel plates to a fixed length. In the past, domestic factories usually used manual measurement or installation of length measuring rollers to detect the fixed length of steel plates. Due to the high risk factor of manual measurement, poor measurement accuracy and low detection efficiency, and the use of length measuring rollers to measure, there are factors such as roller surface wear, which leads to large measurement errors and inaccurate sizing, resulting in a large loss of industrial production benefits. Therefore, it is necessary to design a steel plate non-destructive testing technology based on line structured light to solve the real-time online length measurement of steel plates, so as to realize the automation of the segmented shearing process, speed up the rolling rate, reduce manual participation, and improve detection efficiency and real-time detection. Summary of the invention

[0003] In order to solve the problems raised in the above background technology, the present invention mainly provides an online real-time measurement method of steel plates based on line structured light.

[0004] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an online real-time measurement method for steel plates based on line structured light, wherein the camera tracks the laser line irradiated to the surface of the steel plate in real time and continuously collects the current state image, and transmits the collected real-time image back to the host computer for analysis and processing, and calculates the length of the steel plate according to the coordinate transformation model; the host computer transmits the steel plate length information to the roller control system to realize the start and stop control of the roller, and then cooperates with the shearing machine to realize segmented shearing. The present invention proposes a new detection method in calibration and sizing by designing a software operating system, which can improve the detection accuracy of steel plates and realize the precise sizing of long steel plates on the production line. At the same time, the length measurement system assists the shearing machine to realize accurate shearing of steel plates.

[0005] In a preferred embodiment, the online measurement method comprises the following specific steps:

[0006] Step 1: Calibrate several cameras. After adjusting the camera and laser to the appropriate posture, take the shear blade of the shearing machine as the starting point, lay two 50m long steel rulers on both sides of the laser line on the roller surface along the exit direction of the steel plate, and use the flat white plastic sheet as the calibration object and find its center line position and mark it; secondly, place the plastic sheet on the two steel ruler surfaces within the field of view of the camera to be calibrated, and then align the center line of the plastic sheet with the scale of one of the steel rulers, and record the actual scale value of the point in the software system. In the same way, find multiple different positions under each camera field of view to record the actual scale value; in addition, in the pixel coordinate system, the position coordinates of the two ends of the plastic sheet are obtained through the mutation formed on both sides after the laser line hits the plastic sheet, and the average value is calculated by the software system to obtain the middle position coordinates, and then the conversion relationship between the pixel coordinates and the world coordinates is used to calculate the actual position corresponding to the center line;

[0007] Step 2: After the calibration is completed, according to the ambient light and the reflection of the roller surface on site, set the exposure time of each camera to between 4000-5000μs and the gain to between 0-23.9dB in the software operating system to minimize external factors and meet the measurement accuracy requirements;

[0008] Step 3: By setting the ROI area, observe the real-time position of the steel plate in the image, track the position of the plate head in real time, and output the actual distance between the head of the steel plate and the shear blade of the shearing machine through coordinate calculation.

[0009] Step 4: Record the background of the images collected by each camera. Record once when the roller is stationary, and once when the roller is rotating before the steel plate passes through the shearing machine and enters the field of view of the first camera. This can better consider the impact of redundant bright spots in the image caused by vibration when the roller is rotating on the length measurement of the steel plate;

[0010] In a preferred embodiment, let the steel plate head tracking point be A, then its coordinates in the world coordinate system (X W ,Y W ,Z W ) has the following relationship with the coordinates (u, v) in the pixel coordinate system:

[0011] (1)

[0012] Where: s is the value of the point in the z direction in the camera coordinate system, mm; , f is the focal length of camera 3, mm, dx, dy are the physical sizes of a single pixel, mm; u0, v0 are the centers of the image plane, pixel; is the rotation matrix from the world coordinate system to the camera coordinate system, The translation matrix from the world coordinate system to the camera coordinate system.

[0013] In the actual calibration process, in the world coordinate system O-XwYwZw where the laser line is located, only the value of the Xw axis direction needs to be known, and the values ​​of the other directions are all 0; in the pixel coordinate system o-uv where the laser line is located, only the value of the u axis needs to be known, and the value of the v axis is treated as 0, so the above formula can be changed to:

[0014] (2)

[0015] Where: , ;

[0016] By converting the contents on both sides of formula (2), we can get the following formula:

[0017] (3)

[0018] Where: , , ;

[0019] Converting formula (3) into matrix form yields:

[0020] (4)

[0021] By selecting the coordinates of n points (n ≥ 3) in the world coordinate system and the camera coordinate system in space, the conversion relationship between the world coordinate system and the pixel coordinate system of point A can be obtained: , complete the camera 3 calibration in practical applications.

[0022] Transforming formula (3) yields:

[0023] (5)

[0024] In formula (5), X is the actual position coordinate of the tracking point A; R0, R1, R2 are conversion coefficients; and u is the pixel coordinate in the horizontal direction.

[0025] In a preferred embodiment, according to the on-site steel plate thickness of 10-50 mm, taking into account the actual size of the acquired image and the measurement target, the ROI area is set to 4096*130 pixels, which can greatly improve the image processing speed, remove irrelevant image information, observe the image in real time and quickly track the position of the steel plate, and solve the problem of slow image processing speed.

[0026] In a preferred embodiment, due to the surface characteristics of the hot-rolled steel plate, there are breakpoints in the laser line on the surface of the steel plate in the image captured in real time by the camera. Therefore, by traversing the pixel data in each column of the ROI image area, the leftmost position of the laser line is preferentially found, and then the pixel data is traversed from the left to the right row by row starting from the leftmost position to find intermittent pixel data, and finally the pixels between the breakpoints are filled with data.

[0027] In a preferred embodiment, in order to ensure the reliability of measurement, there is an overlapping area between adjacent cameras, and a method based on grayscale value comparison is used for the overlapping area. This method is to use weighted grayscale processing on the overlapping area images collected by two cameras with overlapping areas arranged continuously, and compare the average grayscale values ​​of the two. The image with a larger grayscale value is set as the processable image area, and the other is discarded to achieve the processing division of the overlapping area, avoiding multiple processing of the overlapping area and inconsistent results.

[0028] In a preferred embodiment, during the background recording process, the laser line is irradiated at the center of the roller, and the camera collects images of the roller when it is static and when the roller is rotating when no steel plate passes. At this time, the image only contains the laser line irradiated to the roller surface and the roller gap, and the laser line highlight points in the v direction in the image are recorded in the image background through an image traversal method; when the steel plate enters the roller, the camera collects the image with the steel plate in real time, and the bright spot at the head of the steel plate is obviously higher than the bright spot recorded in the previous background. At this time, the v value corresponding to the bright spot position at the head of the steel plate in the real-time image is subtracted from the v value corresponding to the position in the background image record to remove the low bright spot recorded in the previous background; and the central grayscale value of the bright spot appearing in the real-time collected image is >15, so the bright spot is the point required for real-time tracking.

[0029] In a preferred embodiment, the grayscale centroid method is used to refine the laser center line, and the formula is as follows:

[0030]

[0031] where g i is the gray value of the pixel involved in the operation, u i is the corresponding horizontal pixel coordinate, and U is the horizontal grayscale centroid coordinate. Accurate positioning is achieved by fitting the grayscale center point set.

[0032] In a preferred embodiment, the calculated length of the steel plate is fed back to the roller control system, and the driving motor advances the steel plate to a specified length position.

[0033] In a preferred embodiment, the host computer sends a shearing instruction, and the shearing machine starts precise segmented shearing.

[0034] The beneficial effects of the present invention are:

[0035] 1. The present invention achieves accurate length measurement of steel plates by fixing a camera and a laser beside a roller table through a bracket, which has low cost and convenient operation and maintenance.

[0036] 2. The present invention utilizes digital image processing technology to perform real-time processing and analysis of the on-site steel plate conditions, thereby realizing automation and intelligence of the production line and enabling real-time detection of the steel plate length.

[0037] 3. The present invention can greatly improve production efficiency and reduce manual labor intensity through the configuration of intelligent equipment, combined with the developed software system and real-time online detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is the overall structure diagram of the length measurement system.

[0039] Figure 2 It is a diagram of the components of each system.

[0040] Figure 3 It is a flow chart of real-time length detection of steel plates.

[0041] Figure 4 This is the principle diagram of steel plate length measurement.

[0042] Figure 5 This is the calibration data recording interface diagram. DETAILED DESCRIPTION

[0043] By referring to the exemplary embodiments, the objects and functions of the present invention and the methods for achieving these objects and functions will be explained. However, the present invention is not limited to the exemplary embodiments disclosed below; it can be implemented in different forms. The essence of the description is only to help those skilled in the relevant art to further understand the specific details of the present invention. In the following, embodiments of the present invention will be described with reference to the accompanying drawings. The same reference numerals represent the same or similar parts, or the same or similar steps.

[0044] This will be described below by means of specific embodiments. Figure 1-5 :A method for online real-time measurement of steel plates based on line structured light, the method is that the camera 3 tracks the laser line irradiated to the surface of the steel plate in real time and continuously collects the current state image, and transmits the collected real-time image back to the host computer for analysis and processing, and calculates the length of the steel plate according to the coordinate transformation model; the host computer transmits the steel plate length information to the roller control system to realize the start and stop control of the roller, and then cooperates with the shearing machine 1 to realize segmented shearing. The present invention proposes a new detection method in calibration and sizing by designing a software operating system, which can improve the detection accuracy of steel plates and realize the precise sizing of long steel plates on the production line. At the same time, the length measurement system assists the shearing machine 1 to realize accurate shearing of steel plates.

[0045] The camera 3 is calibrated by placing a plastic sheet with a flat surface on two steel ruler surfaces within the corresponding field of view of the camera 3 to be calibrated, and then aligning the center line of the plastic sheet with one of the steel ruler scales, and recording the actual scale value of the point in the software system. In the same way, find multiple different positions under the field of view of each camera 3 to record the actual scale value.

[0046] In the pixel coordinate system, the position coordinates of the two ends of the plastic sheet are obtained through the mutations formed on both sides after the laser line hits the flat plastic sheet. The middle position coordinates are obtained by calculating the average value through the software system, and then the actual position corresponding to the center line is calculated using the conversion relationship between the pixel coordinates and the world coordinates.

[0047] The real-time image needs to set the ROI area, observe the real-time position of the steel plate in the image, track the plate head position in real time, and output the actual distance between the head of the steel plate and the shear blade of the shearing machine through coordinate calculation.

[0048] The coordinate transformation model is the transformation between the world coordinate system and the pixel coordinate system, specifically as follows: the coordinates in the world coordinate system (X W ,Y W ,Z W ) has the following relationship with the coordinates (u, v) in the pixel coordinate system:

[0049]

[0050] Where X is the actual position coordinate of tracking point A; R0, R1, R2 are conversion coefficients; u is the pixel coordinate in the horizontal direction. If the pixel coordinate of the target laser point is known, the coordinate of the target laser point in the world coordinate system can be obtained by the derived coordinate conversion relationship.

[0051] The overlapping areas between the cameras 3 are processed by a method based on gray value comparison to avoid inconsistent measurement results.

[0052] The laser line thinning method adopts the grayscale centroid method to extract the grayscale centroid of the laser line image, thereby improving the measurement accuracy.

[0053] The background record is to record the coordinate values ​​of the highlight points in the v direction in the collected image, make a difference between the v values ​​of the same position in the image with and without the steel plate, and compare the grayscale value of the center of the highlight point.

[0054] The length of the steel plate is transmitted to the roller control system through the upper computer controller, and the shearing machine 1 is used to achieve precise shearing.

[0055] The specific working process of the present invention is as follows: First, the arranged camera 3 is calibrated. The positions of the camera 3 and the laser 2 are adjusted to appropriate angles respectively, so that the laser line emitted by the laser 2 can irradiate the center position of the roller 4 surface. The shear blade 1 of the shearing machine is used as the starting position, and the 0m position of two 50m long steel rulers is fixed at the starting position, and they are spread out at a certain interval on the roller surface along the steel plate outlet direction to ensure the straightness of the steel rulers, and at the same time, ensure that the laser line is in the middle of the two steel rulers; a regular white plastic sheet with a flat surface is used as the calibration object, and its center line position is found and marked, and the white plastic sheet is placed on the surface of the two steel rulers within the corresponding field of view of the camera 3 to be calibrated so that the laser line can irradiate the white plastic sheet. Then align the center line of the white plastic sheet with one of the steel ruler scales, and record the actual scale value of the point in the software system. In the same way, find multiple different positions under the field of view of each camera 3 to record the actual scale value; then obtain the position coordinates of the two ends of the white plastic sheet in the pixel coordinate system through the mutation formed on both sides after the laser line hits the white plastic sheet, and calculate the average value of the two ends through the software system to obtain the middle position coordinates. After calculating the actual position corresponding to the center line through the conversion relationship between different coordinate systems, compare the actual position coordinates with the position coordinates calculated by the system to achieve high-precision positioning;

[0056] Secondly, set the parameters for all cameras 3. According to the influence of ambient light and roller surface reflection on site, set the exposure time of each camera 3 to between 4000-5000μs and the gain size to between 0-23.9dB in the software operating system to minimize external influencing factors; then set the ROI area in the software interface. Combined with the requirement of 10-50mm thickness of steel plate detection on site, set the ROI selection area to 4096×130pixel, traverse the pixel data in each column of the ROI area to find the leftmost position of the laser line, traverse the pixel data from left to right from the leftmost position, set the position with less than 10 pixels between two consecutive pixel data as the breakpoint position of the laser line, and fill the pixels between the two points with data. The image is displayed as the completion of the laser line breakpoints until all the pixel data in the ROI area are completed, so that the laser line is in a continuous state. In addition, the center position of the laser line is refined by the grayscale centroid method, and the grayscale centroid of the completed laser line image is extracted. The formula is as follows:

[0057]

[0058] where g i is the gray value of the pixel involved in the operation, u i is the corresponding pixel coordinate in the horizontal direction, and U is the grayscale centroid coordinate in the horizontal direction.

[0059] Then, the background of the image captured by each camera 3 is recorded. When the roller 4 is stationary and when the roller 4 is rotating, the background image of the steel plate is collected before it passes through the shearing machine 1 and enters the field of view of the first camera, and the high point in the v direction corresponding to u in the image is recorded; when the steel plate enters the roller 4, the camera 3 collects the image with the steel plate in real time, and the bright spot at the head of the steel plate is obviously higher than the bright spot recorded in the previous background. At this time, the v value corresponding to the bright spot position of the plate head in the real-time image is subtracted from the v value corresponding to the position in the background image record, and the low bright spot recorded in the previous background is removed; and the central grayscale value of the bright spot in the real-time collected image is >15, Therefore, the bright spot is the point that needs to be tracked in real time; and the central grayscale value of the bright spot in the real-time collected image is >15, so the bright spot is the point that needs to be tracked in real time, observe the real-time position of the steel plate in the image, track the plate head position in real time and output the actual distance between the head of the steel plate and the shearing blade of the shearing machine 1 through coordinate calculation, and then calculate the length of the steel plate; finally, the detected length of the steel plate is transmitted to the roller control system, and the roller 4 is driven to rotate through the PLC program segment. When the desired length position is reached, the roller 4 stops, and the shearing machine 1 is started for segmented shearing, and the segmented steel plate is transported to the corresponding cooling bed for cooling.

Claims

1. A method for online real-time measurement of steel plates based on line structured light, characterized in that: The steps include: Step 1: Calibrate several cameras (3); Adjust the position of the camera (3) and the laser (2) so that the laser line is aligned with the center of the roller and remains horizontal, and the camera field of view is able to observe the five rotating rollers and the laser line in the collected image is horizontal; Taking the shear blade of the shearing machine (1) as the starting point, two 50m long steel rulers are laid flat on both sides of the laser line on the surface of the roller (4) along the exit direction of the steel plate, and a white plastic sheet with a smooth surface is used as the calibration object, and its center line position is found and marked; secondly, the plastic sheet is placed on the two steel ruler surfaces within the corresponding field of view of the camera (3) to be calibrated, and then the center line of the plastic sheet is aligned with the scale of one of the steel rulers, and the actual scale value of the point is recorded in the software system. In the same way, multiple different positions are found in the field of view of each camera (3) to record the actual scale value; in addition, in the pixel coordinate system, the position coordinates of the two ends of the plastic sheet are obtained through the mutation formed on both sides after the laser line hits the plastic sheet, and the middle position coordinates are obtained by calculating the average value through the software system, and then the actual position corresponding to the center line is calculated by using the conversion relationship between pixel coordinates and world coordinates; Step 2: After the calibration is completed, according to the ambient light and the reflection of the roller surface at the scene, the exposure time of each camera (3) is set to between 4000-5000 μs and the gain is set to between 0-23.9 dB in the software operating system to minimize the external influence factors and meet the measurement accuracy requirements; Step 3: By setting the ROI area, observe the real-time position of the steel plate in the image, track the position of the plate head in real time, and output the actual distance between the head of the steel plate and the shear blade of the shearing machine through coordinate calculation; Step 4: Record the background of the image collected by each camera (3); The data are recorded once when the roller conveyor (4) is stationary, and once when the roller conveyor (4) is rotating before the steel plate passes through the shearing machine (1) and enters the field of view of the first camera (3).

2. The method for online real-time measurement of a line structured light steel plate according to claim 1, characterized in that: Coordinates in the world coordinate system (X W ,Y W ,Z W ) has the following relationship with the coordinates (u, v) in the pixel coordinate system: Where X is the actual position coordinate of the tracking point A; R0, R1, R2 are conversion coefficients; u is the pixel coordinate in the horizontal direction; If the pixel coordinates of the target laser point are known, the coordinate position of the target laser point in the world coordinate system can be obtained by the coordinate transformation relationship.

3. The method for online real-time measurement of steel plates based on line structured light according to claim 1, characterized in that: According to the on-site steel plate thickness of 10-50mm and considering the size of the actual collected image and the measurement target, the ROI area is set to 4096*130pixel.

4. The method for online real-time measurement of steel plates based on line structured light according to claim 1, characterized in that: Traverse each column of pixel data in the ROI image area to first find the leftmost position of the laser line, then start from the leftmost position and traverse the laser line pixel data from left to right in rows to find intermittent pixel data, and finally fill the pixels between the breakpoints with data.

5. The method for online real-time measurement of steel plates based on line structured light according to claim 1, characterized in that: To ensure the reliability of measurement, there is an overlapping area between adjacent cameras (3), and a method based on gray value comparison is used for the overlapping area; The method is to use weighted grayscale processing on overlapping area images captured by two cameras (3) arranged in succession and having overlapping areas, and compare the average grayscale values ​​of the two, wherein the image with the larger grayscale value is set as the processable image area, and the other is discarded, thereby avoiding multiple processing of the overlapping area and inconsistent results.

6. The method for online real-time measurement of steel plates based on line structured light according to claim 1, characterized in that: During the background recording process, the laser line is irradiated at the center of the roller (4), and the camera (3) collects images of the roller (4) when no steel plate passes through and when the roller (4) is rotating. At this time, only the laser line irradiated on the roller surface and the roller gap is in the image. The laser line highlight points in the v direction in the image are recorded in the image background through the image traversal method; when the steel plate enters the roller, the camera (3) collects the image with the steel plate in real time, and the bright spot at the head of the steel plate is obviously higher than the bright spot recorded in the previous background. At this time, the v value corresponding to the bright spot position at the head of the steel plate in the real-time image is subtracted from the v value corresponding to the position in the background image record, and the low bright spot recorded in the previous background is removed; and the central grayscale value of the bright spot appearing in the real-time collected image is greater than 15, so the bright spot is the point required for real-time tracking.

Images