Method and System for Measuring the Towering Height of the End Face of a Hot-Rolled Steel Coil Based on 3D Point Cloud

Abstract

The present invention provides a method and system for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud, which can solve the problem that quality inspection personnel cannot enter the site to accurately measure the tower coil height when the temperature of the hot-rolled coil is too high. The method includes: Step 1, obtaining the point cloud image of the end face of the steel coil; Step 2, segmenting the point cloud image region; Step 3, respectively segmenting the point cloud images of the bottom and top of the tower of the end face of the steel coil by using the Euclidean distance between each point cloud coordinate, calculating the number of point clouds in the segmented multiple regions and arranging them in ascending order, and preliminarily positioning the image regions of the bottom and top of the tower; Step 4, setting the tower height measurement error value, calculating the mean value of the z coordinate values of all the point clouds in the preliminarily obtained image regions of the bottom and top of the tower, intercepting the set region near the mean value, and accurately positioning the bottom and top regions of the tower; Step 5, calculating the tower height.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional measurement, and particularly relates to a method and system for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud. Background Art

[0002] The quality of hot-rolled steel coils is the most important quality index in the hot continuous rolling coiling area. During the coiling process of hot-rolled coils, due to inappropriate tensions of the coiling belts, inaccurate equipment installation accuracy, and other reasons, there are quality problems such as flat coils, tower-shaped coils, loose coils, and poor coiling on the end faces of hot-rolled steel coils. Since the occurrence rate of tower-shaped defects in the steel coils produced by the hot-rolling production line is relatively high in the coil shape defects, and the temperature of the just-produced steel coils reaches 700 - 800 degrees, the current measurement of the tower height of hot-rolled coils mainly relies on manual visual inspection or measurement using contact-type measuring devices, which cannot ensure the safety of quality inspection personnel and the durability of the measuring devices. Therefore, the present invention proposes a non-contact method for measuring the tower height of the end face of a hot-rolled coil based on 3D point cloud. Summary of the Invention

[0003] In order to solve the technical problems proposed in the background art, the present invention provides a method and system for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud, which can solve the problem that quality inspection personnel cannot enter the site to accurately measure the tower coil height when the temperature of the hot-rolled coil is too high.

[0004] To achieve the above object, the present invention is implemented by adopting the following technical solutions:

[0005] A method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud includes the following steps:

[0006] Step 1, obtaining a point cloud image of the end face of the steel coil;

[0007] Step 2, segmenting the point cloud image region, and using the point cloud coordinate values to segment the image into two regions, where the bottom image of the tower-shaped end face of the steel coil and the top image of the tower shape are respectively in different regions;

[0008] Step 3, respectively segmenting the point cloud images of the bottom and top of the tower-shaped end face of the steel coil by using the Euclidean distance between each point cloud coordinate, calculating the number of point clouds in the segmented multiple regions and arranging them in ascending order, and initially positioning the bottom image region and the top image region of the tower shape;

[0009] Step 4, setting the tower height measurement error value "ε", calculating the mean value of the z coordinate values of all the point clouds in the initially obtained bottom image region and top image region of the tower shape, intercepting the set region near the mean value, and accurately positioning the bottom region and the top region of the tower shape;

[0010] Step 5, calculating the tower height.

[0011] Further, in the step 1, obtaining the point cloud image of the steel coil end face specifically includes: according to the size of the steel coil and the required image resolution, configuring one or more laser profilometers, using a single laser profilometer to scan multiple times or using multiple laser profilometers to scan and take pictures simultaneously multiple times, and after preprocessing the collected images such as point cloud stitching and point cloud denoising, obtaining a complete point cloud image of the steel coil end face.

[0012] Before collecting the image, by adjusting the pose of the laser profilometer, ensure that the laser line of the laser profilometer is perpendicular to the end face of the measured steel coil.

[0013] Further, in the step 2, the point cloud image region segmentation specifically includes: calculating the maximum value Max, the minimum value Min, and the average value Average of the maximum and minimum values of the z coordinate values of the tower-shaped point cloud image of the steel coil end face, satisfying:

[0014] Average = (Max + Min) / 2

[0015] Divide the image into two regions, respectively set as region P and region Q, where region P is in the range of z coordinate values [Average, Max], and region Q is in the range of z coordinate values [Min, Average]; obviously, the bottom and top images of the tower shape on the steel coil end face are respectively in region P and region Q, or region Q and region P. Different cameras will result in different regions where the bottom and top images of the tower shape on the steel coil end face are located.

[0016] Further, in the step 3, the preliminary positioning of the bottom and top regions of the tower shape specifically includes:

[0017] Set a positive number δ, satisfying that the value of δ is not greater than the x-axis resolution of the laser profilometer, calculate the Euclidean distance between each point cloud coordinate in region P and region Q obtained in step 2. If the Euclidean distance between two point cloud coordinates is less than the positive number δ, then these two point clouds are in the same region, otherwise they are not in the same region;

[0018] Calculate the number of point clouds in each region obtained from region P and region Q respectively and arrange them in ascending order. Starting from the region with the largest number of point clouds, visualize the point cloud images of each region, and continuously adjust the value of the positive number δ to preliminarily obtain region P1 of the bottom image of the tower shape on the steel coil end face in region P and region Q1 of the top image of the tower shape, or region P1 of the top image of the tower shape on the steel coil end face in region P and region Q1 of the bottom image of the tower shape.

[0019] Further, in the step 4, the precise positioning of the bottom and top regions of the tower shape specifically includes:

[0020] Set the tower height measurement error value "ε", calculate the average value of the z - coordinate values of all the point clouds in the bottom image area of the coil end face tower obtained initially in step 3, and denote it as A; similarly, calculate the average value of the z - coordinate values of all the point clouds in the bottom image area of the coil end face tower obtained initially, and denote it as B; intercept the area where the z - coordinate value of the point cloud is in the range of [A - ε / 4, A + ε / 4] from the bottom image area of the coil end face tower obtained initially, and let this area be the accurately positioned tower top area; similarly, intercept the area where the z - coordinate value of the point cloud is in the range of [B - ε / 4, B + ε / 4] from the bottom image area of the coil end face tower obtained initially, and let this area be the accurately positioned tower bottom area.

[0021] Further, in step 5, the specific calculation of the tower height is as follows: Calculate the average value of the z - coordinate values of the point cloud data in the accurately positioned tower bottom area and the accurately positioned tower top area respectively in step 4. The absolute value of the difference between the two average values is the measured value of the tower height, and this value is within the given error range.

[0022] The present invention also provides a hot - rolled coil end face tower height measurement system based on 3D point cloud. The system includes a laser profiler and a processor;

[0023] The laser profiler is used for image acquisition. According to the size of the coil and the required image resolution, one or more laser profilers are configured. Single - laser - profiler multiple scans or multiple - laser - profiler simultaneous scans are used for multiple shootings;

[0024] The processor splices the collected images through point cloud to obtain a complete point cloud image of the coil end face; calculates the tower height of the coil end face by running the foregoing method.

[0025] The present invention also provides a processor of a computer. The processor runs the foregoing method and finally calculates the tower height of the coil end face.

[0026] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0027] 1. The present invention provides a hot - rolled coil end face tower height measurement method and system based on 3D point cloud, which can solve the problem that quality inspection personnel cannot enter the site to accurately measure the tower height of the coil when the temperature of the hot - rolled coil is too high.

[0028] 2. The present invention is applicable to the complex production environment of steel mills, with convenient equipment installation, simple work processes, and easy - to - implement algorithms;

[0029] 3. The present invention realizes non - contact measurement of the tower height of the coil end face in real - time under the online production conditions of hot - rolled coils; 4. The method of the present invention ensures that the measurement result is within the given error range. Description of the Drawings

[0030] Figure 1 It is the implementation flowchart of measuring the tower height of the end face of the hot-rolled steel coil of the present invention;

[0031] Figure 2 It is the side structure schematic diagram of the installation position of the laser profiler and the end face of the steel coil in the embodiment of the present invention;

[0032] Figure 3 It is the point cloud diagram of the simulated object of the end face of the steel coil in the embodiment of the present invention;

[0033] Figure 4 It is the point cloud diagram of the area where the bottom of the tower shape of the end face of the steel coil is located after the point cloud image is segmented in the embodiment of the present invention;

[0034] Figure 5 It is the point cloud diagram of the area where the bottom of the tower top of the end face of the steel coil is located after the point cloud image is segmented in the embodiment of the present invention;

[0035] Figure 6 It is the point cloud diagram of accurately positioning the bottom area of the tower shape in the embodiment of the present invention;

[0036] Figure 7 It is the point cloud diagram of accurately positioning the top area of the tower shape in the embodiment of the present invention;

[0037] Figure 8 It is the result display diagram of the tower height of the simulated object of the end face of the steel coil in the embodiment of the present invention;

[0038] Figure 9 It is a system diagram of measuring the tower height of the end face of the hot-rolled steel coil based on 3D point cloud of the present invention.

[0039] In the figure: 1 - steel coil, 11 - bottom of the tower shape, 12 - top of the tower shape, 13 - core of the steel coil, 2 - laser profiler, 3 - processor. Specific implementation manners

[0040] The following provides a detailed description of the specific implementation manners provided by the present invention in conjunction with the accompanying drawings.

[0041] As Figure 1-2 shown, a method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud includes the following steps:

[0042] Step 1, obtain the point cloud image of the end face of the steel coil 1;

[0043] Step 2, segment the point cloud image area, divide the image into two areas, respectively set as area P and area Q, the images of the bottom 11 and the top 12 of the tower shape of the end face of the steel coil 1 are respectively in area P and area Q, or area Q and area P;

[0044] Step 3: Use the Euclidean distance between each point cloud coordinate to segment areas P and Q respectively, calculate the number of point clouds in the segmented multiple regions, and sort them in ascending order to initially locate the image region of the bottom 11 of the tower shape and the image region of the top 12 of the tower shape;

[0045] Step 4: Set the tower shape height measurement error value "ε", calculate the mean value of the z coordinate values of all point clouds in the initially obtained image region of the bottom 11 of the tower shape and the image region of the top 12 of the tower shape, intercept the set region near the mean value, and accurately locate the region of the bottom 11 of the tower shape and the region of the top 12 of the tower shape;

[0046] Step 5: Calculate the tower shape height, ensuring that the calculation result is within the given error range.

[0047] Specifically, in the above-mentioned step 1, obtaining the end face point cloud image of the steel coil 1 is specifically as follows: According to the size of the steel coil 1 and the required image resolution, configure one or more laser profilometers 2. Use a single laser profilometer 2 to scan multiple times or use multiple laser profilometers 2 to scan and take pictures simultaneously. After preprocessing such as point cloud stitching and point cloud denoising on the collected images, a complete end face point cloud image of the steel coil 1 is obtained.

[0048] Before collecting the image, by adjusting the pose of the laser profilometer 2, ensure that the laser line of the laser profilometer 2 is perpendicular to the end face of the measured steel coil 1.

[0049] As Figure 2 shown, in the embodiment of the present invention, 1 laser profilometer 2 is used. Before collection, by adjusting the pose of the laser profilometer 2, ensure that the laser line of the laser profilometer 2 is perpendicular to the end face of the measured steel coil 1. The attributes of the measured steel coil 1 include the bottom 11 of the tower shape, the top 12 of the tower shape, and the steel coil core 13. The obtained end face point cloud image of the simulated object steel coil 1 is as Figure 3 shown.

[0050] Specifically, the segmentation of the point cloud image region in the above-mentioned step 2 is as follows: Calculate the maximum value Max, the minimum value Min of the z coordinate values of the tower shape point cloud image on the end face of the steel coil 1, and the average value Average of the maximum value and the minimum value, satisfying:

[0051] Average = (Max + Min) / 2

[0052] The image is divided into two regions, namely region P and region Q. Region P is within the range of the z - coordinate value [Average, Max], and region Q is within the range of the z - coordinate value [Min, Average]. Obviously, the images of the bottom 11 and the top 12 of the end - face tower of the steel coil 1 are respectively in region P and region Q, or region Q and region P. Different cameras will result in different regions where the images of the bottom 11 and the top 12 of the end - face tower of the steel coil 1 are located. For the convenience of description, it is now assumed that the images of the bottom 11 and the top 12 of the end - face tower of the steel coil 1 are respectively in region P and region Q. Region P is as Figure 4 , and region Q is as Figure 5 shown.

[0053] Specifically, step 3 for obtaining the regions of the bottom 12 and the top 11 of the tower is as follows:

[0054] S30: Since the x - axis resolution of the laser profiler 2 used in this implementation scheme is 0.035 mm, a positive number δ equal to 0.03 mm is taken here. Calculate the Euclidean distance between each point - cloud coordinate in regions P and Q obtained in step S2. If the Euclidean distance between two point - cloud coordinates is less than the positive number δ, then these two point - clouds are in the same region; otherwise, they are not in the same region. Specifically, if the coordinate of point - cloud a in the point - cloud image is (a1, a2, a3) and the coordinate of point - cloud b is (b1, b2, b3), the Euclidean distance between them is as follows:

[0055]

[0056] If ρ < δ, then point - cloud a and point - cloud b are in the same region; otherwise, they are not in the same region. Here, using the operator in the machine - vision software halcon, referring to the connection_object_model_3d operator, select Feature as 'distance_3d', Value as δ, and the return value is a handle containing multiple different regions.

[0057] S31: Calculate the number of point - clouds in each region for the multiple segmented regions obtained in regions P and Q respectively. Refer to the get_object_model_3d_params operator in halcon, select the parameter GenParamName as 'num_points', and the return value is an array containing the number of point - clouds in each region;

[0058] S32: Sort in ascending order, that is, sort the array obtained in step S31 in ascending order;

[0059] S33: Starting from the area with the largest number of point clouds, visualize the point cloud images of each area, continuously adjust the value of the positive number δ, and initially obtain the image P1 of the bottom 11 of the tower shape of the end face of the steel coil 1 in area P and the image Q1 of the top 12 of the tower shape in area Q. Use the select_object_model_3d operator in halcon to select each area, and use the Visualize_object_model_3d operator to visualize the selected area.

[0060] Specifically, for step 4 to accurately locate the bottom 11 and top 12 areas of the tower shape, the specific steps are as follows:

[0061] S40: Since the laser line is perpendicular to the end face of the steel coil 1, calculate the average value of the z coordinate values of all the point clouds in the image P1 of the bottom 11 of the tower shape on the end face of the steel coil 1 obtained in step S33 above, and denote it as A; similarly, calculate the average value of the z coordinate values of all the point clouds in the image Q1 of the top 12 of the tower shape on the end face of the steel coil 1, and denote it as B. Refer to the get_object_model_3d_params operator in halcon, select the parameter GenParamName as 'point_coord_z', and the return value is an array P_array containing the z coordinate values of all the point clouds in the image P1 of the bottom 11 of the tower shape on the end face of the steel coil 1; A := mean(P_array) is the average value of the array. Obtain the average value B in the same way.

[0062] S41: Set the tower shape height measurement error value "ε", intercept the area where the z coordinate value of the point cloud is in the range of [A - ε / 4, A + ε / 4] from area P1, and mark this area as area P2, as Figure 6 shown. Refer to the region selection operator select_points_object_model_3d in halcon, set the parameter attribute Attrib to 'point_coord_z', set MinValue to A - ε / 4, set MaxValue to A + ε / 4, and the return value is area P2; in the same way, intercept the Q2 area where the z coordinate value of the point cloud is in the range of [B - ε / 4, B + ε / 4] from area Q1, as Figure 7 shown.

[0063] Specifically, for step 5 to calculate the tower shape height: Calculate the average value of the z coordinate values of each point cloud in area P2 again for area P2 and area Q2 obtained in step S41 above, denote it as P2_mean, and calculate the average value of the z coordinate values of each point cloud in area Q2, denote it as Q2_mean. The tower shape height value H is:

[0064] H = |P2_mean - Q2_mean|

[0065] Obviously, the tower shape height value H is within the given error ε range, as Figure 8 shown.

[0066] As Figure 9 shown, the present invention further provides a hot-rolled steel coil end face tower height measurement system based on 3D point cloud. The system includes a laser profiler 2 and a processor 3.

[0067] The laser profiler 2 is used for image acquisition. According to the size of the steel coil 1 and the required image resolution, one or more laser profilers 2 are configured, and one laser profiler 2 is used for multiple scans or multiple laser profilers 2 are used for simultaneous scans and multiple shots are taken.

[0068] The processor 3 splices the collected images through point cloud to obtain a complete end face point cloud image of the steel coil 1; the tower height of the end face of the steel coil 1 is calculated by running the foregoing method.

[0069] The present invention further provides a processor 3 of a computer. The processor 3 runs the foregoing method and finally calculates the tower height of the end face of the steel coil 1.

[0070] The above embodiments are implemented on the premise of the technical solution of the present invention, and detailed implementation manners and specific operation processes are given. However, the protection scope of the present invention is not limited to the above embodiments. The methods used in the above embodiments are all conventional methods unless otherwise specified.

Claims

1. A method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud, characterized in that, The steps include: Step 1: Obtain a point cloud image of the steel coil end surface; Step 2: Segment the point cloud image into two regions using the point cloud coordinate values. The images of the steel coil end face tower bottom and tower top are located in different regions. Step 3: Use the Euclidean distance between the coordinates of each point cloud to segment the point cloud images of the tower bottom and tower top of the steel coil end face, calculate the number of point clouds in the segmented regions and arrange them in ascending order, and preliminarily locate the tower bottom image region and the tower top image region; Step 4: Set the measurement error value of the tower height Calculate the mean value of the z - coordinate values of all the point clouds in the initially obtained tower bottom image area and tower top image area, intercept the set area near the mean value, and accurately locate the tower bottom area and tower top area; Step 5: Calculate the tower height.

2. The method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud according to claim 1, wherein In the step 1, the point cloud image of the end face of the steel coil is obtained as follows: according to the size of the steel coil and the required image resolution, a single or multiple laser profilers are configured, a single laser profiler is used to scan multiple times or multiple laser profilers are used to scan multiple times simultaneously, and the collected images are subjected to point cloud stitching and point cloud denoising preprocessing to obtain a complete point cloud image of the end face of the steel coil.

3. A method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud according to claim 2, characterized in that, Before collecting images, the position of the laser profiler is adjusted to ensure that the laser line of the laser profiler is perpendicular to the end face of the steel coil being measured.

4. The method for measuring the tower height of a hot-rolled steel coil end face based on 3D point cloud according to claim 1, characterized in that: In step 2, the point cloud image region segmentation is specifically as follows: calculating the maximum value Max, the minimum value Min, and the average value Average of the maximum and minimum values of the z coordinate value of the steel coil end pyramid point cloud image, which satisfies: Average=(Max+Min) / 2 The image is divided into two areas, set as P area and Q area respectively, where the z coordinate value of P area is in the range of [Average,Max], and the z coordinate value of Q area is in the range of [Min,Average]. Obviously, the images of the tower bottom and tower top of the steel coil end face are in the P area and Q area, or the Q area and P area respectively. Different cameras will lead to different areas where the images of the tower bottom and tower top of the steel coil end face are located.

5. A method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud according to claim 4, characterized in that In step 3, the preliminary positioning of the tower bottom and tower top areas is specifically as follows: Set a positive number δ, which is no greater than the x-axis resolution of the laser profiler, and calculate the Euclidean distance between the coordinates of each point cloud in the P area and the Q area obtained in step 2. If the Euclidean distance between the coordinates of two point clouds is less than the positive number δ, then the two point clouds are in the same area, otherwise they are not in the same area. The number of point clouds in each of the multiple segmented areas obtained in the P area and the Q area is calculated and arranged in ascending order. Starting from the area with the largest number of point clouds, the point cloud image of each area is visualized, and the value of the positive number δ is continuously adjusted to preliminarily obtain the pyramidal bottom image P1 area of the steel coil end face in the P area and the pyramidal top image Q1 area of the Q area, or the pyramidal top image P1 area of the steel coil end face in the P area and the pyramidal bottom image Q1 area of the Q area.

6. The method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud according to claim 1, wherein, In step 4, the precise positioning of the tower bottom and tower top areas is specifically as follows: Set the tower height measurement error value ", calculate the average value of the z - coordinate values of all point clouds in the tower - shaped bottom image area of the steel coil end face initially obtained in step 3, and denote it as A; similarly, calculate the average value of the z - coordinate values of all point clouds in the tower - shaped bottom image area of the steel coil end face initially obtained, and denote it as B; intercept the area where the z - coordinate value of the point cloud is in [A - / 4, A + / 4] from the tower - shaped bottom image area of the steel coil end face initially obtained, and let this area be the accurately positioned tower - shaped top area; similarly, intercept the area where the z - coordinate value of the point cloud is in [B - / 4, B + / 4] from the tower - shaped bottom image area of the steel coil end face initially obtained, and let this area be the accurately positioned tower - shaped bottom area.

7. A method for measuring the tower height of the end face of a hot-rolled steel coil based on 3D point cloud according to claim 1, characterized in that In step 5, the tower height is calculated specifically by calculating the mean of the z coordinate values of the point cloud data of the tower bottom area and the tower top area accurately located in step 4, respectively. The absolute value of the difference between the two means is the tower height measurement value, and the absolute value is within a given error range.

8. A hot-rolled steel coil end face tower height measurement system based on 3D point cloud, characterized in that, Includes a laser profiler and processor; The described laser profiler is used for image acquisition. According to the size of the steel coil and the required image resolution, one or multiple laser profilers are configured. Single laser profiler is used for multiple scans or multiple laser profilers are used for simultaneous multiple scans and multiple shootings. The described processor splices the acquired images through point cloud to obtain a complete point cloud image of the end face of the steel coil; the tower height of the end face of the steel coil is calculated by running the method described in any one of claims 1-7.

9. A processor of a computer, characterized in that, The described processor runs the method described in any one of claims 1-7, and finally calculates the tower height of the end face of the steel coil.

Images