CN117531834A - Visual positioning-based steel coil head coordinate conversion method - Google Patents
Visual positioning-based steel coil head coordinate conversion method Download PDFInfo
- Publication number
- CN117531834A CN117531834A CN202311045199.5A CN202311045199A CN117531834A CN 117531834 A CN117531834 A CN 117531834A CN 202311045199 A CN202311045199 A CN 202311045199A CN 117531834 A CN117531834 A CN 117531834A
- Authority
- CN
- China
- Prior art keywords
- image
- steel coil
- head
- camera
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 135
- 239000010959 steel Substances 0.000 title claims abstract description 135
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000006243 chemical reaction Methods 0.000 title claims abstract description 31
- 230000000007 visual effect Effects 0.000 title claims abstract description 23
- 238000010008 shearing Methods 0.000 claims abstract description 43
- 238000012545 processing Methods 0.000 claims abstract description 21
- 238000004804 winding Methods 0.000 claims abstract description 3
- 230000011218 segmentation Effects 0.000 claims description 21
- 238000005530 etching Methods 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 12
- 238000001514 detection method Methods 0.000 claims description 10
- 230000004807 localization Effects 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 5
- 238000005286 illumination Methods 0.000 claims description 4
- 230000001131 transforming effect Effects 0.000 claims description 4
- 230000002146 bilateral effect Effects 0.000 claims description 3
- 238000003709 image segmentation Methods 0.000 claims description 3
- 210000001503 joint Anatomy 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 238000011426 transformation method Methods 0.000 claims 2
- 238000004519 manufacturing process Methods 0.000 abstract description 6
- 230000008569 process Effects 0.000 abstract description 3
- 238000005520 cutting process Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 239000000498 cooling water Substances 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007730 finishing process Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000012856 packing Methods 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 229910000859 α-Fe Inorganic materials 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B15/00—Arrangements for performing additional metal-working operations specially combined with or arranged in, or specially adapted for use in connection with, metal-rolling mills
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B15/00—Arrangements for performing additional metal-working operations specially combined with or arranged in, or specially adapted for use in connection with, metal-rolling mills
- B21B15/0007—Cutting or shearing the product
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B38/00—Methods or devices for measuring, detecting or monitoring specially adapted for metal-rolling mills, e.g. position detection, inspection of the product
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B15/00—Arrangements for performing additional metal-working operations specially combined with or arranged in, or specially adapted for use in connection with, metal-rolling mills
- B21B15/0007—Cutting or shearing the product
- B21B2015/0014—Cutting or shearing the product transversely to the rolling direction
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B15/00—Arrangements for performing additional metal-working operations specially combined with or arranged in, or specially adapted for use in connection with, metal-rolling mills
- B21B2015/0057—Coiling the rolled product
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a steel coil head coordinate conversion method based on visual positioning, and relates to the technical field of steel coil treatment processes. The device comprises a barrel expansion suitable for winding a steel coil, wherein a shearing mechanical arm is arranged on one side of the barrel expansion, the device further comprises an image processing unit, a plurality of image acquisition units facing the barrel expansion are arranged above the barrel expansion along the circumference, and the adjacent two image acquisition units have at least 20% of vision overlap ratio; the image processing unit receives the image information of the image acquisition unit and recognizes and positions the head coordinates of the conversion steel coil so as to enable the shearing mechanical arm to conduct shearing action. By the scheme of the invention, automatic production can be realized for steel ring production, and the working efficiency is improved.
Description
Technical Field
The invention relates to the technical field of steel coil treatment processes, in particular to a steel coil head coordinate conversion method based on visual positioning.
Background
The production of the steel coil requires the procedures of heating, rolling, wire laying, cooling and the like. In the rolling process, if the time of feeding the finished raw materials into the pinch roller is not matched with the opening time of the cooling water tank, the front end part of the produced steel coil is not subjected to water penetration treatment, so that ferrite grains in the front end of the steel coil grow, the tissue components are uneven, and the performance indexes such as tensile strength and yield strength can not meet the use standard in use. Therefore, in the finishing step before shipment, the coil ends need to be sheared and recovered according to factory production regulations.
At present, most steel rabbets manually cut and recycle the heads of the steel coils for a plurality of circles. However, the high-temperature and high-heat severe production environment of the steel enterprise is unfavorable for long-time work of workers, so that the working efficiency is low, the labor cost is increased, and potential safety hazards are generated. Therefore, the steel coil front end automatic cutting out is realized by developing the steel coil front end intelligent cutting out technology, the production efficiency is improved, and the potential safety hazard is reduced. In the finishing process of the steel coil, the steel coil is required to be packaged, and then sheared and recovered. In the process of packing the steel coil, the condition that the front end of the steel coil is inserted into the steel coil often occurs, and the front end of the steel coil cannot be automatically sheared by the mechanical arm. Meanwhile, considering the limitation of the working space of the mechanical arm in the automatic shearing, two or more mechanical arms are generally required to be arranged to finish the automatic shearing operation, the cost is too high, and the application of the automatic shearing technology at the front end of the steel coil is limited.
Disclosure of Invention
The invention discloses a steel coil head coordinate conversion method based on visual positioning, which is simple in structure and aims to solve the problems that the existing steel coil shearing device is limited greatly and has high cost.
The invention adopts the following scheme:
the application provides a steel coil head coordinate conversion method based on visual positioning, which comprises a barrel expanding suitable for winding a steel coil, wherein a shearing mechanical arm is arranged on one side of the barrel expanding, the method further comprises an image processing unit, a plurality of image acquisition units facing the barrel expanding are arranged above the barrel expanding along the circumference, and the adjacent two image acquisition units have at least 20% of vision overlap ratio; the image processing unit receives the image information of the image acquisition unit and is adapted to perform the steps of:
s1, identifying head coordinate information of a positioning image steel coil;
s2, judging whether the head of the steel coil is positioned in the working space of the shearing mechanical arm; if yes, coordinate conversion and positioning are carried out; if not, firstly splicing the images, finishing splicing the image pixels in the overlapping area of the adjacent images acquired by the adjacent image acquisition units, performing preprocessing fitting after finishing the butt joint, fitting the single coil where the head of the steel coil is located, and performing coordinate conversion and positioning after finishing the fitting;
and S3, moving the shearing mechanical arm to the coordinates of the head of the steel coil positioned in the S2 or the coordinates of the single coil where the head of the steel coil is located, and shearing the single coil where the head of the steel coil is located.
Further, in S1, the step of identifying and positioning the head coordinate information of the image steel coil is as follows:
a1, transforming the size, and transforming the size of the picture into a specified size after obtaining the picture;
a2, carrying out bilateral filtering on the picture with the transformed size to remove picture noise;
a3, self-adaptive histogram equalization, namely, redistributing the brightness value of the image according to the histogram by calculating the local histogram in the steel coil image, and changing the contrast of the image so as to correct uneven illumination in the image and strengthen local details;
a4, graying the image, and carrying out graying treatment on the image by using a component method to obtain a gray image;
a5, detecting and obtaining an angular point through a FAST algorithm, wherein the angular point is the head of the steel coil; and determining the optimal segmentation threshold value of the foreground and the background of the steel coil in the gray image by using an Ojin method, and taking the difference value between the gray value of the steel coil and the optimal segmentation threshold value as the threshold value of a FAST algorithm.
Further, the algorithm execution steps of the Ojin method are as follows:
a51, calculating a gray image histogram, counting the number of pixel points on 256 pixel values of 0-255, and normalizing;
a52, setting a segmentation threshold value i, and counting the proportion w of pixel points with gray values in the range of 0-i in the image to the whole image according to the segmentation threshold value 0 And an average gray value u 0 And the proportion w of the pixel points in the whole image with gray values in the range of i-255 1 And an average gray value u 1 At the same time, the total average gray level u of the image is calculated according to the following formula 2 And inter-class variance g:
u 2 ＝w 0 *u 0 +w 1 *u 1 ，
g＝w 0 (u 0 -u 2 )2+w 1 (u 1 -u 2 )2；
and A53, increasing a gray value by the segmentation threshold i, and continuing to execute the previous step until the last gray value, and outputting the optimal threshold of the image segmentation foreground and background by taking i corresponding to the maximum inter-class variance g.
Further, in A5, the coil diameter information is converted into a radius of a detection circle in a FAST corner detection algorithm through mapping.
Further, the preprocessing fit comprises the following steps:
b1, intercepting an image, namely expanding n pixel units leftwards and rightwards in width according to pixel coordinates of the head of the steel coil, and intercepting the image in height according to the height of the whole spliced image;
b2, graying, namely graying the intercepted image;
b3, threshold processing, namely obtaining an optimal segmentation threshold value of the steel coil and the background by adopting an OTSU algorithm, and binarizing the intercepted image by using the threshold value to obtain a black-and-white image;
etching and expanding, namely etching and then expanding the black-and-white image according to a kernel function of 3 multiplied by 1, removing discrete interference points in the black-and-white image by etching, and reconnecting the areas with broken etching by expanding;
b5, skeleton algorithm processing, namely continuously corroding the boundary of the black-and-white image by adopting a skeleton algorithm until only a single pixel structure of the black-and-white image is left, so as to obtain the real skeleton of the image;
removing outliers, namely calculating the median of the abscissa of all points by taking each pixel point on the skeleton in the skeleton image as a data point, simultaneously calculating the deviation between each data and the median, and removing the interference outliers in the skeleton image by taking the median of the deviation of 3 times as a standard;
b7, kmeans clustering, namely performing clustering sampling on skeleton pixel points in the skeleton image;
and B8, polynomial fitting, namely performing polynomial fitting on a plurality of clustering points sampled by the B7 to obtain a single steel coil fitting result.
Further, the conversion formula of the three-dimensional coordinates is:
wherein A is an internal reference matrix of a camera obtained by camera calibration, R is a conversion rotation matrix between a world coordinate system established on the expansion barrel and a camera coordinate system which can be calculated according to the Rodrigues variation, T is a translation matrix converted between the world coordinate system established on the expansion barrel and the camera coordinate system which can be calculated according to the Rodrigues variation, zc is a shearing position obtained by a 3D camera under the camera coordinate system, u and v are coordinates of the shearing position, and R, T is a conversion matrix between the world coordinate system and the camera coordinate system.
Further, the image acquisition device is a 3D camera, the 3D cameras are provided with 5, and are circumferentially distributed around the central axis of the expansion barrel.
The beneficial effects are that:
the method starts from the actual engineering requirement, the characteristics of the front end head of the steel coil are analyzed by utilizing a visual positioning method, and according to the spatial arrangement of cameras and pixel coordinate information of the head of the steel coil, coordinate conversion calculation is directly carried out on the condition that the shearing position of the head of the steel coil is in the working space of the mechanical arm, so that the spatial coordinate of the optimal shearing position of the head of the steel coil is obtained; and under the condition that the shearing position of the steel wire head is not in the working space of the mechanical arm, three-dimensional space coordinate conversion of the steel wire head is realized through image splicing, image pixel processing and fitting, and the space coordinate of the optimal shearing position of the steel wire head in the working space of the mechanical arm is obtained.
Drawings
Fig. 1 is a schematic flow chart of a method for converting coordinates of a head of a steel coil based on visual positioning according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of preprocessing fitting of a method for converting coordinates of a head of a steel coil based on visual localization according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a positional relationship between an image acquisition unit and a barrel expansion in a method for converting coordinates of a head of a steel coil based on visual positioning according to an embodiment of the present invention;
fig. 4 is a schematic diagram showing an arrangement of an image acquisition unit on a barrel according to a method for converting coordinates of a head of a steel coil based on visual localization according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a world coordinate system of a barrel expansion in a method for converting coordinates of a head of a steel coil based on visual positioning according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a mounting position of a mechanical arm in a method for converting coordinates of a head of a steel coil based on visual positioning according to an embodiment of the present invention;
fig. 7 is a flow chart of a head recognition method in a method for converting coordinates of a head of a steel coil based on visual localization according to an embodiment of the present invention;
icon: an image acquisition unit 1, a first camera 11, a second camera 12, a third camera 13, a fourth camera 14, a fifth camera 15, a barrel 2 and a cutting mechanical arm 3.
Detailed Description
With reference to fig. 1, this embodiment provides a method for converting coordinates of a head of a steel coil based on visual positioning, which includes a barrel 2 adapted to wind the steel coil, a shearing mechanical arm 3 is disposed on one side of the barrel 2, and an image processing unit is further included, a plurality of image acquisition units 1 facing the barrel 2 are disposed above the barrel 2 along a circumference, and at least 20% of field of vision overlap exists between two adjacent image acquisition units 1; the image processing unit receives the image information of the image acquisition unit 1 and is adapted to perform the steps of:
s1, identifying head coordinate information of a positioning image steel coil;
s2, judging whether the head of the steel coil is positioned in the working space of the shearing mechanical arm 3; if yes, coordinate conversion and positioning are carried out; if not, firstly splicing the images, finishing splicing the image pixels in the overlapping area of the adjacent images acquired by the adjacent image acquisition unit 1, performing preprocessing fitting after finishing the butt joint, fitting the single coil where the head of the steel coil is located, and performing coordinate conversion and positioning after finishing the fitting;
and S3, the shearing mechanical arm 3 moves to the coordinates of the head of the steel coil positioned in the S2 or the coordinates of the single coil where the head of the steel coil is located, and shears the single coil where the head of the steel coil is located.
As shown in fig. 3 to 6, in this embodiment, the image processing unit is an existing image processing system or processing software, the image capturing device is a 3D camera or an industrial color camera, the 3D camera or the industrial color camera is provided with 5 cameras and is circumferentially distributed around the central axis of the expansion barrel 2, specifically, the cameras include a first camera 11, a second camera 12, a third camera 13, a fourth camera 14 and a fifth camera 15, and the cameras are installed while the installation coordinates of each camera are already strictly calculated, so that the field of view of each camera is fixed, there is at least 20% overlapping area between two adjacent cameras, the image captured by the camera is a fixed part of the expansion barrel 2, and the pixel positions of the same part in the image are fixed. In one embodiment, the cutting mechanical arm 3 is mounted on one side of the fourth camera 14 and the fifth camera 15, when the steel coil head is in the view range of the fourth camera 14 and the fifth camera 15, the cutting mechanical arm 3 can directly cut the steel coil head by moving, and when the steel coil head is in the view range of the first camera 11, the second camera 12 and the third camera 13, the cutting mechanical arm 3 cannot move to the position where the steel coil head is located, and at this time, the single steel wire where the steel coil head is located in the view range of the fourth camera 14 and the fifth camera 15 can be cut to meet the requirement by identifying the single steel wire where the steel coil head is located.
In this embodiment, as shown in fig. 7, first, the head of the steel coil needs to be identified, that is, in S1, the steel coil may be identified by a FAST algorithm, or the method of detecting and identifying the head of the steel coil may be applied, which is disclosed as CN114820585a, by a deep learning method or by the head of the steel coil. The method for identifying the FAST algorithm is provided, and the basic principle of the FAST algorithm is as follows: when the number of the pixels in the adjacent area of a certain pixel point and the pixels in different areas of the point is enough, the point is considered as a corner point. Specifically, the definition in this embodiment is as follows: and comparing the pixel gray values of the P point serving as the center with the pixel gray values of 16 pixel points on the neighborhood with the radius of 3 pixels, and if the difference value between the pixel gray values of more than 75% of the pixel points and the pixel gray values of the P point is larger than a set threshold value, considering the point as a corner point.
In this embodiment, since the steel coil head belongs to a corner point with a larger size in the image, a suitable radius of circumference of the pixel point compared with the P point and a threshold value need to be defined, otherwise, a problem of inaccurate identification occurs. Specifically, for the radius of the circumference of the pixel point, according to the principle of small-hole imaging, the distance between the steel coil and the image acquisition unit 1 is unchanged, so that the number of pixels occupied by the diameter of the steel coil in the picture does not change greatly. Therefore, the steel coil diameter information can be converted into the radius of the detection circle in the FAST corner detection algorithm through mapping. The head of the steel coil with the measured diameter of 8mm takes up 18 pixels in the gray level image, and the diameter of the steel coil produced by factories has 8 specifications in total of 6-20 mm, and the corresponding pixels in the gray level image can be obtained by conversion according to the proportional relation, specifically: according to the principle of pinhole imaging, after the camera is calibrated, the size of the diameter of the steel coil head in the gray image and the size of the actual steel coil head are approximately in direct proportion, that is, the size of the steel coil diameter in the gray image can be estimated according to the actual steel coil size and calibrated parameters, namely: Φ1/Φ2=δ×d1/d2, where d1, d2 represent the diameter or radius of the circumference of the steel coil in the gray image, Φ1, Φ2 are the diameter of the actual steel coil head, d1, d2 represent the radius of the circumference of the coil in the gray image, and the parameter δ is a correction parameter, determined according to the installation position of the camera and the camera parameters; if d2 corresponding to phi 2 is required, phi 2 is substituted into the formula. The diameter of the steel coil is known before algorithm detection, and the circumference radius of the pixel point compared with the P point in the FAST algorithm can be obtained according to the diameter of the steel coil. For example, a steel coil having a diameter of 8mm should have a radius of 10 pixels for the pixel point to be compared with the P point, and the number of points on the circumference to be compared with the P point should be increased to 56.
In this embodiment, since the FAST algorithm threshold is set according to the picture condition, it is proposed to determine the optimal segmentation threshold of the foreground and the background of the picture steel coil by using the oxford method (OTSU), and use the difference between the gray value of the steel coil and the optimal segmentation threshold as the FAST algorithm threshold, so that the threshold can be automatically input according to the image information, without manual input, and the input efficiency is improved. The algorithm execution steps of the discipline method are as follows:
a51, calculating a gray image histogram, counting the number of pixel points on 256 pixel values of 0-255, and normalizing;
a52, setting a segmentation threshold value i, and counting the proportion w of pixel points with gray values in the range of 0-i in the image to the whole image according to the segmentation threshold value 0 And an average gray value u 0 And the proportion w of the pixel points in the whole image with gray values in the range of i-255 1 And an average gray value u 1 At the same time, the total average gray level u of the image is calculated according to the following formula 2 And inter-class variance g:
u 2 ＝w 0 *u 0 +w 1 *u 1 ，
g＝w 0 (u 0 -u 2 )2+w 1 (u 1 -u 2 )2；
and A53, increasing a gray value of the segmentation threshold i, continuing to execute the previous step until the last gray value, taking i corresponding to the maximum inter-class variance g as the optimal segmentation threshold output of the image segmentation foreground and background, and inputting the difference value between the gray value of the steel coil and the optimal segmentation threshold as the threshold of the FAST algorithm. The gray pixel value of the steel coil in the gray image is in the range of 150-210. Note that since an image employing this algorithm is susceptible to noise, it is necessary to perform noise reduction processing, that is, image gradation processing, on the image.
In the method, the size of the shot picture is larger, so that the detection speed can be influenced, and in S1, the size of the picture can be changed into a smaller size, the detection algorithm is quickened while the detection effect is not influenced, and meanwhile, the difference between the pixel numbers occupied by the steel coils with the same diameter can be ensured to be as small as possible. In addition, since the picture shot on site contains various noises, if the noise in the picture is removed by using the conventional gaussian filtering, the edge of the steel coil in the picture becomes blurred, so in the embodiment, the processing is performed by adopting a bilateral filtering mode in the S3, and the sharpness and smoothness of the edge information of the steel coil can be maintained. Meanwhile, as the change of illumination in the picture affects the result of the FAST algorithm, in S4, local histograms in the steel coil image are calculated, the brightness values of the image are redistributed according to the histograms, and the contrast of the image is changed, so that uneven illumination in the image is corrected, and local details are enhanced. In S5, the gray level processing is carried out on the image by using a component method to obtain a gray level image, and further, the highlighting has better stability when threshold calculation is carried out by a discipline method in the FAST algorithm. After being detected by the FAST algorithm, the double-sided filtering is finally carried out again, so that the boundary is clearer.
After head identification, according to the information of the head of the steel coil, the head needs to be judged, when the head is just located in the working range of the shearing mechanical arm 3, the head is directly sheared, when the head is not located in the working range of the shearing mechanical arm 3, the part, extending to the working range of the shearing mechanical arm 3, of the single steel wire of the steel coil where the head is located needs to be identified, and then the part is sheared. Specifically, when the head is not in the working range of the shearing mechanical arm 3, the images acquired by the plurality of cameras are spliced, for example, when the steel coil head is positioned on the right side of the fourth camera 14, the fifth camera 15 or the third camera 13, the steel coil head is in the working space of the shearing mechanical arm 3, and the three-dimensional coordinates of the head can be obtained directly through coordinate conversion; when the head of the steel coil is positioned at the left side of the first camera 11, the second camera 12 or the third camera 13, firstly, the images are spliced according to the position of the head of the steel coil, and if the head is in the view of the first camera 11, the images of the first camera 11 and the fifth camera 15 are spliced; if the head is in the field of view of the second camera 12, the second camera 12 and the third camera 13 are spliced, after the splicing is completed, the single coil image of the head is preprocessed and fitted according to the pixel coordinate information of the head, and then coordinate conversion positioning is carried out, so that the three-dimensional coordinates of the shearing position of the shearing mechanical arm 3 are obtained.
As shown in connection with fig. 2, the preprocessing fit described herein includes the steps of:
b1, intercepting an image, namely expanding n pixel units leftwards and rightwards in width according to pixel coordinates of the head of the steel coil, and intercepting the image in height according to the height of the whole spliced image; where n may be 25.
B2, graying, namely graying the intercepted image;
b3, threshold processing, namely obtaining an optimal segmentation threshold value of the steel coil and the background by adopting an OTSU algorithm, and binarizing the intercepted image by using the threshold value to obtain a black-and-white image; the OTSU algorithm is the oxford method in head recognition to obtain the optimal segmentation threshold.
Etching and expanding, namely etching and then expanding the black-and-white image according to a kernel function of 3 multiplied by 1, removing discrete interference points in the black-and-white image by etching, and reconnecting the areas with broken etching by expanding;
b5, skeleton algorithm processing, namely continuously corroding the boundary of the black-and-white image by adopting a skeleton algorithm until only a single pixel structure of the black-and-white image is left, so as to obtain the real skeleton of the image;
removing outliers, namely calculating the median of the abscissa of all points by taking each pixel point on the skeleton in the skeleton image as a data point, simultaneously calculating the deviation between each data and the median, and removing the interference outliers in the skeleton image by taking the median of the deviation of 3 times as a standard; thus solving the problem that the intercepted image contains a target coil skeleton and other parts of interference coil skeletons;
b7, kmeans clustering, namely clustering and sampling skeleton pixel points in the skeleton image by adopting a kmeans clustering algorithm; because the skeleton image still contains more data points after the outliers are removed, fitting the data points can lead to longer fitting time, and a small number of interference points can lead to deviation of fitting results, the skeleton pixel points can be clustered and sampled in the skeleton image through a kmeans clustering algorithm, and fitting is carried out in the next step;
and B8, polynomial fitting, namely performing polynomial fitting on a plurality of clustering points sampled by the B7 to obtain a single steel coil fitting result. The fitting expression is x= 1254.8446-7.4359 ×10 (-2) y+2.4656 ×10 (-5) y-2, where x and y represent the coordinates of a single coil where the coil head is located in the working space of the mechanical arm, and the pixel coordinates (u, v) of the shearing position of the shearing mechanical arm 3 can be obtained.
After coordinate fitting, the coordinate needs to be converted into a three-dimensional coordinate, and the conversion formula of the three-dimensional coordinate is as follows through a world coordinate system established on a barrel expansion 2 mechanism:
wherein X is W 、Y W 、Z W For the three-dimensional coordinates of the shearing position, a is an internal reference matrix of the camera obtained by calibrating the camera, R is a conversion rotation matrix between the world coordinate system established on the expansion bucket 2 and the camera coordinate system which can be calculated according to the rogows, T is a translation matrix converted between the world coordinate system established on the expansion bucket 2 and the camera coordinate system which can be calculated according to the rogows, zc is a shearing position obtained by the 3D camera under the camera coordinate system, u, v are coordinates of the shearing position, and R, T is a conversion matrix between the world coordinate system and the camera coordinate system.
By the method, the shearing position of the head part of the steel coil can be positioned in the working space of the mechanical arm, and when the shearing position of the head part of the steel coil is not in the working space of the mechanical arm, the shearing position of the head part of the steel coil can be positioned by image splicing and fitting after preprocessing a single coil. And a world coordinate system is established on the barrel expansion 2 mechanism, and three-dimensional space positioning of the shearing position is realized according to the conversion relation between the camera coordinate system and the world coordinate system and the depth information of the 3D camera, so that the coordinates under the world coordinate system are obtained, and automatic positioning and shearing are realized.
It should be understood that: the above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above examples, and all technical solutions belonging to the concept of the present invention belong to the protection scope of the present invention.
The description of the drawings in the embodiments above illustrates only certain embodiments of the invention and should not be taken as limiting the scope, since other related drawings may be made by those of ordinary skill in the art without the benefit of the inventive faculty.
Claims (7)
1. The steel coil head coordinate conversion method based on visual positioning comprises a barrel expansion suitable for winding a steel coil, wherein a shearing mechanical arm is arranged on one side of the barrel expansion, and the method is characterized by further comprising an image processing unit, a plurality of image acquisition units facing the barrel expansion are arranged above the barrel expansion along the circumference, and the adjacent two image acquisition units have at least 20% of visual field overlap ratio; the image processing unit receives the image information of the image acquisition unit and is adapted to perform the steps of:
s1, identifying head coordinate information of a positioning image steel coil;
s2, judging whether the head of the steel coil is positioned in the working space of the shearing mechanical arm; if yes, coordinate conversion and positioning are carried out; if not, firstly splicing the images, finishing splicing the image pixels in the overlapping area of the adjacent images acquired by the adjacent image acquisition units, performing preprocessing fitting after finishing the butt joint, fitting the single coil where the head of the steel coil is located, and performing coordinate conversion and positioning after finishing the fitting;
and S3, moving the shearing mechanical arm to the coordinates of the head of the steel coil positioned in the S2 or the coordinates of the single coil where the head of the steel coil is located, and shearing the single coil where the head of the steel coil is located.
2. The visual localization-based steel coil head coordinate transformation method as set forth in claim 1, wherein in S1, the step of identifying localization image steel coil head coordinate information is as follows:
a1, transforming the size, and transforming the size of the picture into a specified size after obtaining the picture;
a2, carrying out bilateral filtering on the picture with the transformed size to remove picture noise;
a3, self-adaptive histogram equalization, namely, redistributing the brightness value of the image according to the histogram by calculating the local histogram in the steel coil image, and changing the contrast of the image so as to correct uneven illumination in the image and strengthen local details;
a4, graying the image, and carrying out graying treatment on the image by using a component method to obtain a gray image;
a5, detecting and obtaining an angular point through a FAST algorithm, wherein the angular point is the head of the steel coil; and determining the optimal segmentation threshold value of the foreground and the background of the steel coil in the gray image by using an Ojin method, and taking the difference value between the gray value of the steel coil and the optimal segmentation threshold value as the threshold value of a FAST algorithm.
3. The method for converting the coordinates of the head of the steel coil based on visual localization according to claim 2, wherein the algorithm of the oxford method is performed as follows:
a51, calculating a gray image histogram, counting the number of pixel points on 256 pixel values of 0-255, and normalizing;
a52, setting a segmentation threshold value i, and counting the proportion w of pixel points with gray values in the range of 0-i in the image to the whole image according to the segmentation threshold value 0 And an average gray value u 0 And the proportion w of the pixel points in the whole image with gray values in the range of i-255 1 And an average gray value u 1 At the same time, the total average gray level u of the image is calculated according to the following formula 2 And inter-class variance g:
u 2 ＝w 0 *u 0 +w 1 *u 1 ，
g＝w 0 (u 0 -u 2 )2+w 1 (u 1 -u 2 )2；
and A53, increasing a gray value by the segmentation threshold i, and continuing to execute the previous step until the last gray value, and outputting the optimal threshold of the image segmentation foreground and background by taking i corresponding to the maximum inter-class variance g.
4. A method of converting a head coordinate of a steel coil based on visual localization according to claim 3, wherein in A5 the steel coil diameter information is converted into a radius of a detection circle in a FAST corner detection algorithm by mapping.
5. The visual localization-based steel coil head coordinate transformation method of claim 1, wherein the preprocessing fit comprises the steps of:
b1, intercepting an image, namely expanding n pixel units leftwards and rightwards in width according to pixel coordinates of the head of the steel coil, and intercepting the image in height according to the height of the whole spliced image;
b2, graying, namely graying the intercepted image;
b3, threshold processing, obtaining the optimal segmentation threshold value of the steel coil and the background by adopting an OTSU algorithm,
binarizing the intercepted image by using the threshold value to obtain a black-and-white image;
etching and expanding, namely etching and then expanding the black-and-white image according to a kernel function of 3 multiplied by 1, removing discrete interference points in the black-and-white image by etching, and reconnecting the areas with broken etching by expanding;
b5, skeleton algorithm processing, namely continuously corroding the boundary of the black-and-white image by adopting a skeleton algorithm until only a single pixel structure of the black-and-white image is left, so as to obtain the real skeleton of the image;
removing outliers, namely calculating the median of the abscissa of all points by taking each pixel point on the skeleton in the skeleton image as a data point, simultaneously calculating the deviation between each data and the median, and removing the interference outliers in the skeleton image by taking the median of the deviation of 3 times as a standard;
b7, kmeans clustering, namely performing clustering sampling on skeleton pixel points in the skeleton image;
and B8, polynomial fitting, namely performing polynomial fitting on a plurality of clustering points sampled by the B7 to obtain a single steel coil fitting result.
6. The visual positioning-based steel coil head coordinate conversion method according to claim 1, wherein the conversion of three-dimensional coordinates is shown as:
wherein A is an internal reference matrix of a camera obtained by camera calibration, R is a conversion rotation matrix between a world coordinate system established on the expansion barrel and a camera coordinate system which can be calculated according to the Rodrigues variation, T is a translation matrix converted between the world coordinate system established on the expansion barrel and the camera coordinate system which can be calculated according to the Rodrigues variation, zc is a shearing position obtained by a 3D camera under the camera coordinate system, u and v are coordinates of the shearing position, and R, T is a conversion matrix between the world coordinate system and the camera coordinate system.
7. The visual positioning-based steel coil head coordinate conversion method according to claim 1, wherein the image acquisition device is a 3D camera, the 3D cameras are provided with 5, and are circumferentially distributed around the central axis of the barrel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311045199.5A CN117531834A (en) | 2023-08-18 | 2023-08-18 | Visual positioning-based steel coil head coordinate conversion method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311045199.5A CN117531834A (en) | 2023-08-18 | 2023-08-18 | Visual positioning-based steel coil head coordinate conversion method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117531834A true CN117531834A (en) | 2024-02-09 |
Family
ID=89790650
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311045199.5A Pending CN117531834A (en) | 2023-08-18 | 2023-08-18 | Visual positioning-based steel coil head coordinate conversion method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117531834A (en) |
-
2023
- 2023-08-18 CN CN202311045199.5A patent/CN117531834A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111047615B (en) | Image-based straight line detection method and device and electronic equipment | |
| JP2008158958A (en) | Road surface determination method and road surface determination device | |
| EP3534333A1 (en) | Method for calibrating the position and orientation of a camera relative to a calibration pattern | |
| CN111308987B (en) | Automatic uncoiling control system of uncoiler based on image processing and detection method | |
| CN109886958B (en) | Method, device and system for automatically identifying welding seam position based on square tube | |
| CN115631116B (en) | Aircraft power inspection system based on binocular vision | |
| CN106355597A (en) | Monocular vision based image processing method for automatic measuring robot for steel plate folding angle | |
| CN112361990A (en) | Laser pattern extraction method and device, laser measurement equipment and system | |
| CN114241438B (en) | Traffic signal lamp rapid and accurate identification method based on priori information | |
| CN117531834A (en) | Visual positioning-based steel coil head coordinate conversion method | |
| CN110310239B (en) | Image processing method for eliminating illumination influence based on characteristic value fitting | |
| CN117058182A (en) | Target detection frame optimization method for visual ranging | |
| CN116843761A (en) | Bolt size measuring method and device, electronic equipment and storage medium | |
| CN114842335B (en) | Grooving target identification method and system for construction robot | |
| JP3039241B2 (en) | Feature point extraction method for ring-shaped work | |
| CN113469980B (en) | Flange identification method based on image processing | |
| CN115830049A (en) | Corner point detection method and device | |
| CN110060270B (en) | Edge detection method suitable for polygonal tubular object with low imaging quality | |
| CN112767425A (en) | Parking space detection method and device based on vision | |
| CN109272462B (en) | Image self-correction method, device and storage device suitable for metal workpiece identification | |
| CN110046528B (en) | Dot-shaped DataMatrix two-dimensional code identification method | |
| CN117115102A (en) | Machine vision-based steel coil head identification method | |
| JPH1010051A (en) | Applying condition detecting device using vision sensor | |
| US20240020846A1 (en) | Corner point detection method and apparatus | |
| EP4276755A1 (en) | Image segmentation method and apparatus, computer device, and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |