JP4896828B2 - Shape detection method and shape detection apparatus - Google Patents

Description

  The present invention relates to a shape detection method, and more particularly to a shape detection method used for detecting a shape profile of a rolled material or the like. The present invention also relates to a shape detection apparatus that detects a shape profile of a rolled material by extracting a light cutting line by the extraction method.

  Conventionally, a process for processing a material such as a metal material into a strip-shaped rolled material has been widely performed. For example, metal materials such as steel, aluminum, and copper are processed into a thick plate and a thin plate through a rolling process. At this time, there is a shape profile as one of the indexes that influence the quality of the strip-shaped material before and after passing through the rolling process (hereinafter simply referred to as a rolled material). For example, if the shape profile of the rolled material is abnormal, that is, if the rolled material is greatly waved or has uneven unevenness at the center or end thereof, it will hinder subsequent processes such as rolling, conveying, and processing. And may cause quality degradation such as scratches, distortion, and abnormal thickness on the product surface. Therefore, it is required to reduce the shape abnormality of the rolled material by measuring such shape profile abnormality and feeding back the measurement information to the rolling process.

Such a rolled material is usually conveyed and processed on a line along the long side direction. Therefore, the long side direction is referred to as “feed direction”, and the short side direction perpendicular thereto is referred to as “width direction”. As described above, the belt is conveyed along the feeding direction. As a method for measuring the shape of the belt at a high speed in a non-contact manner, a light cutting method is used.
For example, in Patent Document 1, a plate-shaped light projecting unit and a special camera for image capturing are installed above a belt-like body. The plate-shaped light projecting unit irradiates the plate-shaped light beam in the width direction of the belt-shaped body, and the special camera captures a light cutting line that is an image formed by reflecting the plate-shaped light beam on the surface of the belt-shaped body. The shape profile of the rolled material is detected from the positional relationship between the plate-like light source and the camera and the position of the light cutting line on the image.

If the plate-like light projecting unit and the special camera are relatively stationary with respect to the strip, the profile of the strip surface can be known only with respect to the coordinates on the light cutting line. However, when the plate-like light projecting unit and the special camera are stationary and the belt-like body is conveyed in the direction of the arrow, the effect of scanning the surface of the belt-like body with the light cutting line can be obtained. As a result, the three-dimensional shape of the rolled material along the feed direction can be measured.
One of the important technical points in the surface shape detection method using the light cutting method is to accurately extract the position of the light cutting line from the image of the light cutting line. In Patent Document 1, the following method is used. First, a laser is used as a plate-like light source, and an optical filter is installed in front of the camera lens to image only the reflected light of the light source laser. A clear light section line image is obtained by cutting light components other than the light section line image, and the imaging accuracy is improved. Subsequently, the image is divided into light and dark by binarization processing, and it is determined that a series of bright points are light cutting lines, and the coordinates of the light cutting line position on the image are extracted.

In Patent Document 2, similarly to Patent Document 1, laser slit light is used, and a disturbance light component is removed by combining an optical filter with a camera lens. In the extraction of the light section line, first, the light section line is scanned in a direction crossing at a right angle to obtain the maximum luminance point (peak point), and the approximate curve is obtained using data in the vicinity of the peak point. The continuation of the peak positions of these approximate curves is the position of the optical cutting line.
Patent Document 3 also shows a method for measuring the shape of a belt-like body by capturing an optical cutting line with a special camera using a fan-shaped beam. Here, in particular, a method is shown in which the shape of the band can be measured even when the height of the band is changed due to the vibration of the production line.
Japanese Patent No. 2913903 JP-A-7-324915 JP 2006-189315 A

  In any of the above-mentioned patent documents, a plate-like or fan-like beam is irradiated onto a belt-like body, and the reflected beam is imaged with a camera to form an optical cutting line. Usually, such a plate-like or fan-shaped beam is generated by expanding spot light irradiated from a laser point light source only in one direction with a cylindrical lens or the like. However, the light from the laser point light source that is usually available has a non-uniform light quantity distribution in the cross section of the light beam, so that the brightness of the beam irradiated to the belt is different between the central portion and the side edge portion. In addition, since ambient light such as illumination light includes light of the same wavelength as the irradiated beam, even if a filter is installed on the camera, an image with a mixture of light cutting lines and ambient ambient light can be displayed. I will take an image.

Furthermore, since the reflectance of light locally changes due to adhesion of scale, cooling water, or the like to the surface of the rolled material, the brightness of the light section line varies or decreases depending on the location. This makes it impossible to stably obtain a light cutting line in the width direction on the image. In particular, in a portion where the luminance of the light cutting line is low, the luminance may be reduced to a level equivalent to the luminance of the disturbance light component, so that the light cutting line is covered with the disturbance light component, which is the conventional technique described in the patent literature. With this method, it is not possible to extract the light section line stably.
In order to obtain an accurate shape of the belt-like body, there is a problem that the light section line must be detected as accurately as possible from such an image. The present invention has been made in view of the above problems, and an object thereof is to provide a method for accurately detecting a light section line.

In order to achieve the above object, the present invention has taken the following measures. That is, the technical means for solving the problems in the present invention is a shape detection method for extracting a light cutting line from reflected light of slit light projected on a rolled material and detecting a shape profile of the rolled material by a light cutting method. Then, an image including reflected light of the slit light projected on the rolled material is captured, a luminance distribution in the captured image is detected, and a predetermined region including a portion having the highest luminance in the image is detected. is set as the first search small areas a search small areas to limit the scope of searching the optical cutting line starts searching for the light section lines (called search start region) within the image, included in the search small areas The coordinates of the light cutting line to be extracted are extracted, the average value and the variance value of the coordinates in the search subregion from which the light cutting line is extracted are obtained, and the search subregion is adjacent based on the obtained average value and dispersion value. Next search subregion Constant and extracts the coordinates of the light section lines included in said next search small areas, and detects the shape profile of the rolled material from all of the light section line coordinate extracted.

In here, depending on the variance value of the the search light section line in the small area coordinates may change the size of the next search small areas.
Further, the position of the next search subregion may be changed according to the average value of the coordinates of the light cutting line in the search subregion.
Here, the extraction of the light section line is performed by dividing the search subregion into a plurality of search lines, creating a brightness pattern for each of the plurality of search lines, and setting the peak position of the brightness pattern as the position of the light section line. It may be performed by judging.

Here, the light cutting line is extracted by dividing the search subregion into a plurality of search lines, creating a luminance pattern for each of the plurality of search lines, and setting the barycentric position of the luminance pattern as the position of the light cutting line. It may be performed by judging.
Further, the technical means for solving the problems in the present invention is a shape detection device for measuring a shape profile of a rolled material by a light cutting method, and an irradiation unit for irradiating the rolled material with slit light and a projection on the rolled material. An imaging unit that captures an image including reflected light of the slit light, and a luminance distribution in the image captured by the imaging unit, and a predetermined region including a portion having the highest luminance in the image, A search subregion that limits the search range of the light section line in the image and is set as the first search subregion (referred to as a search start region) for starting the search for the light section line, and the light included in the search subregion The coordinates of the cutting line are extracted, the average value and the variance value of the coordinates in the search subregion from which the optical cutting line is extracted are obtained, and the next adjacent to the search subregion based on the obtained average value and variance value Exploration And an image recognition unit that extracts the coordinates of the light cutting line included in the next search sub-region and detects the shape profile of the rolled material from the coordinates of all the extracted light cutting lines. Is.

Here, the image recognition unit detects a luminance distribution in the captured image, and sets a search small region including a portion having the highest luminance in the image as a search start region for starting a search for an optical section line. Also good.
Here, the image recognition unit may change the size of the next search subregion according to the variance value of the coordinates of the light section line in the search subregion.
Further, the image recognition unit may change the position of the next search subregion according to the average value of the coordinates of the light section line in the search subregion.

Here, the image recognition unit divides the search subregion into a plurality of search lines, creates a luminance pattern for each of the plurality of search lines, and determines that the peak position of the luminance pattern is the position of the light cutting line. May be.
Here, the image recognition unit divides the search sub-region into a plurality of search lines, creates a luminance pattern for each of the plurality of search lines, and determines that the barycentric position of the luminance pattern is the position of the light cutting line. May be.

  According to the present invention having such a feature, since the light cutting line extraction process can be started from an area where the brightness of the light cutting line is high and can be extracted clearly, and the result can be reflected in the surrounding area one after another. Even in a region where the brightness of the cutting line is low, it is possible to accurately extract the light cutting line without being affected by the disturbance light component.

The shape detection method and shape detection apparatus according to the present invention will be described below. Here, application to a hot rolling line will be described as an example.
FIG. 1 shows a hot rolling line and discloses a device configuration from a rolling mill 1 to a shape detection device 2, a cooling device 3, and a winding device 4. In addition, in the transfer direction of a rolling material, the side (winding device 4 side) which is transferred is called a downstream side, and the opposite side (rolling machine 1 side) is called an upstream side.
The rolling mill 1 is for rolling the rolled material 5 and is disposed on the most upstream side. The pair of work rolls 6 and 6, the pair of backup rolls 7 and 7 for backing up the work rolls 6 and 6, and the rolling mill 1 and a rolling control device 8 for controlling the rolling speed of the rolled material 5. The cooling device 3 cools the rolled material 5 and includes a plurality of continuous cooling banks 9 and 9, and discharges cooling water from cooling nozzles 9 a provided in the respective cooling banks 9 and 9. Thus, the rolled material 5 is cooled. The winding device 4 winds the rolled material 5 cooled by the cooling device 3.

As shown in FIGS. 1 and 2, the shape detection device 2 measures the shape profile of the rolled material 5 rolled by the rolling mill 1, and with respect to the shape profile obtained by the measurement, ear waves, medium elongation, The characteristics of the shape profile of the rolled material 5 can be detected by analyzing how much flatness patterns such as composite elongation and quarter elongation are included.
The shape detection device 2 is disposed between the rolling mill 1 and the cooling device 3, and an irradiation unit 10 that irradiates slit light, an imaging unit 11 that images reflected light from the irradiated slit light, and shape detection control. Device 12.

  The irradiating unit 10 generates slit light by passing point laser light through a cylindrical lens, and irradiates the slit light toward the surface so as to face the width direction of the rolled material 5. The imaging unit 11 captures reflected light reflected from the surface of the rolled material 5 when irradiated with slit light, and is composed of, for example, a two-dimensional CCD camera. In addition, the lens of the imaging unit 11 is provided with a filter 13 that transmits only the wavelength of the slit light from the irradiation unit 10. The captured image is transmitted to the shape detection control device 12. The angle (holding angle) formed by the laser projection axis of the irradiation unit 10 and the optical axis of the imaging unit 11 is set in a range of approximately 30 ° to 150 °.

The shape detection control device 12 analyzes the image transmitted from the imaging unit 11 and detects the shape profile of the rolled material 5 from the image or controls the imaging unit 11. The shape detection control device 12 controls the imaging unit 11. 14, an analysis computer 15 that analyzes an image, and a monitor 16 that displays a captured image or the like.
The control computer 14 includes an imaging control unit 17 that adjusts the luminance of the imaging unit 11 and controls lens aperture and the like, and an image storage unit 18 that stores the captured image. The image storage unit 18 is composed of, for example, a frame memory, and stores a captured image with 640 × 480 pixels.

  The analysis computer 15 includes an image recognition unit 19 and an analysis unit 20. The image recognition unit 19 extracts a line that is a light cutting line from the reflected light (slit light) in the image stored in the image storage unit 18, and coordinates the line in two dimensions (x-axis, y-axis). Store as image coordinate data. The x axis in the image coordinate data corresponds to the width direction of the rolled material 5, and the y axis in the image coordinate data corresponds to the transfer direction of the rolled material 5. Further, the image recognizing unit 19 obtains a shape profile using the principle of triangulation method based on the image coordinate data, normalizes the obtained shape profile, and normalizes the two-dimensional (x axis, z axis) coordinate data. Remember as.

The horizontal axis (x-axis) in the normalized coordinate data stored in the image recognition unit 19 corresponds to the width direction of the rolled material 5, and the vertical axis (z-axis) in the normalized coordinate data is the rolling It corresponds to the thickness direction of the material 5. In this way, a normalized shape profile as shown in FIG. 3 is obtained.
A method in which the image recognition unit 19 and the analysis unit 20 extract the light section line from the image stored in the image storage unit 18 in the process of detecting the shape profile as described above will be described in detail below.

FIG. 4 is a diagram specifically showing the configuration of the rolled material 5, the irradiation unit 10, the imaging unit 11, and the shape detection control device 12 in FIG. 1. The imaging unit 11 captures an image of a portion indicated by oblique lines in the rolled material 5 and sends the image to the shape detection control device 12. In addition, the arrow of FIG. 4 shows the feed direction of the rolling material 5. FIG.
FIG. 5 is a diagram illustrating an image 21 a captured by the imaging unit 11. As described above, the image 21a includes 640 pixels in the x-axis direction and 480 pixels in the y-axis direction, and coordinates are assigned to each pixel. The light cutting line 22 is reflected by the rolled material 5, and the light cutting lines 23 and 24 are reflected by the background of the rolled material 5. Since the optical cutting lines 23 and 24 appear at certain coordinate positions in the screen and are irrelevant to the shape profile of the rolled material 5, the image recognition unit 19 previously stores the positions of the optical cutting lines 23 and 24. The position of the rolled material 5 is detected and subtracted from the image 21a. By doing so, the image recognition unit 19 obtains an image 21b as shown in FIG. The obtained image 21b is divided into a plurality of search regions of 10 pixels in the x-axis direction and 480 pixels in the y-axis direction, and a search subregion for extracting a light section line is set in a part of each search region To do. Hereinafter, a series of processes for extracting the light section line will be described with reference to the flowchart of FIG.

FIG. 7 is a flowchart showing a processing procedure when the image recognition unit 19 detects the light section line 22. First, the image recognizing unit 19 sets 480 pixels having the same x coordinate in the image 21b and arranged in a straight line in the y-axis direction as one search line, and sets one pixel for every 640 search lines. Two luminances are detected (S1). The maximum luminance pixel which is the highest luminance pixel is detected for each search line, and the luminance patterns shown in FIG. 8 are created by arranging the maximum luminances of the detected 640 pixels (S2).
Next, a “search start area” which is a small search area and is searched first is set (S3). This search start area is set to a search area including a pixel having the highest luminance in the luminance pattern shown in FIG. 8 created in S2. The width of the search start area in the x-axis direction matches the width of the search area (x = 10), and the width in the y-axis direction is ± 5 in the y-axis direction with reference to the y coordinate of the pixel with the highest luminance. The pixel width (y = 10).

Only within the set search start region (the search region in the second and subsequent processes), the maximum luminance pixels in each of the ten search lines are detected, and the coordinates of the detected ten maximum luminance pixels are used as the light cutting line 22. The upper point is determined (S4).
An average value of the y coordinates of the ten maximum luminance pixels detected in S4 and a variance value of the y coordinates (hereinafter simply referred to as a variance value) are obtained (S5).
A search subregion (next search subregion) for obtaining the optical cutting line position next to a position adjacent to the search subregion (the search subregion) using the average value and the variance value of the y coordinate obtained in S5. Set (S6). The width of the adjacent search small area in the x-axis direction matches the width of the search area (x = 10). Further, the width in the y-axis direction is set to a width of ± (5 pixels + dispersion value) in the y-axis direction centered on the average value of the y-coordinate obtained in S5 and using the dispersion value obtained in S5. Is set.

Although the width in the y-axis direction in the search subregion set in S6 is ± (5 pixels + dispersion value), the range in the y-axis direction is set to ± by using an arbitrary coefficient multiplied by the dispersion value. It may be determined as (A pixel + α × dispersion value) or β × dispersion value (A, α, β are arbitrary values).
The following processing is executed for each of the 10 search lines only within the search small area set in S6. That is, one search line in the search small area is selected, and the pixel having the highest luminance on the line is detected (S7).
Next, it is determined whether or not the luminance of the detected pixel is greater than or equal to a threshold value (S8). If it is less than the threshold value in S8, it is determined that the end of the light cutting line 22 has been detected, and the series of processing ends. If the luminance is equal to or higher than the threshold, it is determined that the coordinates of the pixel detected in S7 are included in the light cutting line 22 (S9). It is determined whether or not detection of pixels on all search lines in the search subregion has been completed (S10). If the detection has not been completed, the process returns to S7 to detect the pixel having the highest luminance in the next search line, and if the detection has been completed, the process returns to S5 to average the y-coordinates of the 10 pixels detected in the small search area. Find the value and variance.

  In a series of processes for detecting the light section line 22, in S6, the size of the search small area is changed according to the dispersion value as shown in FIG. FIG. 9 shows a small search area indicated by a thick line in the image 21b of FIG. Usually, the light cutting line 22 has a lower luminance as it is reflected at the end portion of the rolled material 5, and at the same time, the dispersion value of the y coordinate becomes larger. From this, the size of the search subregion becomes larger as it reaches the end of the light section line 22 as shown in FIG. However, if the dispersion value does not change even at the end portion of the light section line 22, the size of the search small area does not change. In addition, when the variance value that increases as it reaches the end of the light section line 22 decreases toward the end, the size of the next search subregion becomes smaller than the size of the search region.

As shown in FIG. 10, in S6, the size of the next search subregion is changed by the variance value, and the average value of the coordinates of the light section line in the search subregion is set as the center position of the next search subregion. Adopted to determine the range of the next search subregion.
In S7, the pixel with the highest luminance is detected, and in S9, the coordinates are points included in the light cutting line 22. In this regard, in S7 or S9, a luminance pattern of the search line as shown in FIG. 11 is created, and the coordinates of the barycentric point B of the luminance pattern are included in the light cutting line 22 instead of the coordinates of the point A having the maximum luminance. It can also be.

  A method for obtaining the y coordinate of the barycentric point B will be described with reference to FIG. First, it is assumed that the luminance pattern shown in FIG. 12 is obtained in S7 or S9. Here, let the luminance value at the point of the y coordinate value Yi be Ii. If the starting point i = s and the ending point of the luminance pattern are i = e, the barycentric position Yg of the luminance pattern is obtained by the following equation.

Here, the search line for extracting the luminance pattern is regarded as one thin beam, the y coordinate is assigned to the position of the beam, and the luminance is regarded as the density of the beam at that position. By doing this, the total weight of the beam is calculated as Σ (Ii), and the primary moment of the beam is calculated as Σ (Yi × Ii). Yg is calculated.
Thus, by obtaining the barycentric point of the luminance pattern, the position of the maximum luminance can be determined with high accuracy even when the position of the maximum luminance is shifted due to the noise of the CCD of the imaging unit 11. It is also possible to determine the position of the maximum luminance by obtaining an approximate curve near the maximum luminance.

According to the light cutting line extraction method and the shape detection device as described above, the search small area is set in the captured image and the light cutting line is extracted, so that the amount of data processing necessary for the light cutting line extraction is small. I'm sorry. Also, in the captured image, when the brightness of the light cutting line varies depending on the location due to adhesion of scales etc. on the surface of the rolled material or specular reflection by cooling water, or the light cutting line and disturbance light that is natural light are mixed However, since the size of the search subregion is variable depending on the luminance dispersion value, the light section line can be obtained stably.
In addition, this invention is not limited to said embodiment.

For example, although the hot rolling line has been exemplified in the embodiment, application to a cold rolling line and a wire rod rolling line, which are similar rolling processes, is also possible.
Moreover, if the shape detection apparatus concerning this invention is installed in the upstream of a rolling mill, the shape profile of the rolling material before entering into a rolling mill can be detected. Therefore, the rolling material referred to in the present embodiment is not limited to a thick plate or a thin plate after rolling, but also includes, for example, members such as a slab, a bloom, and a billet before rolling.

It is a schematic diagram of the rolling line provided with the shape detection apparatus. It is a block diagram of a shape detection apparatus. It is the figure which normalized the coordinate data. It is the figure which showed the structure of the rolling material, irradiation part, imaging part, and shape detection control apparatus in embodiment of this invention. It is a figure which shows the image which the imaging part in embodiment of this invention imaged It is a figure which shows the state which divided the image which the imaging part in embodiment of this invention imaged into the search area | region. It is a flowchart which shows the operation | movement of the optical section line detection in embodiment of this invention. It is a figure which shows the luminance pattern of the maximum luminance in embodiment of this invention. It is the figure which showed the search small area | region in the image imaged by the imaging part in embodiment of this invention. It is the figure which showed the search small area | region in the image imaged by the imaging part in embodiment of this invention. It is a figure which shows an example of the luminance pattern of the search line in embodiment of this invention. It is a figure which shows an example of a luminance pattern.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Rolling machine 2 Shape detection apparatus 3 Cooling apparatus 4 Winding apparatus 5 Rolled material 6 Work roll 7 Backup roll 8 Rolling control apparatus 9 Cooling bank 10 Irradiation part 11 Imaging part 12 Shape detection control apparatus 13 Filter 14 Control computer 15 For analysis Computer 16 Monitor 17 Imaging control unit 18 Image storage unit 19 Image recognition unit 20 Analysis unit 21a, 21b Image 22, 23, 24 Optical cutting line

Claims (10)

A shape detection method for extracting a light cutting line from reflected light of slit light projected on a rolled material, and detecting a shape profile of the rolled material by a light cutting method,
Take an image containing the reflected light of the slit light projected on the rolled material,
Detecting a luminance distribution in the captured image;
A predetermined area including the highest luminance portion in the image is a search small area that limits a range for searching for a light cutting line in the image and is a first search small area for starting a search for a light cutting line Set,
Extract the coordinates of the light section line included in the search subregion,
Obtain the average value and variance value of the coordinates in the search subregion from which the light section line has been extracted,
Set the next search subregion adjacent to the search subregion based on the obtained average value and variance,
Extract the coordinates of the light section line included in the next search subregion,
A shape detection method, comprising: detecting a shape profile of a rolled material from coordinates of all extracted light cutting lines. The shape detection method according to claim 1 , wherein the size of the next search subregion is changed in accordance with a variance value of the coordinates of the light cutting line in the search subregion. Wherein the search in accordance with the average value of the coordinates of the light section lines in a small area, the shape detecting method according to claim 1 or 2, characterized in that changing the position of the next search small areas. The extraction of the light section line is as follows:
Dividing the search subregion into a plurality of search lines;
Creating a luminance pattern for each of the plurality of search lines;
Shape detecting method according to any one of claims 1 to 3, characterized in that it is carried out by determining the peak position of the brightness pattern and the position of the light section lines. The extraction of the light section line is as follows:
Dividing the search subregion into a plurality of search lines;
Creating a luminance pattern for each of the plurality of search lines;
Shape detecting method according to any one of claims 1 to 3, characterized in that it is carried out by determining the center of gravity of the intensity pattern and the position of the light section lines. A shape detection device for measuring a shape profile of a rolled material by an optical cutting method,
An irradiation unit for irradiating the rolled material with slit light;
An imaging unit that captures an image including reflected light of slit light projected on the rolled material;
Detecting a luminance distribution in the image captured by the imaging unit, a predetermined region including the highest portion brightness in the image, the search small areas to limit the scope of searching the light section line in said image Set as the first search subregion to start the search for the light section line, extract the coordinates of the light section line included in the search subregion, and average the coordinates in the search subregion from which the light section line was extracted Find the value and variance, set the next search subregion adjacent to the search subregion based on the calculated average and variance, and extract the coordinates of the light section line included in the next search subregion And an image recognition unit for detecting a shape profile of the rolled material from the coordinates of all the extracted light cutting lines. The shape detection apparatus according to claim 6 , wherein the image recognition unit changes the size of the next search subregion according to a variance value of the coordinates of the light section line in the search subregion. The shape detection device according to claim 6 or 7 , wherein the image recognition unit changes a position of a next search subregion according to an average value of coordinates of light cutting lines in the search subregion. . The image recognition unit divides the search subregion into a plurality of search lines, creates a luminance pattern for each of the plurality of search lines, and determines that the peak position of the luminance pattern is the position of the light cutting line. The shape detection device according to any one of claims 6 to 8 , wherein: The image recognition unit divides the search subregion into a plurality of search lines, creates a luminance pattern for each of the plurality of search lines, and determines that the barycentric position of the luminance pattern is the position of the light cutting line. The shape detection device according to any one of claims 6 to 8 , wherein:

Images