JP7351282B2 - Method for superimposing captured images of steel materials and method for detecting markings on steel materials - Google Patents

Description

The present invention takes a steel material extending in the longitudinal direction as a subject, and superimposes two captured images in the longitudinal direction, which are obtained by capturing images using at least two imaging devices and have different and overlapping ranges along the longitudinal direction. The present invention relates to a method for superimposing captured images of steel materials to be combined and a method for detecting markings on steel materials.

Generally, a round steel material having a circular cross section (referred to as a round billet) is manufactured directly or by blooming rolling. Round steel materials often have some kind of flaws (casting flaws, holes, linear flaws, cut flaws, etc.) on their surfaces during the manufacturing process. These flaws become a hindrance in the subsequent process, so before sending the round steel material to the subsequent process, a so-called "care work" is performed to remove the surface flaws by grinding. In addition, long steel products such as round steel bars and product steel pipes may develop surface flaws during the manufacturing process, so maintenance work is performed on them.

As a conventional surface flaw care device for round steel materials, for example, one shown in Patent Document 1 is known.
The surface flaw cleaning device for a round steel material disclosed in Patent Document 1 includes a surface flaw inspection device that can freely move back and forth along the axial direction of a rotating round steel material and detects surface flaws by contacting the round steel material; The device is equipped with a marking device that sprays a marking liquid to mark the position of a surface flaw, and a work deck that cleans the marked surface flaw.
According to the surface flaw care device for round steel materials disclosed in Patent Document 1, the position of detected surface flaws is marked, so that the operator can accurately judge the grinding position and perform the grinding work quickly. can.

However, in the case of the surface flaw cleaning device for round steel materials disclosed in Patent Document 1, an operator visually finds a marking applied to the position of a detected surface flaw and grinds the spot.
Here, the size and shape of the marking applied to the position of the surface flaw on the round steel material may change depending on the spraying condition of the marking liquid, and it may be difficult for an operator to visually detect the marking. Therefore, in the method of visually detecting markings by an operator, there is a high risk of missing the markings, and there is a problem that the grinding efficiency is low.

In order to solve this problem, it is desirable to automatically detect markings.
In order to automatically detect markings, one imaging device is installed at a position where the surface of the round steel material is within the field of view (for example, directly above the round steel material), and the entire surface of the round steel material is detected by rotating the round steel material. Measure the circumference. Thereafter, the round steel material is moved in the longitudinal direction of the round steel material, and the entire circumference is measured again. Then, it is conceivable to measure the round steel material in the longitudinal direction by repeating the measurement of the entire circumference and the movement in the longitudinal direction. The reason for this is that the field of view in the longitudinal direction of the round steel material in one imaging device does not cover the entire range to be imaged in the longitudinal direction of the round steel material.

JP2002-28724A

However, if one imaging device is used to measure the longitudinal direction of a round steel material, as mentioned above, it is necessary to repeatedly measure the entire circumference of the round steel material and move it in the longitudinal direction. The problem is that it takes a lot of time.
On the other hand, a method can be considered in which a plurality of imaging devices are installed along the longitudinal direction of the steel material and images captured by each imaging device are combined through image processing.
However, if the diameter of each round steel material to be measured differs, the working distance (distance between the surface of the round steel material and the imaging device) changes, so in that case, it is difficult to use two imaging devices. However, there is no established method for overlapping two captured images obtained by capturing images of different and overlapping regions along the longitudinal direction.

Therefore, the present invention has been made in order to solve this conventional problem, and its purpose is to take a longitudinally extending steel material as a subject and to obtain longitudinal images obtained by imaging using at least two imaging devices. To provide a method for superimposing captured images of a steel material and a method for detecting markings on a steel material, which can appropriately superimpose two captured images having different ranges and overlapping ranges in the longitudinal direction according to a working distance. It is in.

In order to achieve the above object, a method for superimposing captured images of steel materials according to one aspect of the present invention uses as a subject a steel material extending in the longitudinal direction, and obtains images in the longitudinal direction obtained by capturing images using at least two imaging devices. A method for superimposing two captured images having different ranges and overlapping ranges along the longitudinal direction of a steel material, the method comprising A working distance, which is a distance between each object and the steel material serving as the subject, is calculated, and a superimposition position of the two captured images is set on each of the two captured images according to the calculated working distance, and the The gist is to overlap each of the two captured images at an overlapping position set on each of the two captured images.

Further, a marking detection method for a steel material according to another aspect of the present invention is a marking detection method for a steel material that detects a marking applied at a position of a flaw formed on a surface of a steel material extending in a longitudinal direction, the method comprising: The surface of the steel material is imaged using two imaging devices to obtain two captured images having different and overlapping ranges along the longitudinal direction, and the two The gist of this method is to superimpose two captured images obtained from two imaging devices in the longitudinal direction, and extract markings from the superimposed captured images of steel material.

According to the method for superimposing captured images of a steel material and the method for detecting markings on a steel material according to the present invention, a steel material extending in the longitudinal direction is taken as a subject, and images obtained by capturing images using at least two imaging devices, which differ along the longitudinal direction. It is possible to provide a method for superimposing captured images of a steel material and a method for detecting markings on a steel material, which can appropriately superimpose two captured images having a range and an overlapping range in the longitudinal direction according to a working distance.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a front view of a schematic configuration of a steel marking detection apparatus to which a method of superimposing captured images of steel and a method of detecting markings of steel according to an embodiment of the present invention is applied. FIG. 2 is a diagram showing a schematic configuration of the steel marking detection device shown in FIG. 1, viewed from the right side. 2 is a flowchart showing the procedure of a method for detecting markings on steel materials using the marking detection device for steel materials shown in FIG. 1. FIG. 4 is a flowchart showing a procedure for superimposing captured images in the flowchart shown in FIG. 3. FIG. An actual captured image (a) when a round steel material (steel material) is imaged using adjacent imaging devices as one, an image captured by each of the adjacent imaging devices (b), and a superimposition of images captured by adjacent imaging devices. A diagram for explaining a pattern 1 (c) with deviation, a pattern 2 (d) with a deviation in the superposition of images captured by adjacent imaging devices, and a pattern (e) with a successful superposition of images captured by adjacent imaging devices. It is. FIG. 4 is a diagram for explaining the principle of superimposition in the process of superimposing captured images in the flowchart shown in FIG. 3; It is an explanatory view for calculating working distance which is a distance between an imaging device and a round steel material (steel material). It is an explanatory view when the steel material used as the object of marking detection is a square billet.

Embodiments of the present invention will be described below with reference to the drawings. The embodiments shown below illustrate devices and methods for embodying the technical idea of the present invention. It is not limited to the embodiments described below. Furthermore, the drawings are schematic. Therefore, it should be noted that the relationships, ratios, etc. between thickness and planar dimensions are different from those in reality, and the drawings also include portions where the relationships and ratios of dimensions are different.

FIG. 1 shows a schematic configuration of a steel marking detection apparatus to which a method for superimposing captured images on steel and a method for detecting markings on steel according to an embodiment of the present invention is applied. A marking detection device 1 for a steel material shown in FIG. 1 is installed downstream of a marking device (not shown) that marks the position of a surface flaw on a round steel material (round billet) S as a steel material. The marking device is installed in the middle of the conveyance line that conveys the round steel material S to the downstream process, and the marking device applies markings to the positions of surface flaws on the round steel material S, and the round steel material S with the marking applied is transferred. (not shown) to the marking detection device 1. In this embodiment, the size of the round steel material S to be marked is any size between the minimum diameter φ90 mm and the maximum diameter φ450 mm, and in FIGS. 1 and 2, the round steel material S1 with the maximum diameter is , the round steel material with the smallest diameter is indicated by S2.

The marking detection device 1 detects the marking M (see FIG. 5(a)) applied to the position of the surface flaw on the round steel material S, and the marking detection device 1 detects the marking M (see FIG. 5(a)) applied to the position of the surface flaw of the round steel material S. It is provided with a plurality of turning rollers 2 that are supported so as to be maintained in the same direction and rotated at a predetermined rotational speed (in the present embodiment, for example, at about 1500 mm/s in the circumferential direction). A pulse generator 18 as a rotation angle detection device for detecting the rotation angle of the round steel material S is installed on the turning roller 2. The rotation speed of the turning roller 2 is inputted from the pulse generator 18 to the marking extraction section 9, which will be described later, and the marking extraction section 9 extracts the rotation angle from the imaging start point of the round steel material S from the input rotation speed of the turning roller 2. Detect. The turning roller 2 is installed on a truck (not shown), and by moving the truck in the longitudinal direction, the round steel material S placed on the turning roller 2 can be moved in the longitudinal direction.

The marking detection device 1 includes a plurality of (two in this embodiment) imaging devices 3a and 3b that capture images of the surface of the round steel material S rotating in the circumferential direction on the turning roller 2, and a computer system 7. , and a display device 10.
The round steel material S is a steel material called a round billet, which has a circular cross section extending in the longitudinal direction (axial direction). As shown in FIG. 2, the plurality of (two) imaging devices 3a, 3b are arranged at predetermined intervals along the longitudinal direction of the round steel material S.

In the marking detection device 1, as shown in FIGS. 1 and 2, a plurality of first support members 14 are attached to a plurality of support legs 13 erected on a pedestal 12 so as to be perpendicular to the support legs 13. It is being A second support member 15 is attached to these first support members 14 so as to be perpendicular to the first support members 14 . Furthermore, a plurality of third support members 16 are attached at positions above the portions of the plurality of support legs 13 to which the first support members 14 are attached so as to be orthogonal to the support legs 13 . Further, a fourth support member 17 is attached to these third support members 16 so as to be perpendicular to the third support members 16 .
The tip of the fourth support member 17 extends along the longitudinal direction of the round steel material S, as shown in FIG. It is supported by

Each of the adjacent imaging devices 3a and 3b is composed of a line sensor camera, and as shown in FIG. There is. Each imaging device 3a, 3b is configured such that the direction in which the imaging line of the line sensor camera constituting each imaging device 3a, 3b extends and the axial direction of the round steel material S match, and the optical axis L3 of the line sensor camera and the round steel material The longitudinal direction of S forms a right angle, and furthermore, as shown in FIG. will be installed. That is, the optical axis L3 of the line sensor camera is set in the vertical direction, and is set so that the optical axis L3 passes through the vertex of the round steel material S held horizontally in the longitudinal direction.

As shown in FIGS. 6 and 7, the installation height of the line sensor cameras constituting each imaging device 3a, 3b is such that the working distances WD1, WD2 between the lenses of each line sensor camera and the round steel material S are predetermined. distance (the working distance WD1 (Φmax) = WD2 (Φmax) in the case of the round steel material S1 with the maximum diameter: 900 mm, the working distance WD1 (Φmin) = WD2 (Φmin) in the case of the round steel material S2 with the minimum diameter: 1270 mm is set to
When selecting the line sensor cameras constituting each of the imaging devices 3a and 3b, the depth of field is calculated and A camera is selected that has a lens that can be focused even at the top position P2. In the case of this embodiment, the depth of field is set to 771 mm, and a lens is selected that can focus even when the diameter of the round steel material S1 with the largest diameter is Φ450 mm and the diameter of the round steel material S2 with the smallest diameter is Φ90 mm.

The line sensor camera constituting each imaging device 3a, 3b moves the uppermost position (specific position) P in the circumferential direction of the surface of the round steel material S rotating in the circumferential direction on the turning roller 2 at a predetermined period. Capture images for one round. Each imaging device 3 detects the uppermost position P of the surface of the round steel material S rotating in the circumferential direction (the uppermost position of the surface of the round steel material S1 with the largest diameter is P1, and the uppermost position of the surface of the round steel material S2 with the smallest diameter is P2). The position is imaged for one rotation of the round steel material S at a predetermined period. In this embodiment, the rotational speed of the round steel material S is 1500 mm/s, and the imaging cycle with the line sensor camera is 1/2381 s so that the entire circumferential surface of the round steel material S can be imaged without any gaps. There is.

Furthermore, the imaging ranges L3a, L3b (see FIG. 6) of the respective imaging devices 3a, 3b are determined by the diameter of the round steel material S, that is, the working distance, which is the distance between each imaging device 3a, 3b and the surface of the round steel material S. It changes depending on WD1 and WD2 (see FIGS. 6 and 7). As shown in FIG. 2, the imaging ranges of the imaging devices 3a and 3b in the case of the round steel material S1 with the maximum diameter are L3a (Φmax), L3b (Φmax), and the imaging ranges of the imaging devices 3a and 3b in the case of the round steel material S2 with the minimum diameter. The imaging ranges are L3a (Φmin) and L3b (Φmin) (only L3b (Φmin) is shown).
The distance H along the longitudinal direction between the two imaging devices 3a and 3b (distance between the center lines of the two imaging devices 3a and 3b) is such that the imaging ranges L3a and L3b of each imaging device 3a and 3b are made of round steel. Regardless of the diameter of the round steel material S in the longitudinal direction of S (both in the case of the round steel material S1 with the largest diameter and in the case of the round steel material S2 with the smallest diameter), the sizes are set so that they always overlap. Ru.

Here, the reason why the imaging devices 3a and 3b are line sensor cameras is because the surface of the round steel material S is circular. This is because the distance to the surface of the round steel material S varies along the circumferential direction, and the appearance of the shape of the marking applied to the surface of the round steel material S changes. If the imaging devices 3a and 3b are configured with line sensor cameras and image the position of the uppermost position P on the surface of the round steel material S rotating in the circumferential direction along the axial direction, the round steel material S can be captured from the line sensor camera. There is no change in the distance to the topmost position P on the surface of , and there is no such inconvenience. Therefore, by using the imaging devices 3a and 3b as line sensor cameras, it is possible to appropriately detect the shape of the marking applied to the surface of the round steel material S that is the object of imaging.
The line sensor cameras constituting each imaging device 3a, 3b are connected to a camera control device 4 (not shown) that controls a power source, an imaging cycle, and the like.

Further, the marking detection device 1 includes an illumination device 5 for each of the imaging devices 3a and 3b, as shown in FIGS. 1 and 2.
Each lighting device 5 is rotatably attached to the tip of the first support member 14 described above.
Each illumination device 5 is composed of two rows of bar illumination that continuously illuminate the surface of the round steel material S, particularly in the vicinity of the top position P to be imaged. The color of the lighting is close to white. As shown in FIG. 1, the angle θ between the optical axis L5 of the lighting device 5 and the vertical line VL varies from near the top position P2 of the round steel material S2 with the smallest diameter to the topmost position P1 of the round steel material S1 with the largest diameter. It is adjustable so that it can be irradiated in all cases up to the vicinity of .

Each lighting device 5 is connected to a lighting control device 6 (not shown) that controls a lighting power source, lighting brightness, and the like.
The computer system 7 also generates two captured images obtained by connecting in the circumferential direction the captured images of a plurality of uppermost positions (specific positions) P captured by each of the line sensor cameras constituting each of the imaging devices 3a and 3b. It includes a captured image superimposing section 8 that superimposes each of the images in the longitudinal direction, and a marking extraction section 9 that extracts the marking M from the captured images superimposed by the captured image superimposing section 8. Each of the line sensor cameras, the pulse generator 18, and the host computer 11 constituting each imaging device 3a, 3b are connected to the computer system 7. Information such as the distance H between the two imaging devices 3a and 3b (see FIG. 6(a)) and the diameter of the round steel material S coated with the marking M to be detected is input from the host computer 11 to the computer system 7. Ru.

This computer system 7 is a computer system having an arithmetic processing function for realizing each function of the captured image superimposing section 8 and the marking extracting section 9 by executing a program on computer software. This computer system is configured with a ROM, RAM, CPU, etc., and realizes each of the above-mentioned functions on software by executing various dedicated programs stored in advance in the ROM, etc.
Further, the display device 10 displays the marking M extracted by the marking extraction section 9, the circumferential position and longitudinal position of the marking M, and the like.

Next, a method for superimposing captured images of steel materials and a method for detecting markings on steel materials using the marking detection device 1 for round steel materials S will be described with reference to FIGS. 3 and 4.
First, in step S1 shown in FIG. The surface of the steel material S is imaged (imaging step).

In this imaging step, each of the line sensor cameras constituting the imaging devices 3a and 3b arranged along the longitudinal direction of the round steel material S scans the surface of the round steel material S rotating in the circumferential direction on the turning roller 2 in the circumferential direction. The uppermost position (specific position) P is imaged for one round of the round steel material S at a predetermined period. At this time, the imaging range of the imaging device 3a is L3a, and the imaging range of the imaging device 3b is L3b, as shown in FIG. 6(a). Two captured images E3a and E3b (see FIGS. 6(b) and (c)) captured using the two imaging devices 3a and 3b have different ranges and overlapping ranges along the longitudinal direction of the round steel material S. .

Next, in step S2, the captured image superimposing unit 8 of the computer system 7 collects the captured circles obtained from each of the line sensor cameras configuring the imaging devices 3a and 3b arranged along the longitudinal direction of the round steel material S. Two captured images E3a and E3b of the steel material S are superimposed in the longitudinal direction (captured image superimposition step).
In this captured image superimposition step, the captured image superimposition unit 8 specifically displays captured images of a plurality of uppermost positions (specific positions) P captured by each of the line sensor cameras configuring each imaging device 3a, 3b. The two captured images E3a and E3b obtained by connecting the images in the circumferential direction are superimposed on each other in the longitudinal direction.

Here, as shown in FIG. 5, when an actual captured image E (see FIG. 5(a)) is obtained when the round steel material S is imaged using the adjacent imaging devices 3a and 3b as one, one imaging device 3a (images joined in the circumferential direction) E3a and the image taken by the other imaging device 3b (images joined in the circumferential direction) E3b are superimposed, and when the superposition is successful (Fig. 5e)), the actual captured image E and the superimposed image of the captured image E3a and the captured image E3b are the same.

On the other hand, when the captured image E3a of one imaging device 3a and the captured image E3b of the other are superimposed, if the superimposition position set on each captured image E3a, E3b is inappropriate and the superposition fails, the As shown in FIG. 5(c) and FIG. 5(d), the captured image E of the actual object and the superimposed image of the captured image E3a and the captured image E3b do not match. The reason for this is that the overlapping position must be set on each captured image E3a, E3b according to the working distances WD1, WD2 of each imaging device 3a, 3b. This is because it is set on images E3a and E3b.

Therefore, in the captured image superimposition step in this embodiment, that is, in the captured image superposition method, the captured image superimposition unit 8 executes steps S21 to S28 shown in FIG.
First, in step S21, the captured image superimposition unit 8 combines two captured images E3a and E3b (FIG. 6) of the round steel material S captured by each of the line sensor cameras configuring the two adjacent imaging devices 3a and 3b. b), (c)) is captured (image capture step).
In this image capturing step, the captured image superimposing unit 8 specifically images a specific position P in the circumferential direction of the round steel material S rotating in the circumferential direction on the turning roller 2 by each line sensor camera at a predetermined period. Captured images of a plurality of specific positions P are captured, and the captured images of the plurality of specific positions P captured by each line sensor camera are connected in the circumferential direction to obtain captured images E3a and E3b.

Next, in step S22, the captured image superimposition unit 8 selects the captured images from each of the captured image pickup devices 3a, 3b based on the distance H along the longitudinal direction between the line sensor cameras configuring the two image pickup devices 3a, 3b. The overlapping position T in the longitudinal direction of the two captured images E3a and E3b of the round steel material S by the line sensor camera constituting the is determined (overlapping position determining step). The distance H along the longitudinal direction between the two imaging devices 3a and 3b is the distance H between the center lines CL3a and CL3b of the two imaging devices 3a and 3b (see FIG. 6(a)).

In this superimposition position determining step, the captured image superimposition unit 8 specifically positions the position corresponding to the intermediate position in the longitudinal direction between the two line sensor cameras in each of the captured images E3a and E3b, around the circumference. The overlapping position T in the longitudinal direction of the images E3a and E3b of the round steel material S taken by the line sensor cameras connected in the direction is determined. The intermediate position in the longitudinal direction between the two line sensors is a position that is 1/2 of the distance H toward the imaging device 3b along the longitudinal direction from the position of the imaging device 3a. In this embodiment, the distance H along the longitudinal direction between the two line sensor cameras is set to, for example, 800 (mm), and 1/2 of the distance H from each line sensor camera = 400 (mm). ) is the intermediate position, and the position directly below this intermediate position in the captured images E3a and E3b is the longitudinal overlap of the captured images E3a and E3b of the round steel material S. The alignment position T is determined.

Here, as shown in FIGS. 6(a), (b), and (c), an image E3a of the round steel material S by the line sensor camera constituting the imaging device 3a, and an image E3a taken by the line sensor camera constituting the imaging device 3b. The captured image E3b of the round steel material S overlaps with the round steel material S in the longitudinal direction. In order to superimpose each captured image E3a, E3b, it is necessary to set where on each captured image E3a, E3b the superimposition position T determined above is located. When setting the overlapping position T on each captured image E3a, E3b, the pixel resolution of each line sensor camera is determined by the working distances WD1, WD2 (see FIG. 7) between each line sensor camera and the round steel material S. changes, it is necessary to set the working distances WD1, WD2 and the pixel resolution of each line sensor camera in consideration.

For this reason, when step S22 is completed, the process moves to step S23, and the captured image superimposing unit 8 first selects each of the two imaging devices 3a and 3b and the round steel material S to be photographed according to the dimensions of the round steel material S to be photographed. The working distances WD1 and WD2, which are the distances between the working distance and the working distance S, are calculated (working distance calculating step).
Calculation of the working distances WD1 and WD2 will be explained with reference to FIG. 7. Here, since the methods for calculating the working distances WD1 and WD2 are the same, the method for calculating the working distance WD will be simply described. The working distance WD is calculated by the following equation (1).
WD=W-φ-Δw...(1)
Here, W is the distance between the apex of the turning roller 2 and the line sensor camera that constitutes each imaging device 3a, 3b, Φ is the diameter of the round steel material S to be detected, and Δw is the amount of error due to the turning roller installation. be.

Further, the error amount Δw due to the installation of the turning roller is calculated by the following equation (2).
Δw=r-(h-t)...(2)
Here, r is the radius of the round steel material S, h is the vertical height from the center of the round steel material S to the center of the turning roller 2, and t is the radius of the turning roller 2.
Further, the vertical height h from the center of the round steel material S to the center of the turning roller 2 is calculated by the following equation (3).
h=sinβ×(r+t)...(3)
Here, β is the angle between the straight line connecting the center of the turning roller 2 and the center of the round steel material S and the horizontal axis.
Furthermore, this β is calculated by the following equation (4).
β=arccos((p/2)/(r+t))...(4)
Here, p is the arrangement pitch of the turning rollers 2.

Then, each value of W, Φ, r, t, and p is input from the host computer 11 to the captured image superimposition unit 8, and the captured image superposition unit 8 inputs the input W, Φ, r, t, and The working distance WD is calculated by introducing each value of p into equations (1), (2), (3), and (4). The imaging devices 3a and 3b may be installed such that the working distance WD is the same value for the imaging device 3a and the imaging device 3b, and W is the same value for both the imaging device 3a and 3b. , when the imaging devices 3a and 3b are installed so that W is different, the working distance for each of the imaging devices 3a and 3b is determined individually.

Next, in step S24, the captured image superimposition unit 8 adjusts the longitudinal superposition position T of the two captured images E3a and E3b determined in step S22 according to the working distances WD1 and WD2 calculated in step S23. , are set on each of the two captured images E3a and E3b captured in step S21 (overlapping position setting step).
In step S24, the imaging device superimposition unit 8 first calculates the pixel resolution (mm/pix) of the line sensor camera forming each imaging device 3a, 3b based on the working distances WD1, WD2 calculated in step S23. do.

In this calculation, the pixel resolution (mm/pix) of each line sensor camera is determined for each working distance (mm), and the value for each working distance (mm) is stored in the storage unit (not shown) of the computer system 7. The pixel resolution (mm/pix) of the line sensor camera is stored. In this embodiment, as shown in Table 1, the value of pixel resolution (mm/pix) is determined for each diameter of the round steel material S corresponding to the working distance (mm). For example, when the diameter of the round steel material S is 200 mm, the pixel resolution of each line sensor camera is 0.805 (mm/pix).

Next, in step S24, the captured image superimposition unit 8 merges the end portions E3ae, Positions shifted by the pixel numbers napix and nabpix calculated from E3be are set as overlapping positions T3a and T3b of the images E3a and E3b of the round steel material S captured by each line sensor, respectively.

Specifically, as shown in FIG. 6(b), a position corresponding to the intermediate position in the longitudinal direction between the line sensor cameras in the captured image E3a is set as an overlapping position T (T3a), and the captured image E3a is captured. The number of pixels napix along the longitudinal direction from the end E3ae on the device 3b side to the overlapping position T (T3a) is determined. This number of pixels napix is determined by the distance H between the two imaging devices 3a and 3b, the working distance WD1 of the imaging device 3a calculated in step S23, the angle of view αa, and the pixel resolution (mm/pixel) of the imaging device 3a calculated above. ) from the following equation (5).
napix=[WD1×tan(αa/2)-H/2]/pixel resolution (WD1)...(5)

Similarly, as shown in FIG. 6(c), the position corresponding to the intermediate position in the longitudinal direction between the line sensor cameras in the captured image E3b is set as an overlapping position T (T3b), and the imaging device of the captured image E3b The number nbpix of pixels along the longitudinal direction from the end E3be on the 3a side to the overlapping position T (T3b) is determined. This number of pixels nbpix is determined by the distance H between the two imaging devices 3a and 3b, the working distance WD2 of the imaging device 3b calculated in step S23, the angle of view αb, and the pixel resolution (mm/pix) of the imaging device 3b calculated above. ) from the following equation (6).
nbpix=[WD2×tan(αb/2)-H/2]/pixel resolution (WD2)...(6)
Thereby, step S24 ends, and the process moves to step S25.

Then, in step S25, the captured image superimposition unit 8 converts each of the two captured images E3a and E3b of the round steel material S by each line sensor camera into the round steel material S by each line sensor, as shown in FIG. 6(d). The captured images E3a and E3b of S are superimposed so that superimposition positions T3a and T3b set on each of them are superimposed on each other (superimposition step). Thereby, the two captured images E3a and E3b are superimposed at the superposition position T.

Here, in the above-mentioned step S24 (superimposition position setting step), the pixel resolution (mm/pix) of each line sensor camera is not calculated based on the working distances WD1 and WD2, but the pixel resolution (mm/pix) is set uniformly. pix), and the numbers of pixels napix and nabpix in the captured images E3a and E3b of the round steel material S by each line sensor camera are calculated based on this uniform pixel resolution (mm/pix). Then, the position shifted by the number of pixels napix, nabpix calculated from the end of each of the images E3a, E3b of the round steel material S captured by each line sensor is changed to When the overlapping positions T3a and T3b are set, the overlapping positions T3a and T3b of the captured images E3a and E3b are inappropriate positions, respectively.

Therefore, in step S24, by calculating the pixel resolution (mm/pix) of each line sensor camera based on the working distances WD1 and WD2, changes in pixel resolution can be reliably dealt with, and each line sensor The overlapping positions T3a and T3b of the captured images E3a and E3b of the round steel material S can be appropriately set.
Next, in step S26, the captured image superimposing unit 8 performs noise removal processing on the superimposed captured images E3a and E3b except for the marking M (including M3a+M3b).
Next, in step S27, the captured image superimposing unit 8 performs a binarization process on the superimposed captured images E3a and E3b from which noise has been removed.

Thereafter, in step S28, the captured image superimposition unit 8 outputs the binarized image from which noise has been removed to the marking extraction unit 9. Thereby, step S2 (captured image superimposition step) ends.
After step S2 is completed, as shown in FIG. 3, the process moves to step S3.
In step S3, the marking extraction unit 9 extracts the marking M from the superimposed captured images E3a and E3b of the round steel material S (marking extraction step).
In this marking extraction step, the marking M and the circumferential and longitudinal positions of the marking M are extracted.

Finally, in step S4, the display device 10 displays the marking extraction results output from the marking extraction unit 9 of the computer system 7, that is, the markings M and the circumferential and longitudinal positions of the markings M.
Further, after steps S1 to S4 are completed, the cart is moved in the longitudinal direction and placed on the turning roller 2 until the imaging range L3b captured by the imaging device 3b becomes the imaging range L3a captured by the imaging device 3a. Move the placed round steel material S in the longitudinal direction, perform steps S1 to S4, and repeat steps S1 to S4 and the longitudinal movement of the round steel material S until the marking M of the entire length of the steel material S is detected. Just repeat.

As described above, according to the captured image superimposition method and the marking detection method according to the present embodiment, each of the two imaging devices 3a and 3b and the round steel material S to be photographed are adjusted according to the dimensions of the round steel material S to be photographed. Working distances WD1 and WD2 are calculated (step S23). Then, the overlapping position T of the two captured images E3a, E3b is set on each of the two captured images E3a, E3b according to the calculated working distances WD1, WD2 (step S24). Further, each of the two captured images E3a and E3b is superimposed at superimposition positions T(T3a) and T(T3b) set on each of the two captured images E3a and E3b (step S25).

As a result, two images having different and overlapping ranges along the longitudinal direction are obtained by imaging the round steel material S extending in the longitudinal direction using two longitudinally adjacent imaging devices 3a and 3b. It is possible to provide a method for superimposing captured images of a steel material and a method for detecting markings on a steel material, which can appropriately superimpose the captured images E3a and E3b in the longitudinal direction according to the working distance WD.

Although the embodiments of the present invention have been described above, the present invention is not limited thereto, and various changes and improvements can be made.
For example, in this embodiment, two imaging devices 3a and 3b are arranged along the longitudinal direction of the round steel material S, but two or more imaging devices may be arranged.
In this case, steps S1 to S4 may be executed when overlapping two captured images including the same longitudinal portion of the round steel material S obtained from any two imaging devices in the longitudinal direction. Steps S1 to S4 may be executed when two captured images adjacent in the direction are superimposed in the longitudinal direction.
Furthermore, in this embodiment, the steel material is a round steel material S, and the imaging devices 3a and 3b are line sensor cameras, but steel materials other than the round steel material S and imaging devices other than line sensor cameras may be used.

In this case, in the image capturing step, each of the captured images E3a and E3b of the steel material captured by the two imaging devices 3a and 3b is captured, and in the overlapping position determination step, the longitudinal direction between the two imaging devices 3a and 3b is captured. Based on the distance H along the longitudinal direction, the overlapping position T in the longitudinal direction of the captured images E3a and E3b of the steel material captured by the respective imaging devices 3a and 3b is determined. In addition, in the working distance calculation step, working distances WD1 and WD2, which are the distances between each of the two imaging devices 3a and 3b and the round steel material S to be photographed, are calculated according to the dimensions of the round steel material S to be photographed. . Then, in the superposition position setting step, the superposition position T in the longitudinal direction of the two determined captured images E3a and E3b is set according to the working distances WD1 and WD2 calculated in step S23. It is set on each of the captured images E3a and E3b. Furthermore, in the superimposition step, each of the two images E3a and E3b of the round steel material S taken by each line sensor camera is placed at a superimposition position set on each of the images E3a and E3b of the round steel material S taken by each line sensor. Overlap T3a and T3b so that they overlap each other.

In addition, as an example in which the steel material is other than the round steel material S, there is an example in which a square billet S with a square cross section as shown in FIG. Preferably, it is an area sensor camera.
Furthermore, the steel material to be imaged may not have markings applied to the positions of flaws formed on its surface.
In addition, in the overlapping position determination step, when determining the overlapping position T in the longitudinal direction of each captured image E3a, E3b of the steel material captured by each imaging device 3a, 3b, the longitudinal direction between the two imaging devices 3a, 3b is determined. It does not necessarily have to be based on the distance H along the direction.

In addition, in the superposition position determination step, a position corresponding to the intermediate position in the longitudinal direction between the two line sensor cameras in each of the captured images E3a and E3b is set as a circle formed by each line sensor camera connected in the circumferential direction. The superimposition position T in the longitudinal direction of each of the captured images E3a and E3b of the steel material S is determined, but the superposition position T is determined to be a position directly below the intermediate position in the longitudinal direction between the two line sensor cameras. It is not necessarily necessary to do so.
Further, the two imaging devices 3a and 3b do not necessarily need to be installed so that the imaging line runs along the longitudinal direction of the round steel material S.

1 Marking detection device 2 Turning rollers 3a, 3b Imaging device 4 Camera control device 5 Illumination device 6 Illumination control device 7 Computer system 8 Captured image superimposition section 9 Marking extraction section 10 Display device 11 Host computer 12 Pedestal section 13 Support leg 14 No. 1 support member 15 second support member 16 third support member 17 fourth support member 18 pulse generator P uppermost position (specific position)
S Round steel material (steel material)

Claims (5)

A steel material that uses a steel material extending in the longitudinal direction as a subject, and superimposes two captured images having different and overlapping ranges along the longitudinal direction, obtained by imaging using at least two imaging devices, in the longitudinal direction. A method for superimposing captured images, the method comprising:
Calculating the working distance of each imaging device, which is the distance between each of the two imaging devices and the steel material serving as the subject, according to the dimensions of the steel material serving as the subject;
setting an overlapping position of the two captured images on each of the two captured images according to the calculated working distance of each imaging device ;
Each of the two captured images is superimposed at an overlapping position set on each of the two captured images,
When setting the overlapping position of the two captured images on each of the two captured images, calculating the pixel resolution of each of the two imaging devices based on the calculated working distance of each imaging device, of the two captured images based on the distance between the two imaging devices, the calculated working distance of each imaging device, the angle of view of each imaging device, and the calculated pixel resolution of each of the two imaging devices. The number of pixels from each edge to the overlapping position of the two captured images is calculated, and a position shifted by the calculated number of pixels from each edge of the two captured images is calculated. A method for superimposing captured images of steel materials, characterized by setting the images at superimposed positions . The steel material is a round steel material, each of the two imaging devices is a line sensor camera, and the two imaging devices are installed so that an imaging line follows the longitudinal direction. 1. The method of superimposing captured images of steel materials according to 1. Each of the two captured images is a plurality of captured images obtained by capturing images of specific positions in the circumferential direction of the round steel material rotating in the circumferential direction on the turning roller at a predetermined period by the line sensor camera, and the captured images are connected in the circumferential direction. 3. The method of superimposing captured images of steel materials according to claim 2, wherein the images are obtained by 4. The method for superimposing captured images of steel materials according to any one of claims 1 to 3, wherein the steel material to be photographed is coated with markings at the positions of flaws. A method for detecting markings on a steel material that detects markings applied at positions of flaws formed on the surface of a steel material extending in the longitudinal direction, the method comprising:
imaging the surface of the steel material using at least two imaging devices to obtain two captured images having different and overlapping ranges along the longitudinal direction;
By the method for superimposing captured images of steel materials according to claim 4, superimposing two captured images obtained from the two imaging devices in the longitudinal direction,
A method for detecting markings on steel materials, characterized by extracting markings from superimposed captured images of steel materials.

Images