JP5765040B2 - Wrinkle detection method and wrinkle detection device - Google Patents

Description

  The present invention relates to a flaw detection technique for detecting a surface flaw of a predetermined size or more existing on a surface to be inspected such as a steel surface in a process of manufacturing and inspecting a steel product. In particular, the present invention relates to a technique effective for detecting wrinkles in an environment where water such as water droplets may adhere to the surface to be inspected.

In the conventional steel product inspection process, defects are detected by eddy current flaw detection and leakage magnetic flux flaw detection, etc. Image analysis is used to detect wrinkles.
The above-described flaw detection method using leakage magnetic flux is used in fields such as circle devotion. However, when a cross-sectional shape is complicated, for example, a shape steel material, the magnetic flux that is the premise of inspection is used. There is a problem in that it may be difficult to keep it constant. In addition, since a certain amount of time is required to hold the magnetic flux, there is a problem that there is a restriction when incorporating the flow into the manufacturing process.
Here, in Patent Document 1, inspecting the surface defect of the surface to be inspected by imaging the surface to be inspected in a state where the surface to be inspected is illuminated to improve the reflection of the surface to be inspected. Is disclosed.

JP 2010-38759 A

In the technique described in Patent Document 1, light from illumination is reflected by a surface to be inspected, and the reflected light is acquired by an imaging unit. However, each arrangement is set to an optimum value from the viewpoint of the S / N ratio. Is set. However, this method has a problem in that it is easily affected by the surface properties of the surface to be inspected, and the surface properties may cause over-detection or false detection.
In particular, according to experiments by the present inventors, it has been confirmed that there is a possibility that a water droplet placed on the surface to be inspected may be misrecognized as a soot.
The present invention has been made paying attention to the above points, and has an object to reduce the influence of the surface properties of the surface to be inspected and to detect a target wrinkle with higher accuracy.

In order to solve the above-described problems, the present invention detects wrinkles by performing image processing on a captured image after reducing the influence of the surface properties of the surface to be inspected.
That is, the invention described in claim 1 is a wrinkle detection method for detecting a surface defect present on the surface to be inspected based on a captured image obtained by photographing the surface to be inspected of a rolled material being conveyed in a steel process line. ,
The surface to be inspected is irradiated with light having an illuminance of 1000 lux or more so as to mask irregularities with a depth of 1.0 mm or less so that halation occurs at least in the wrinkle detection region of the captured image. And

Next, the invention described in claim 3 includes an illuminating device that irradiates light onto a surface to be inspected of a rolled material being conveyed in a steel process line, and a camera that photographs the surface to be inspected. A wrinkle detection device that detects a surface defect present on a surface to be inspected based on a captured image,
The illumination device irradiates the surface to be inspected with light having an illuminance of 1000 lux or more so as to generate halation in at least a wrinkle detection region of the captured image and mask an unevenness having a depth of 1.0 mm or less. And
Next, the invention described in claim 4 is a cooling step in which the flaw detection device uses water for cooling at least a part of the step of manufacturing a rolled steel material in the configuration described in claim 3. It is provided in the exit side of this, or the exit side of a washing process.

According to the present invention, unlike a normal idea, by increasing the amount of light to be irradiated and causing halation, the surface roughness, etc., consisting of minute irregularities as surface properties in the captured image is skipped. This makes it possible to make the kite to appear more clearly.
Further, even if water such as water droplets adheres to the surface to be inspected, the contrast difference of water such as water droplets is reduced by causing halation. For this reason, it is possible to reduce or eliminate erroneous detection of water droplets or the like as wrinkles.

It is a conceptual diagram which shows the structure of the wrinkle detection apparatus based on embodiment based on this invention, Comprising: (a) is the side view, (b) is the perspective view. It is a figure which shows the example of the captured image in a comparative example. It is a figure which shows the example of the captured image in an Example. It is a figure which shows the example of the captured image (with water wetting) in a comparative example. It is a figure which shows the example of the captured image (with water wetting) in an Example.

Next, an embodiment according to the present invention will be described with reference to the drawings.
FIG. 1 is a conceptual diagram illustrating a configuration of a wrinkle detection device according to the present embodiment.
In this embodiment, the case where a wrinkle detection apparatus is arrange | positioned at the inspection line which is a part of manufacturing processes, such as a steel plate, is illustrated. And the wrinkle detection apparatus of this embodiment inspect | inspects the surface of the said rolling material, while conveying rolling materials, such as a steel plate, with a conveyance apparatus (not shown) in an inspection line. That is, the surface of the rolled material is used as the surface to be inspected. In the following example, the upper surface of the conveyed rolled material is exemplified as the inspection surface, but the lower surface of the rolled material may be the inspection surface.
Note that the surface to be inspected may not be flat. A surface such as a shape steel or a steel bar may be used as a surface to be inspected.

Moreover, the soot detecting device of the present invention is a soot detecting device suitable for detecting soot on the exit side of the cooling process using water for cooling at least part of the process of manufacturing steel or the exit side of the cleaning process. is there.
Here, the “outgoing side of the cooling process or the outgoing side of the cleaning process” does not have to be the cooling process or the cleaning process. That is, even if any other processing process is performed from the cooling process or the cleaning process to the soot detecting device of the present embodiment, the surface to be inspected is caused by the cooling process or the cleaning process at the position of the soot detecting device. If there is a possibility that water droplets are attached to the surface, it is regarded as the exit side of the cooling process or the exit side of the cleaning process.

(Constitution)
As shown in FIG. 1, the wrinkle detection device of this embodiment includes a lighting device 1, a camera 2, and an image processing unit 3.
Here, the rolling material 4 such as a target steel plate is conveyed along the longitudinal direction by a table roller or other conveying device (not shown).
The illuminating device 1 is arranged so as to be able to irradiate light from an oblique direction inclined in the transport direction H with respect to the wrinkle detection position 5 on the upper surface of the rolled material 4. For example, the irradiation axis 1 </ b> L of the illumination device 1 is arranged to be inclined with respect to the normal (vertical line) at the wrinkle detection position 5 by a preset angle θ on the upstream side in the transport direction H. The angle θ with respect to the normal is set within ± 45 degrees, for example. In this embodiment, θ is set to 30 degrees. As the illumination device 1, a high-frequency fluorescent lamp can be exemplified.

The camera 2 is arranged at a position where the wrinkle detection position 5 on the upper surface of the rolled material 4 can be photographed. In the present embodiment, the photographing axis 2L of the camera 2 is set at the center of the wrinkle detection position 5 from an oblique direction inclined to the downstream side of the conveyance direction H with respect to the wrinkle detection position 5. .
Here, in the present embodiment, the light irradiated from the lighting device 1 is irradiated with the distance from the lighting device 1 to the eyelid detection position 5 so that halation occurs in the captured image obtained by photographing the eyelid detection position 5. The amount of light to be determined is determined by experiments or other methods. In this embodiment, a high-frequency fluorescent lamp of 20 V or higher is used, and the distance from the lighting device 1 to the wrinkle detection position 5 is set within 1160 mm, so that an illuminance of 1000 lux or more can be secured at the wrinkle detection position 5. did. This is because it has been confirmed that halation occurs at least in the image area where the wrinkle detection position 5 is captured in the captured image captured by the camera 2 by setting in this way.

The image processing unit 3 performs image processing of the captured image taken by the camera 2 and performs wrinkle detection determination.
The image processing unit 3 includes an image capturing unit 3A and a wrinkle detection unit 3B.
The image capturing unit 3A acquires the captured image captured by the camera 2 at a preset image capturing period and stores the acquired image in the storage unit.
The eyelid detection unit 3B sequentially inputs captured images from the storage unit, performs binarization processing (edge processing) on the input captured images, and converts the grayscale image of the target captured image into a density difference distribution. Convert to The dividing method in the screen is not particularly limited. For example, the camera 2 is divided in units of pixels, and binarization processing is performed for each divided region.

Next, the wrinkle detection unit 3B compares the converted density difference distribution with a preset threshold value, and determines that there is a surface defect when the threshold value is exceeded.
The image processing unit 3 outputs the determination result of the wrinkle detection unit 3B to, for example, a display unit and displays it on the display unit, or outputs it to a marking device or the like that marks the wrinkle position on the rolled material 4.

(Operation other)
For example, the surface of the rolled material 4 that has arrived at the inspection position in the rolled material 4 that is being conveyed in the steel process line is irradiated with light from the lighting device 1 with a light quantity that can generate halation, and the halation. The surface to be inspected in a state where the light is generated is photographed by the camera 2.
In the picked-up image taken with this camera 2, since halation has occurred, the surface roughness composed of minute irregularities formed on the surface of the rolled material 4 is masked by flying from the image. And the wrinkle more than predetermined depth becomes the picked-up image image | photographed with the predetermined | prescribed density | concentration difference.

In addition, the wrinkle detection apparatus of this embodiment detects the presence or absence of a wrinkle having a predetermined depth or more. In the experiment, irregularities with a depth of 1.0 mm or less were reliably masked by halation, but wrinkles with a depth of 2.0 mm or more were recognized by the captured image without being masked by halation.
As described above, with the wrinkle detection device according to the present embodiment, it is possible to detect a surface defect having a depth of, for example, 2.0 mm or more with higher accuracy while suppressing the influence of the error due to the surface property of the surface to be inspected. .

Note that since wrinkles are detected by image processing, the shape of the detected wrinkles can be detected with higher accuracy.
Moreover, in the case of the rolling material 4, the water droplet may have adhered to the surface by the process using water, such as a cooling process and a washing | cleaning process.
Here, in the case where wrinkle detection is performed using a photographed image shot by irradiating with a light amount capable of clearly photographing the surface properties as in the conventional case, even if it is a water droplet having a size of 1 to 2 mm, the water droplet May be falsely detected as 疵. This is because the irradiated light is irregularly reflected by the water droplets, and an apparent contrast difference is generated, thereby causing a false detection.

On the other hand, in the present embodiment, by generating the above-mentioned halation, it is possible to mask a water droplet that is erroneously detected as the above-described soot, and to detect only the target soot with higher accuracy.
As described above, in the present embodiment, since the influence of the surface property of the surface to be inspected can be reduced, it is possible to detect wrinkles in a form that is difficult with the prior art. Of course, there is also an effect that the false detection rate / overdetection rate can be lowered as compared with the prior art.

In addition, since the influence of the surface texture can be reduced, the equipment introduction cost for eliminating the influence of the surface texture can be reduced. For example, it becomes possible to reduce the drainage capacity of the drainer.
Further, in the automatic determination of surface defects, in the present embodiment, it can be realized only by a process of increasing the amount of light from the illumination device 1 so that the illuminance on the surface to be inspected is increased. That is, since the calculation load on the image processing side is not increased, the surface of the rolled material 4 being conveyed can be inspected continuously in a short time. It is necessary to adjust the threshold value.

In addition, in the said embodiment, the case where the surface defect of the depth more than the objective is detected during conveyance of the rolling material 4 is illustrated. Wrinkle detection may be performed with the surface to be inspected standing still. Alternatively, scanning for detection of wrinkles may be performed by relatively moving the pair of lighting device 1 and camera 2 of the wrinkle detection device.
Further, the target surface is not limited to the surface of the rolled material 4 such as a steel plate or a shape steel. This embodiment is particularly effective for detecting wrinkles on a surface to be inspected in a state where water droplets may be attached.

An example of wrinkle detection based on the wrinkle detection device of the above embodiment will be described.
The setting conditions of the example are as follows.
Target material: Steel plate having a width of 400 mm Surface to be inspected: Steel plate surface Steel sheet transport speed: 4.0 m / s on average
Illuminance at the eyelid detection position 5 by irradiation from the illumination device 1: 1230 lux Camera 2: Shutter speed 1/240, fully open aperture Here, due to the above settings, halation has occurred in the captured image taken by the camera 2 Have confirmed.
As a comparative example, the illuminance at the wrinkle detection position 5 due to irradiation from the lighting device 1 was set to 850 lux, and other conditions were set to the same conditions as in the example to detect wrinkles.

And the surface defect of the magnitude | size of 80 mm x 20 mm x maximum depth 2.0mm occurred on the steel plate surface, and the steel plate surface of the state which is not wetted is made into object, The detection of the wrinkles is carried out in the said Example and comparative example. I tried it under the set conditions.
The threshold values in the image processing unit 3 are calibrated to values that can detect defects.
As a result of the detection, the surface defect could not be detected with high accuracy in the comparative example, but the surface defect could be detected in this example.
This is because in this embodiment, the influence of the surface property of the surface to be inspected is reduced by creating a halation state, so that the contrast difference between the surface defect to be detected and the others can be imaged. Further, for this reason, it is obvious that the false detection rate and the excessive extraction rate are reduced.

An example of the picked-up image image | photographed with the camera 2 in a comparative example and an Example is shown in FIG.2 and FIG.3.
In the captured image of the comparative example (see FIG. 2), unevenness of 1 mm or less is also reflected, and the contrast difference at the boundary position of the surface defect to be detected is small. On the other hand, in the picked-up image (FIG. 3) of the present embodiment, the surface defect to be inspected is masked by halation caused by halation, which is smaller than the target wrinkles on the surface to be inspected.

  Here, regarding the conditions of the above-described embodiment, when the setting of the lighting device 1 is changed so that the illuminance at the wrinkle detection position 5 changes, and it is confirmed whether halation occurs in the captured image captured by the camera 2, 942 In the looks, it was not in the halation state, but it was confirmed that it was in the halation state in 1000 looks or more. This experiment was conducted in the range of 840 to 1143 lux.

Moreover, it was investigated whether or not wrinkle detection was affected by the difference in each area (position) in the captured image causing halation. According to the result of the investigation, although there was variation depending on the position, it was within the allowable range.
In the steel manufacturing process, the surface of a rolled material that has been wetted after water cooling may be subjected to surface inspection. In order to simulate this state, an imaging test was performed after water was applied to the lower side of the surface to be inspected.
The resulting images are shown in FIGS. 4 is a captured image according to the comparative example, and FIG. 5 is a captured image according to the embodiment.

  In the comparative example (see FIG. 4), the wetted portion is projected darkly, and the image has a boundary between the wetted (water droplet) surface and the dry surface. For this reason, in the comparative example (see FIG. 4), the boundary between the surface of water wetting (water droplets) and the dry surface, or the portion where the color changes due to the difference in the thickness of the attached water, etc. There is a high possibility of false detection as a contour of a surface defect. This also contributes to hindering detection of actually detected surface defects.

On the other hand, in the example (see FIG. 5), the boundary between the wet surface and the dry surface is not detected. As described above, when the present invention is used, the influence of water adhering to the surface can be reduced or eliminated. Therefore, in an Example (refer FIG. 5), it turns out that the false detection by a water drop can be reduced significantly compared with a comparative example.
In FIG. 5, a gray portion is visible on the upper end side or the lower end side, but this portion is a portion where no halation occurs. In the captured image shown in FIG. 5, the wrinkle detection region is set on the center side (a region where halation is reliably generated).

DESCRIPTION OF SYMBOLS 1 Illuminating device 1L Irradiation axis | shaft 2 Camera 2L Image | photographing axis | shaft 3 Image processing part 3A Image capture part 3B Soot detection part 4 Rolled material 5 Soot detection position H Conveyance direction

Claims (4)

A wrinkle detection method for detecting surface defects present on the surface to be inspected based on a captured image obtained by photographing a surface to be inspected of a rolled material being conveyed in a steel process line ,
The surface to be inspected is irradiated with light having an illuminance of 1000 lux or more so as to mask irregularities with a depth of 1.0 mm or less so that halation occurs at least in the wrinkle detection region of the captured image.疵 detection method.   2. The method for detecting wrinkles according to claim 1, wherein surface defects are detected for a surface to be inspected on which water may be attached to the surface. A surface defect present on the surface to be inspected based on an image captured by the camera, having an illumination device that irradiates light onto the surface to be inspected of the rolled material being conveyed in the steel process line and a camera that images the surface to be inspected A wrinkle detection device for detecting
The illumination device irradiates the surface to be inspected with light having an illuminance of 1000 lux or more so as to generate halation in at least a wrinkle detection region of the captured image and mask an unevenness having a depth of 1.0 mm or less. A wrinkle detection device. The said wrinkle detection apparatus is provided in the exit side of the cooling process which uses water for cooling at least one part in the process of manufacturing the rolled material of steel, or the exit side of the washing | cleaning process. The wrinkle detection device described.

Images