JPH09145638A - Method and device for detecting surface defect - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect streak like surface defects by accumulating 2-D picture data obtained by, whole irradiating in the travel direction, photodetecting the reflected light from a to-be-examined area and correcting distorting in data caused by irregularity in illumination in the width direction contained in the obtained travel direction accumulation data. SOLUTION: A lighting device 2 illuminates a molten galvanized steel plate which is a to-be-examined material 1 traveling in the longitudinal direction and the inside of a to-be-examined area 13 of the material 1, in the constant illumination in the longitudinal direction compared to the width direction orthogonal to the former. A 2-D CCD camera 3 photodetects a reflection light 13a from the area 13, to generate 2-D picture data representing each pixel of the 2-D picture in the area 13. A data accumulation means 4 accumulates the 2-D picture data obtained from the camera 3 in the longitudinal direction of the material 1, to generate travel direction accumulation data. A data distortion correcting means 5 corrects the data distortion caused by irregular illumination in the width direction of the area 13 contained in the travel direction accumulation data. A surface defect detecting means 6 detects streak like surface defects based on the corrected data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a surface defect inspection method and apparatus for detecting surface defects of a material to be inspected that is running and extends linearly in the running direction.

[0002]

2. Description of the Related Art In the hot dip galvanizing process for producing surface-treated steel sheets used in the fields of automobiles, home appliances, and the pre-process thereof, surface defects extending linearly along the running direction of the steel sheet occur. There is. Since this is a fatal defect for product quality, defects in the running steel sheet in the online process can be detected promptly, and as soon as a defect is detected, necessary measures such as changing the operating conditions can be taken to detect further defects. Preventing the occurrence is an important issue.
At that time, in order to take appropriate measures, it is necessary not only to detect the occurrence of a defect but also to accurately determine the degree of the defect (hereinafter, referred to as a defect level).

It is not only the surface-treated steel sheet that the surface defects extending in a streak pattern along the running direction occur, but also other steel sheets, aluminum sheets, papers, etc. In the case of a color tone defect of a zinc-plated steel sheet, it is particularly difficult to detect the presence / absence of a defect and determine the defect level because the difference in color tone between the defect portion and the sound portion is extremely small.

In the case of detecting a defect based on such a slight difference in color tone, if there is uneven illuminance in the inspection area on the surface of the inspection material, it becomes more difficult to determine the defect level. In order to eliminate the uneven illuminance of the area to be inspected, it is sufficient to eliminate the uneven illumination of the light source that illuminates the area to be inspected.To do so, a multi-point lighting system or a complex lens system that reduces uneven illumination in two dimensions Large-scale capital investment, such as construction of lighting equipment using, is unavoidable, and there is a risk of significant cost increases.

Therefore, conventionally, in detecting a surface defect extending linearly along the traveling direction as described above, there is uneven illuminance in both the traveling direction of the surface of the material to be inspected and the width direction intersecting with the traveling direction. In this case, the unevenness of the illuminance in the running direction and the width direction of the material to be inspected included in the two-dimensional image obtained by imaging the surface of the material to be inspected with the two-dimensional imager is simultaneously corrected by the image processing, and then the defect is extended. It is common to detect a defect by integrating the two-dimensional image signal in the traveling direction and emphasizing the density difference between the defective part and the sound part.

In this case, the illuminance unevenness is corrected by subjecting all image data to a two-dimensional smoothing process and taking the difference between the smoothed image and the original image to obtain a low illumination unevenness which is a main component. This is done by removing the density change of the frequency.
Usually, a moving average method is used as a calculation method for this smoothing process. That is, for two-dimensional image data, first, for each data in each data row in either the vertical or horizontal direction, an operation of obtaining an average value of a plurality of data adjacent to the data is performed. It is executed while moving one direction at a time in one direction, and after the calculation of the data string in that one direction is completed, the same calculation is executed for each data in each data string in the other direction. Dimension smoothing is performed.

Therefore, since the calculation of the moving average method is executed the same number of times as the number of pixels of the area to be inspected, for example, in the case of vertical 400 × horizontal 500 pixels, the moving average calculation is executed 200,000 times. The Rukoto.

[0008]

By the way, in order to apply such a defect detection method to an online process, it is necessary to repeatedly perform image processing on a running inspection material at high speed. In the defect detection method in which the correction is performed by the two-dimensional smoothing process, the moving average calculation as described above is necessary in the smoothing process, and the traveling of the inspected material must be suppressed to a low speed due to the restriction of the processing time. For example, it is not practical as an online defect inspection device.

As a solution to this problem, for example, it is conceivable to shorten the operation time by using a dedicated LSI chip that executes only smoothing processing at high speed, but this leads to an increase in cost and defect inspection accompanying the specialization. There is a problem that the versatility of the device is lost. If there is illuminance unevenness in the running direction within the inspection area but no illuminance unevenness in the width direction, there is a method of integrating the image signals in the running direction to detect the presence or absence of defects. This case is more advantageous than the above-mentioned case in that it is not necessary to correct the uneven illuminance, but this method also has the following problems.

FIG. 8 is a diagram showing a two-dimensional image in the case where there is no illuminance unevenness in the width direction of the region to be inspected and there is illuminance unevenness in the running direction, a diagram showing illuminance unevenness LA in the running direction A, and two. 3 is a profile of running direction integrated data obtained by integrating image data representing each pixel of a three-dimensional image in the running direction. FIG.
In (a), two stripe-shaped defects 30a and 30b having the same defect level extending along the traveling direction A of the inspection area appear at different positions in the traveling direction A on the two-dimensional image. The situation is shown.

Due to the illuminance unevenness LA in the traveling direction A of the region to be inspected, the color tone of the two-dimensional image shown in FIG. 8A is such that the region near the center of the traveling direction A is closer to the front and rear regions of the traveling direction A. It looks bright. Therefore, looking at the profile 15 of the traveling direction integrated data obtained by integrating the two-dimensional image data representing each pixel of this two-dimensional image in the traveling direction A of the inspection area, as shown in FIG. , The peak value 31 corresponding to the defect 30a located in the region near the center of the bright tone
“A” indicates a value higher than the peak value 31b corresponding to the defect 30b located in the front area where the color tone is dark. in this way,
Even if the surface defect has the same defect level, the peak value appearing in the profile 15 of the traveling direction integrated data varies depending on which region of the traveling direction A in the inspected region is located. It is not possible to accurately evaluate the defect level of each of the two defects 30a and 30b from 31b.

In view of the above circumstances, the present invention is capable of promptly detecting the presence or absence of defects extending linearly along the running direction of the material to be inspected, and accurately determining the defect level. An object of the present invention is to provide a defect inspection method and apparatus.

[0013]

According to the surface defect inspection method of the present invention for achieving the above object, a predetermined inspected region on a surface of an inspected material which travels in a predetermined traveling direction is provided with an illuminance within the inspected region. 2D image data that represents each pixel of the 2D image of the inspection area by illuminating with a uniform illuminance in the traveling direction compared to the width direction intersecting with the traveling direction and receiving reflected light from the inspection area with a 2D imager To generate the running direction integrated data by integrating the two-dimensional image data obtained by the two-dimensional imager in the running direction of the material to be inspected, and included in the generated running direction integrated data. It is characterized in that the distortion of the data caused by the uneven illuminance in the width direction is corrected, and the surface defects extending in a streak shape along the running direction of the inspection material are detected from the data in which the distortion is corrected.

Further, according to the surface defect inspection apparatus of the present invention which achieves the above-mentioned object, a predetermined inspection area on the surface of the inspection object traveling in a predetermined traveling direction is set so that the illuminance in the inspection area is in the traveling direction. An illuminator that illuminates the traveling direction with a uniform illuminance compared to the intersecting width direction, and two-dimensional imaging that receives reflected light from the inspection area and generates two-dimensional image data that represents each pixel of the two-dimensional image of the inspection area. Device, data accumulating means for accumulating the two-dimensional image data obtained by the two-dimensional imaging device in the traveling direction of the material to be inspected to generate traveling direction accumulating data, and the traveling direction accumulating data obtained by the data accumulating means. included,
Data distortion correction means for correcting data distortion caused by unevenness of illuminance in the width direction of the inspection area, and stripe data extending from the data whose distortion has been corrected by the data distortion correction means along the running direction of the inspection material. And a surface defect detecting means for detecting a surface defect.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below. FIG. 1 is a schematic diagram showing an embodiment of the surface defect inspection apparatus of the present invention. As shown in FIG. 1, in this surface defect inspection apparatus 10, a hot-dip galvanized steel sheet which is an inspected material 1 traveling in an arrow direction A and an inspected region 13 of the inspected material 1 in an arrow A direction. A lighting device 2 that illuminates the traveling direction A with a uniform illuminance as compared with the intersecting width direction B, and a two-dimensional image representing each pixel of the two-dimensional image of the inspection region 13 by receiving the reflected light 13a from the inspection region 13. A two-dimensional CCD camera 3 for generating data, and a data accumulating means 4 for accumulating two-dimensional image data obtained by the two-dimensional CCD camera 3 in the traveling direction A of the inspected material 1 to generate traveling direction integrated data.
And the inspection area 1 included in the traveling direction integrated data
Data distortion correction unit 5 that corrects data distortion caused by unevenness of illuminance in the width direction B of No. 3, and data whose distortion has been corrected by the data distortion correction unit 5 along the traveling direction A of the inspected material 1 Surface defect detecting means 6 for detecting the surface defect extending to the.

The lighting device 2 has a length of 760 mm and a capacity of 16 mm.
Two fluorescent lamps having a frequency of 20 W and a frequency of 20 kHz are arranged in parallel with the width direction of the material to be inspected 1 and are arranged at an interval of 300 mm, and are installed above the surface of the material to be inspected 1 by a distance of 300 mm. . With this illuminating device 2, the inspection area 13 of 200 mm in the traveling direction A and 250 mm in the width direction B, which is defined on the surface of the inspection material 1, is inspected.
The illuminance inside is illuminated with more uniform illuminance in the traveling direction A than in the width direction B intersecting the traveling direction A, and the illuminance unevenness in the traveling direction A in the inspection region 13 is practically sufficiently suppressed.

The number of pixels of the two-dimensional CCD camera 3 is 432 in the vertical direction.
The number of pixels is 512 pixels, and the vertical direction is aligned with the running direction A of the inspection object 1 and is installed. Although the two-dimensional CCD camera 3 is used to generate the two-dimensional image data in the present embodiment, the device for generating the two-dimensional image data is not limited to the two-dimensional CCD camera. Any two-dimensional image pickup device generally used may be used.

Next, one embodiment of the surface defect inspection method of the present invention, which is performed by using the above-described surface defect inspection apparatus (see FIG. 1), will be described. FIG. 2 shows a two-dimensional CCD for the inspected area on the surface of the traveling inspected material illuminated as described above.
It is a figure which shows the two-dimensional image obtained by imaging with a camera, and a figure showing the illuminance unevenness LB of the width direction B.

In this two-dimensional image, the illuminance unevenness in the traveling direction A is extremely small as described above. However, as shown in FIG. 2, the illuminance unevenness LB in the width direction B is large, which is 1 in the traveling direction A. The ratio of the maximum image density to the minimum image density on the pixel data of the column is only about 5/256, while the width direction B
The ratio of the maximum image density to the minimum image density on the pixel data of one column reaches about 20/256.

In such a two-dimensional image, a streak-like defect 30 called a black streak, which extends in a streak direction along the running direction A of the inspection object 1, appears. FIG. 3 shows one of the two-dimensional image data representing each pixel of the two-dimensional image shown in FIG.
It is a profile of pixel data for columns. Next, the two-dimensional image data representing each pixel of the two-dimensional image is sent to the data integrating means 4 and integrated in the running direction A of the inspection object 1 to generate running direction integrated data.

FIG. 4 is a profile of traveling direction integrated data obtained by integrating the two-dimensional image data representing each pixel of the two-dimensional image shown in FIG. 2 in the traveling direction A. In the profile 14 of pixel data for one column in the width direction shown in FIG.
The peak value 31c corresponding to the defect 30 is smaller than the noise component, and the S / N of the peak value 31c for the noise component
Although the ratio is about 1.3, in the profile 15 of the traveling direction integrated data shown in FIG. 4, the noise component is suppressed by the integration process, and the peak value 31 for the noise component is reduced.
The S / N ratio of d is improved to about 6.5.

As described above, when trying to detect a line-shaped defect extending in one direction, the defect can be easily detected by integrating the image data in the extending direction of the defect. Next, the traveling direction integrated data is sent to the data distortion correction means 5, and the distortion of the data included in the traveling direction integrated data, which is caused by the illuminance unevenness in the width direction B of the inspection region 13, is corrected.

The data distortion correction in this embodiment is performed as follows. First, the smoothing process of the traveling direction integrated data is performed. FIG. 5 is a profile of smoothed data obtained by smoothing the individual data constituting the traveling direction integrated data shown in FIG. 4 by the moving average method with 16 data before and after.

As shown in FIG. 5, the smoothed data profile 16 has a smooth shape in which fine irregularities are smoothed. In this embodiment, the number of pieces of processing data for each average value calculation in the moving average method is 16, but selection of the number of pieces of processing data is important for smoothing processing, and an image corresponding to this number of pieces of processing data is selected. The size in the upper width direction needs to be sufficiently larger than the size in the width direction of the streak defect to be detected and smaller than the change in the background of the image. By doing this, the profile near the defect is hardly changed,
It is possible to extract a main component of illuminance unevenness (a component wavily convex upward) having a period longer than the size in the width direction on the image corresponding to the number of processed data.

Next, the difference between the running direction integrated data of the profile 15 shown in FIG. 4 and the profile 16 shown in FIG. 5 is calculated, and the average value of the running direction integrated data is added to the difference in the illuminance unevenness in the width direction. The generated data distortion is corrected. FIG. 6 is a profile of image data after the distortion of the data included in the traveling direction integrated data, which is caused by the unevenness of the illuminance in the width direction, is corrected.

As shown in FIG. 6, this profile 17
From FIG. 5, the data distortion caused by the uneven illuminance in the width direction is removed, and the profile shape near the peak value 31e corresponding to the defect 30 (see FIG. 2) is stored with almost no influence. (See FIG. 4) The data whose distortion has been corrected by the data distortion correction means 5 (see FIG. 1) is sent to the surface defect detection means 6.

The surface defect detecting means 6 compares the distortion-corrected data with a predetermined threshold value to detect the defect 3
0 is detected. The defect level of the defect 30 in the original inspection area 13 can be determined by obtaining the difference between the detected defect density level and the sound part density level. Next, in the present embodiment, a state of defect detection when a plurality of defects exist at different positions in the traveling direction in the inspection area will be described.

FIG. 7 is a diagram showing a state in which two defects are present in the two-dimensional image in the present embodiment, a diagram showing illuminance unevenness LB in the width direction B, and an image showing each pixel of the two-dimensional image. 3 is a profile of running direction integrated data obtained by integrating data in the running direction. Similarly to FIG. 2, in FIG. 7A, the illuminance unevenness in the traveling direction A of the inspection region 13 is extremely small, but the large illuminance unevenness LB exists in the width direction B, and the central portion in the width direction B is bright. However, it gets darker as it gets closer to the left and right ends. The first defect 30a is located near the center of the inspection region 13 in the width direction B, and the second defect 30b having the same defect level as the first defect 30a is located slightly near the right end.

As shown in FIG. 7B, in the profile 18 of the running direction integrated data obtained by integrating the two-dimensional image data in the running direction, the illuminance unevenness LB component is superimposed over the entire profile 18. , The first defect 30
The peak values 31a and the peak values 31b corresponding to a and the second defect 30b have substantially the same length of protrusion from the neighboring level. Therefore, it can be easily estimated that the defect levels of the two defects 30a and 30b having different positions in the width direction B can be determined to be the same by performing the subsequent data distortion correction process.

As described above, the profile 19 in the case where there is no illuminance unevenness in the width direction B and there is illuminance unevenness LA in the running direction
In FIG. 8, two defects 30a, 30 having different positions in the width direction B are caused by the uneven illuminance LA in the traveling direction A.
The effect of the present invention is clear in comparison with the fact that the defect levels of b were not judged to be the same.

Table 1 is a table showing the determination of streak defects in the hot-dip galvanized steel sheet by the surface defect inspection method and apparatus of this embodiment.

[0032]

[Table 1]

Table 1 shows a comparison between the defect determination result by the surface defect inspection apparatus of the present embodiment and the defect determination result visually by the inspector for a large number of streaky defect samples in the inspection area. ing. The numerical value represents the number of defects judged as streak defects. In Table 1, large, medium, and small indicate the size of a defect, which is one of the defect levels. The way to read this table will be described with reference to an example. For example, in the surface defect inspection apparatus of the present embodiment, the number of streaky defects having a large size is 1.
While it was judged that there were two, there were 10 streaky defects of large size in the visual judgment for the same sample,
This indicates that the number of streak defects in the size is determined to be two. That is, the numerical values on the diagonal line to the lower right of the 3 × 3 matrix indicate that the judgment results of both sides are the same, and the numerical values at the positions other than the diagonal line are not the same. Indicates a number.

The coincidence rate obtained by dividing the number of coincidences 41 of both judgment results (total of numerical values on the diagonal line descending to the right) by the total number of samples 44 is about 93%, and both coincide well. I can say. In this embodiment, the difference method is used to correct the distortion of the data caused by the uneven illuminance in the width direction. It is not necessary to use the difference method described in the embodiment, and the effect of uneven illuminance can be removed by a differentiating method that differentiates in a range sufficiently narrower than the cycle of image density change due to uneven illuminance in the width direction.

In the above embodiment, only the example in which the material to be inspected is a steel plate has been described, but the material to be inspected according to the present invention is not limited to the steel plate, but may be an aluminum plate, paper, etc. Can also be applied to.

[0036]

As described above, according to the surface defect inspection method and apparatus of the present invention, the illuminance unevenness in the inspection area is eliminated in the traveling direction and the width direction of the inspection material as in the conventional method. Since it is not necessary to provide such a large-scale lighting equipment, it is sufficient to illuminate in the traveling direction with a uniform illuminance compared to the width direction,
The scale of the illumination equipment of the inspection device can be simplified, and the cost can be reduced.

Further, unlike the conventional method, it is not necessary to provide an image signal processing device for correcting the illuminance unevenness in both the traveling direction and the width direction, and the illumination in the traveling direction is more uniform than that in the width direction. Since the data distortion correction caused by the uneven illuminance can be performed only in the width direction, the number of moving average calculations in the data distortion correction processing is significantly reduced, and the time required for image signal processing is shortened. It is possible to perform high-speed defect inspection in the online process.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing an embodiment of a surface defect inspection apparatus of the present invention.

[Fig. 2] Two-dimensional C of the inspection area on the surface of the traveling inspection material
It is a figure which shows the two-dimensional image imaged by the CD camera, and a figure showing the illuminance unevenness LB in the width direction B. It is.

3 is a diagram showing a profile of pixel data for one column in the width direction of the two-dimensional image data representing each pixel of the two-dimensional image shown in FIG.

FIG. 4 is a diagram showing a profile of traveling direction integrated data obtained by integrating two-dimensional image data representing each pixel of the two-dimensional image shown in FIG. 2 in a traveling direction A.

FIG. 5 is a diagram showing a profile of smoothed data obtained by smoothing the individual pieces of data forming the traveling direction integrated data shown in FIG. 4 by a moving average method of front and rear 16 data.

FIG. 6 is a diagram showing a profile of image data after correction of data distortion caused by unevenness in illuminance in the width direction, which is included in traveling direction integrated data.

FIG. 7 is a diagram showing a state in which two defects are present in a two-dimensional image according to the present embodiment, a diagram showing illuminance unevenness LB in the width direction B, and image data representing each pixel of the two-dimensional image. It is a figure which shows the profile of the traveling direction integrated data integrated in the direction.

FIG. 8 is a diagram showing a two-dimensional image in the case where there is no illuminance unevenness in the width direction of the region to be inspected and there is illuminance unevenness in the running direction, a diagram showing illuminance unevenness LA in the running direction A, and a two-dimensional image. It is a figure which shows the profile of the traveling direction integrated data which integrated the image data showing each pixel in the traveling direction.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Inspected material 2 Illumination device 3 Two-dimensional CCD camera 4 Data integration means 5 Data distortion correction means 6 Surface defect detection means 10 Surface defect inspection device 13 Inspected area 13a Reflected light 14, 15, 16, 17, 18, 19 Profile 30, 30a, 30b Defects 31a, 31b, 31c, 31d, 31e Peak value

Claims (2)

[Claims] 1. Illumination of a predetermined inspected region on a surface of an inspected material traveling in a predetermined traveling direction with a uniform illuminance in the traveling direction as compared with a width direction in which the illuminance in the inspected region intersects the traveling direction. Then, the reflected light from the inspection area is received by a two-dimensional imager to generate two-dimensional image data representing each pixel of a two-dimensional image of the inspection area, and the two-dimensional image obtained by the two-dimensional imager Image data is integrated in the traveling direction of the inspection material to generate traveling direction integration data, and the distortion of the data included in the generated traveling direction integration data, which is caused by unevenness in the illuminance in the width direction of the inspection region, is corrected. A surface defect inspecting method, which comprises correcting and correcting the distortion to detect a surface defect extending linearly along the running direction of the material to be inspected. 2. Illuminating a predetermined inspected region on the surface of the inspected material traveling in a predetermined traveling direction with a uniform illuminance in the traveling direction as compared with a width direction in which the illuminance in the inspected region intersects the traveling direction. An illuminator, a two-dimensional image pickup device that receives reflected light from the inspection region, and generates two-dimensional image data representing each pixel of a two-dimensional image of the inspection region; Data accumulating means for accumulating the two-dimensional image data in the traveling direction of the inspected material to generate traveling direction accumulating data, and the inspecting area included in the traveling direction accumulating data obtained by the data accumulating means. Data distortion correction means for correcting data distortion caused by uneven illuminance in the width direction, and surface defects extending linearly from the data whose distortion has been corrected by the data distortion correction means along the running direction of the inspected material. Detect Surface defect inspection apparatus is characterized in that a surface defect detecting means that.