JP2002310917A - Defect detecting method and device thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To secure sufficient detection accuracy even through the use of a relatively inexpensive aperture fluorescent lamp as an illuminating device in a defect detecting method and device. SOLUTION: The optical axis of the aperture fluorescent lamp 2 is inclined at an angle from the direction perpendicular to the horizontal direction to an object to be inspected 1 within a range in which illuminating light from the aperture fluorescent lamp 2 is not directly incident onto an image pickup camera 3. A point of view P of the image pickup camera 3 is set at an end of an irradiation range W on the side of the image pickup camera on the object to be inspected irradiated by the aperture fluorescent lamp 2 or within a predetermined range Wp in the vicinity of the end.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect detection technique for irradiating an object to be inspected with light and detecting a surface defect based on a reflected image thereof, and more particularly to a defect detection technique using an aperture fluorescent lamp as a lighting device. About.

[0002]

2. Description of the Related Art Conventionally, in the inspection of defects on a surface such as a painted surface, a defect detecting apparatus has been used which can automatically detect a defect of surface irregularities by combining an image processing technique with a reflection optical system. This defect detection device usually includes an illumination device that irradiates a linear light beam onto the surface of the inspection object, an imaging camera that receives reflected light from the inspection object to capture a reflected image, and a signal (a signal from the imaging camera). Image information)
And an image processing apparatus for detecting a defect on the surface of the inspection object by performing image processing on the object.

In the above defect detection apparatus, the luminance (intensity) of the reflected light on the surface is sent to the image processing apparatus as an image signal, and the defect is detected based on the luminance of the reflected light. Therefore, in order to make it easy to recognize the image signal of the defective portion and to be able to detect even a fine defect, it is necessary to increase the gradation difference of the reflected light between the defective portion and the normal portion,
For this purpose, it is important to suppress diffuse reflection on the surface and increase the reflectance.

Therefore, a conventional defect detection device uses an illumination device that irradiates light from a halogen lamp or the like using an optical fiber light guide or a rod unit, and applies a linear light beam from this illumination device to the surface of the inspection object. On the other hand, it was irradiated from a low angle. Optical fiber light guides and rod units use a condenser lens to reduce the irradiation light distribution angle to 1
It can be set to about 0 deg, and can emit substantially parallel light rays. Therefore, according to these optical fiber light guides and the like, irregular reflection of the surface can be suppressed by irradiating the parallel rays in this way, and the reflectance of the surface can be increased by irradiating the rays from a low angle, and the It is possible to increase the gradation difference of the reflected light between the portion and the defective portion.

[0005]

The optical fiber light guide is of a line type in which the tips of a plurality of optical fibers are arranged in a line to irradiate a linear light beam. A length of about 500 mm is standard. Therefore, if the inspection object is 1
If the inspection object requires an illumination width of 00 to 500 mm, a commercially available optical fiber light guide can be used as the illumination device.

However, the defect detection apparatus is also applied to a surface inspection of a wide inspection object such as an interior panel of a building, and an optical fiber light guide having a long illumination width for such a wide inspection object. Is required. For example, since the interior panel has a width of about 2 m, an optical fiber light guide having an illumination width equivalent to at least 2 m is required. However, an optical fiber light guide that greatly exceeds such a standard size is extremely expensive, and often cannot be adopted due to cost.

The rod unit has a structure in which light is guided to a rod-shaped quartz glass rod via an optical fiber, and a linear light beam is irradiated from a side surface of the quartz glass rod. About 1 m is a standard length. Therefore, if the rod unit is applied to an inspection object having a width of about 2 m, such as the above-mentioned interior panel, it will be custom-made and extremely expensive like an optical fiber light guide. Furthermore, in the case of a rod unit, the illuminance in the length direction is not uniform and is 5 to 10%.
There is also a disadvantage that the degree of unevenness (illuminance unevenness) occurs.

On the other hand, a rod-shaped fluorescent lamp has been conventionally used as a general linear illumination. Fluorescent lamps have the advantage that they are inexpensive and do not have cost restrictions on their length, such as optical fiber light guides. Another advantage is that the uniformity of the illuminance in the length direction is extremely high. Therefore, it is conceivable to use a fluorescent lamp as an illumination device of the defect detection device as a measure for solving the above problem. in this case,
Since a general fluorescent lamp has a structure that irradiates light in all directions of 360 degrees, it cannot irradiate parallel rays unless it is located far from the inspection object. Irradiates light directly into the imaging camera. For this reason, an aperture fluorescent lamp with a limited light distribution direction is actually used.

However, the aperture distribution angle of an aperture fluorescent lamp is limited to 40 deg due to manufacturing problems. Therefore, the parallelism of light rays on the optical axis is lower than that of an optical fiber light guide having a light distribution angle of about 10 deg. Low.
Also, since the light distribution angle is at least 40 deg,
The optical axis angle of the aperture fluorescent lamp for preventing the irradiation light from directly entering the imaging camera is limited to at most 40 deg, and the optical axis angle cannot be greatly inclined like an optical fiber light guide. Therefore, simply replacing the optical fiber light guide or the like with an aperture fluorescent lamp in the conventional defect detection apparatus will result in a detection accuracy (sensitivity).
Will be reduced.

As described above, the aperture fluorescent lamp is inferior to the optical fiber light guide or the like in terms of light beam parallelism and low irradiation angle in a simple comparison, but is superior in cost and uniformity of illuminance as described above. ing. Therefore, even with an aperture fluorescent lamp, it is possible to achieve the same degree of parallelism and low irradiation angle of light rays as an optical fiber light guide, etc., even with an aperture fluorescent lamp. I want to be able to adopt it as a lighting device.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and provides a defect detection method and apparatus capable of securing sufficient detection accuracy while using a relatively inexpensive aperture fluorescent lamp as an illumination device. The purpose is to provide.

[0012]

According to the present invention, an inspection object is illuminated with light from an aperture fluorescent lamp, a reflected image of the illuminated inspection object is picked up by an image pickup camera, and information of the picked-up image is obtained. In order to detect the defect on the surface of the inspection object based on the above, the above object can be achieved by devising the light distribution direction of the aperture fluorescent lamp and the position of the viewpoint by the imaging camera.

That is, according to the defect detection method of the present invention, the optical axis of the aperture fluorescent lamp is inclined from the vertical direction to the horizontal direction with respect to the inspection object within a range where the irradiation light from the aperture fluorescent lamp does not directly enter the imaging camera. In addition, the viewpoint of the imaging camera is set at a predetermined range near or near the imaging camera end of the irradiation range on the inspection object irradiated by the aperture fluorescent lamp.

By tilting the optical axis of the aperture fluorescent lamp as described above, the irradiation angle with respect to the inspection object is reduced and the reflectance is increased, and the position of the viewpoint by the imaging camera is imaged in the irradiation range as described above. By setting the camera side end portion or a predetermined range near the end portion, the substantial irradiation angle at the viewpoint is further reduced, and the reflectance is further increased. At the same time, since the angle between the aperture of the aperture fluorescent lamp and the viewpoint is reduced to a minimum angle or close to the minimum angle, the parallelism of the light beam irradiated to the viewpoint increases.

In this case, as the position of the viewpoint is closer to the imaging camera than the optical axis, the substantial irradiation angle at the viewpoint becomes smaller, and the narrower the angle between the opening of the aperture fluorescent lamp and the viewpoint becomes, the closer the viewpoint becomes. Since the parallelism of the irradiated light beam becomes high, the position of the viewpoint of the imaging camera is the optimum position at the imaging camera side end of the irradiation range. However, since it is difficult to adjust the position of the viewpoint to the end of the imaging camera in this way, it may be substantially in a predetermined range near the end, and preferably, the angle between the opening of the aperture fluorescent lamp and the viewpoint. Is within a range of 6 deg or less. Within such a range, light rays that can be regarded as substantially parallel light rays can be applied to the viewpoint. The light distribution angle of the aperture fluorescent lamp is not particularly limited, but is preferably 40 deg.
An angle of about or less is used.

Further, when the above-described defect detection method is used particularly for defect detection using a painted surface as an object to be inspected, the following first and second images are used with respect to the reflected image picked up by the image pickup camera. It is preferable to detect the defect by performing processing. First, in the first image processing, a processing region is extracted from image information captured by an imaging camera, and ± 3.5 to 4.5σ of the average luminance or the mode luminance of the extracted processing region.
Is a process of performing binarization processing of image information using a predetermined threshold value within the range of, and detecting a defect based on a predetermined feature amount of the binarized particles obtained by the binarization processing. By performing such a process, fine unevenness defects having a thickness of several μm existing on the surface of the painted surface are detected with high accuracy.

In the second image processing, a processing area is extracted from image information picked up by an image pickup camera, and a predetermined threshold value within a range of 25 to 35% of the average luminance or the mode luminance of the extracted processing area. Is a process for performing a binarization process on image information using, and detecting a defect based on a predetermined feature amount of the binarized particles obtained by the binarization process. By performing such processing, a defect due to a paint lump existing on the painted surface is detected with high accuracy.

The feature values used in the first and second image processes determine whether the binarized particles obtained by the binarization process are normal or defective, and classify the defect type. The size, area, and size of the binarized particles
It is preferable to use one or a combination of the maximum / minimum luminance, the aspect ratio, the area ratio, and the number. In particular, it is preferable to use the area and the number of the binarized particles as the feature amount for detecting the defect due to the paint mass in the second image processing.

The present invention also provides an apparatus to which the above-described defect detection method can be applied. The defect detection device includes an aperture fluorescent lamp that irradiates the inspection object with light, an imaging camera that captures a reflection image of the inspection object that has been irradiated with light, and a surface of the inspection object that processes captured image information. Image processing means for detecting the defect of the defect, the optical axis of the aperture fluorescent lamp from the direction perpendicular to the inspection object in a range where the irradiation light from the aperture fluorescent lamp does not directly enter the imaging camera It is set obliquely toward the horizontal direction,
It is characterized in that the viewpoint of the imaging camera is set to a predetermined range near the imaging camera end or near the end of the irradiation range on the inspection target irradiated by the aperture fluorescent lamp.

It is preferable that the viewpoint is set at a position where the angle between the viewpoint and the opening of the aperture fluorescent lamp is 6 deg or less.

[0021]

Embodiments of the present invention will be described below with reference to the drawings. First, the configuration of a defect detection device according to an embodiment of the present invention will be described with reference to FIG. As shown in FIG. 1, the defect detection device according to the present embodiment includes an aperture fluorescent lamp 2 arranged toward a surface (inspection surface) 1 of an inspection object, and a line sensor similarly arranged toward the inspection surface 1. Camera: A camera having a plurality of (for example, 5000) CCD elements arranged side by side in the width direction, hereinafter simply referred to as a camera # 3, and an image processing device 4 connected to the line sensor camera 3.

In the present invention, the light distribution direction of the aperture fluorescent lamp 2 and the position of the viewpoint on the inspection surface 1 by the camera 3 are important. In the present embodiment, the light distribution angle θ of the aperture fluorescent lamp 2 is 40 deg, opening 2a
The optical axis is inclined from the vertical direction to the horizontal direction with respect to the inspection surface 1 so as to face the camera 3 side. The viewpoint P of the camera 3 is set in a predetermined range Wp near the camera 3 side end of the irradiation range W in the irradiation range W on the inspection surface 1 irradiated by the aperture fluorescent lamp 2.

The above-mentioned predetermined range Wp is a range in which the included angle γ between the opening 2a of the aperture fluorescent lamp 2 and the viewpoint P is 6 deg or less. When the viewpoint P is set within the predetermined range Wp, if the illumination height HL of the aperture fluorescent lamp 2 is 50
~ 200mm, distance Lc between camera 3 and viewpoint P is 600
１０００1000 mm, the light receiving angle α of the camera 3 is 55
The angle β of the optical axis can be set in the range of 20 to 40 deg.
However, in order to improve the reflectance by a low irradiation angle, the angle β of the optical axis is preferably set to be as large as possible as long as the irradiation light from the aperture fluorescent lamp 2 does not directly enter the camera 3.

According to such an arrangement, by inclining the optical axis of the aperture fluorescent lamp 2, the irradiation angle with respect to the inspection surface 1 is reduced and the reflectance is increased. By setting within the predetermined range Wp, the substantial irradiation angle φ at the viewpoint P is further reduced, and the reflectance is further improved. At the same time, the angle γ between the opening 2a of the aperture fluorescent lamp 2 and the viewpoint P is 6 de.
g, the light beam applied to the viewpoint P is a substantially parallel light beam having a variation of 6 deg or less.

FIGS. 2 and 3 are schematic diagrams showing the relationship between the brightness of the reflected light and the shape of the inspection surface 1. FIG. Here, as shown in FIG. 2A, the viewpoint P of the camera 3 is shifted to the irradiation range W.
Is set at the end of the camera 3 (the boundary between bright and dark). The light distribution angle θ of the aperture fluorescent lamp 2 is 40 deg similarly to FIG. 1, and the angle of the optical axis is 33 deg.
Is set to Further, the optical axis of the camera 3 is inclined in the horizontal direction with respect to the inspection surface 1 with respect to the specular reflection direction at the bright and dark boundaries, and the light receiving angle α is set to 65 deg here.

Although the light emitted from the aperture fluorescent lamp 2 is scattered and reflected on the inspection surface 1, the scattered reflected light at the viewpoint P has a drop-shaped luminance distribution (reflection) centered on the regular reflection direction as shown in the figure. (Light distribution curve). Therefore, the brightness of the reflected light is strongest in the specular reflection direction, and becomes weaker as the light deviates from the specular reflection direction. For this reason, by setting the optical axis of the camera 3 so as to be inclined from the regular reflection direction on the horizontal plane as described above, the camera 3 can be used as shown in FIGS. 2B, 2C and 3.
As shown in FIGS. 3A and 3B, reflected light having a luminance corresponding to the shape of the inspection surface 1 is incident on the camera 3.

More specifically, as shown in FIG. 2B, when the inspection surface 1 is inclined downward from the aperture fluorescent lamp 2 toward the camera 3, the regular reflection direction at the bright and dark boundaries is also inclined downward. The angle between the reflection direction and the optical axis direction of the camera 3 becomes narrow. As a result, the luminance of the reflected light incident on the camera 3 becomes as shown in FIG.
The luminance of the reflected light from the horizontal reference plane shown in FIG. Conversely, when the inspection surface 1 is inclined upward from the aperture fluorescent lamp 2 toward the camera 3 as shown in FIG. The angle with respect to the optical axis direction of the camera 3 opens. Thereby, the luminance of the reflected light incident on the camera 3 is lower than the luminance of the reflected light from the horizontal reference plane shown in FIG.

When the inspection surface 1 is higher than the reference surface as shown in FIG. 3A, the specular reflection direction at the bright / dark boundary moves upward in parallel, and the axis of the specular reflection direction and the camera are shifted. 3
Away from the optical axis of Thereby, the luminance of the reflected light incident on the camera 3 is lower than the luminance of the reflected light from the horizontal reference plane shown in FIG. Conversely, when the inspection surface 1 is lower than the reference surface as shown in FIG. 3B, the specular reflection direction at the bright / dark boundary moves downward in parallel, and the axis of the specular reflection direction and the camera 3 The optical axis approaches. Thus, the luminance of the reflected light incident on the camera 3 is higher than the luminance of the reflected light from the horizontal reference plane shown in FIG.

As described above, the luminance of the reflected light incident on the camera 3 changes according to the shape of the inspection surface 1. Since the signal (image information) transmitted from the camera 3 corresponds to the luminance of the reflected light, the signal from the camera 3 is subjected to the binarization processing in the image processing device 4 using an appropriate threshold. Thus, a defect on the inspection surface 1 can be detected. In particular, according to the defect detection device according to the present embodiment, the inspection surface 1 can be irradiated with a parallel light beam at a low irradiation angle as in the illumination device using the conventional optical fiber light guide or the like as described above. In addition, it is possible to suppress the spread of the reflection light distribution curve and increase the reflectance in the specular reflection direction, and to detect a fine defect by increasing the gradation difference of the reflected light between the defective portion and the normal portion on the inspection surface 1. Can be.

Next, an image processing method by the image processing apparatus 4 will be described. The image processing device 4 identifies a normal part and a defective part on the inspection surface 1 by performing binarization processing of image information with an appropriate threshold. The threshold value is set according to the gradation difference between the normal part, the defective part, and the luminance. In general, the smaller the gradation difference, the closer the threshold to the luminance of the normal part.

However, when the inspection surface 1 is a painted surface, such as when a painted interior material is an object to be inspected, the painted surface has relatively large surface roughness due to the paint. Even in the normal part, the dispersion of the luminance of the reflected light is large. Therefore, if the threshold value is set near the luminance of the normal part so that even a minute defect can be detected, not only the defective part but also a part of the normal part is binarized to the defect side, and the normal part is erroneously recognized as a defect. There is a risk of doing this.

In addition, the defects on the painted surface include, in addition to those having irregularities on the surface of the substrate serving as a base (irregularities), painting defects due to a lump of paint. Painting defects due to this paint lump are usually treated separately from other fine unevenness defects, but simply by binarizing, coating defects are recognized as normal defects like other unevenness defects. I will. Conventionally, this coating defect has been visually detected by a human, but it is desired that the coating defect be automatically detected in the same manner as the fine unevenness defect.

Therefore, in the present embodiment, the image processing device 4
By performing the defect detection processing in the above with the following algorithm, accurate detection of minute unevenness defects and identification of unevenness defects and paint defects were made possible. First, FIGS. 4 and 5
The outline of the defect detection processing according to the present embodiment will be described with reference to FIG. FIG. 4 shows an example of the luminance distribution of the painted surface. As shown in FIG. 4, in this embodiment, the threshold value used for defect detection differs between the unevenness defect and the painting defect. That is, when the average luminance is m and the luminance variance is σ, the threshold value is m ± 3.5-4.5σ for the unevenness defect. In the conventional method for detecting a defect on a painted surface, the threshold value is set to m ± 5 to 6σ. However, in the present embodiment, the threshold value is set close to the average luminance so that a finer defect can be detected. This is prevented by a determination process based on the amount. On the other hand, for paint defects,
Since the painted lump looks extremely dark against the background, an extremely low value of about 25 to 35% of the average luminance m is set as the threshold. Note that the mode luminance may be used instead of the average luminance as a reference for setting the threshold.

By performing the binarization processing while changing the threshold value as described above, the binarized particles corresponding to the fine irregularities and the coating defects are respectively extracted. However, the binarized particles obtained by the binarization process are only candidates for unevenness defects and painting defects, and may include a part of a normal part. Therefore, in the present embodiment, the unevenness defect and the coating defect are respectively detected from the binarized particles by the image processing based on the feature amount indicating the shape of each defect.

For example, FIG. 5 is a schematic diagram showing an example of the shape of a representative one of the surface defects. As shown in this figure, each defect has a feature in shape, and by performing image processing based on a quantified feature (feature amount), it is possible to identify a normal part. Table 1 below shows the relationship between the type of defect and the corresponding threshold value and feature amount. In the table below, the coating defect is a defect due to a coating lump, and the surface defects A to L correspond to unevenness defects. As shown in this table, each defect corresponds to the size, area, and number of binarized particles, maximum luminance, minimum luminance, and aspect ratio 縦 = min (vertical length,
An appropriate combination of (horizontal length) / max (vertical length, horizontal length) 長 and area ratio ｛= area / (vertical length × horizontal length)｝ is used as a feature amount to enable discrimination from a normal part. In particular, a coating defect can be distinguished from a normal part by using the area and the number of binarized particles as characteristic amounts.

[0036]

[Table 1]

Next, the defect detection processing according to the present embodiment will be described more specifically with reference to the flowcharts of FIGS. First, as shown in FIG. 6, the image processing device 4
Image information (luminance signal) is received from the camera 3, and the received image information is stored in the memory (step S11).
0). Then, when image information (for example, 2048 pixels × 1000 lines) corresponding to a predetermined area on the painted surface is accumulated, the image processing after step S120 is started.

In step S120, the image processing device 4
Performs histogram processing on the image obtained in step S110, and calculates the average luminance m, the most frequent luminance,
And the standard deviation σ are calculated. Note that this histogram processing may be performed on the entire image or on some specific lines. Subsequently, the image processing device 4
Performs shading correction of the image based on the average luminance m in step S130. Then, step S14
At 0, a binarization process is performed with a threshold value (m-4σ) to extract particles that are candidates for a defect darker than a normal part (hereinafter, referred to as a black defect). In step S150, binarization processing is performed using a threshold value (m + 4σ) to extract particles that are candidates for a defect brighter than a normal part (hereinafter, referred to as a white defect).

Next, the image processing device 4 executes step S16.
0, noise processing of the image is performed, and step S14
0, noise is removed from the particles extracted in S150. Then, labeling is performed on the particles remaining after the noise removal (step S170), and particles whose inter-particle distance is within a predetermined distance are regarded as the same particle and integrated (step S170).
180). Then, for each particle, a coordinate, an area, an aspect ratio, an aspect ratio, an area ratio, and a minimum / maximum luminance from the average are calculated (step S190).

When the above preprocessing is completed, the image processing apparatus 4 first performs a coating defect detection process according to the flow shown in FIG. In step S210, the image processing device 4
Determines whether there is a particle whose minimum luminance is smaller than the threshold value (m-0.7 m). If there is a particle that satisfies the condition of step S210, the threshold value (m−0.7) is set for the particle that is a candidate for a black defect extracted in step S140.
m), the binarization processing is performed again (step S220), and subsequently, the area of each particle extracted by the binarization processing is calculated (step S230). Then, in step S240, X0 or more particles satisfy the relationship of (predetermined value A0 ≦ area <predetermined value A1) or particles satisfying the relationship of (predetermined value A1 ≦ area) are X1 (<X0 It is determined whether or not there is at least one particle. If there is a particle that satisfies the condition of step S240, it is determined that the particle is a coating defect.

Next, the image processing device 4 performs detection processing of black defects other than the coating defects, that is, surface defects A to G according to the flow shown in FIG. First, in step S310, {predetermined value S0 ≦ particle size <predetermined value S1, and −
It is determined whether there is a particle satisfying the relationship of 0.7m ≦ (minimum luminance−m) <predetermined value D｝, and if so, the particle is determined to be any black defect. Next, step S320
Then, Δpredetermined value S1 ≦ particle size <predetermined value S2, and
It is determined whether there is a particle that satisfies the relationship of (minimum luminance−m) <predetermined value D｝, and if so, that particle is also determined to be any black defect. Further, in step S330,
It is determined whether there is a particle satisfying the relationship of (predetermined value S2 ≦ particle size), and if so, the particle is also determined to be any black defect.

Next, the image processing apparatus 4 performs detection processing of surface defects H to L, which are white defects, according to the flow shown in FIG. First, in step S410, it is determined whether or not there are X2 or more particles having a predetermined value X satisfying the relationship of (particle size ≦ predetermined value S4).
Is determined. Next, in step S420, (predetermined value S
It is determined whether there is a particle satisfying the relationship of 4 ≦ particle size and aspect ratio <predetermined value R), and if so, the particle is also determined to be a surface defect H. Further, step S430
Then, it is determined whether or not there is a particle satisfying the relationship of (predetermined value S4 ≦ particle size and area ratio <predetermined value B), and if so, the particle is determined to be a surface defect K. And
In step S440, the same determination processing as in steps S310 to S330 of FIG. 8 is performed. If there is a particle that satisfies the determination condition, the particle is determined to be another white defect. As a result of the above processing, the inspection object (painted surface) on which no defect is detected is determined as a non-defective product, and a series of defect detection processing is terminated.

The surface roughness Ra of the painted surface is 0.2 to 1.0 μm
In contrast, the irregularities on the painted surface as shown in FIG. 5 are minute with an inclination angle of the surface irregularities of 0.2 to 1.0 deg and a height difference of about several μm. By acquiring the image information of the painted surface using the optical system shown in FIG. 1 and processing the image information by the above-described algorithm, it becomes possible to accurately detect even such minute defects. Further, by performing the binarization processing while changing the threshold value, it is also possible to accurately identify the unevenness defect on the painted surface and the paint defect without being confused.

As described above, one embodiment of the present invention has been described. However, the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit of the present invention. Needless to say. For example, the defect detection method and apparatus of the present invention are not limited to the detection of a defect on a painted surface as in the embodiment, but can be applied to the detection of a defect on a resin sheet, a metal surface, or the like. According to the present invention, a parallel light beam can be applied to the surface of the inspection object at a low irradiation angle in the same manner as a conventional illumination device using an optical fiber light guide or the like. Detection accuracy can be ensured.

[0045]

As described above in detail, according to the defect detection method and apparatus of the present invention, the optical axis of the aperture fluorescent lamp is inspected within the range where the irradiation light from the aperture fluorescent lamp does not directly enter the imaging camera. To the horizontal direction from the vertical direction, and set the viewpoint of the imaging camera at a predetermined range near or near the imaging camera side end of the irradiation range on the inspection target irradiated by the aperture fluorescent lamp. This makes it possible to irradiate the surface of the inspection object with a light beam that is nearly parallel to the surface of the inspection object at a low irradiation angle. Therefore, it is possible to increase the reflectance on the surface and increase the gradation difference between reflected light between the defective portion and the normal portion. Therefore, there is an effect that even a minute defect can be accurately detected.

In particular, by setting the angle between the aperture of the aperture fluorescent lamp and the viewpoint to be 6 deg or less, it is possible to irradiate the viewpoint with a light beam that can be regarded as a substantially parallel light beam. When the inspection object requires an illumination width equivalent to 2 m, by employing an aperture fluorescent lamp by the defect detection method and apparatus of the present invention, the illuminance uniformity in the width direction is extremely high, and a substantially parallel linear illumination means. Can be obtained at low cost.

When the painted surface is to be inspected, the binarization processing of the image information is performed by using a predetermined threshold value within a range of ± 3.5 to 4.5σ of the average luminance or the mode luminance. By performing the defect detection based on the predetermined feature amount of the binarized particles, it is possible to detect the unevenness defect of the painted surface with high accuracy, and the average luminance or the mode luminance is within the range of 25 to 35%. By performing binarization processing of image information using a predetermined threshold value and detecting a defect based on a predetermined feature amount of the binarized particles, a coating defect on a painted surface can be detected with high accuracy . In this case, by using any or a combination of the size, area, maximum / minimum brightness, aspect ratio, area ratio, and number of the binarized particles as the feature amount, the unevenness defect and the coating defect can be detected with higher accuracy. Can be.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a defect detection device according to an embodiment of the present invention.

2A and 2B are schematic diagrams illustrating a relationship between the brightness of reflected light and the shape of an inspection surface according to an embodiment of the present invention, where FIG. 2A illustrates a reflection state on a horizontal plane (reference plane), and FIG. FIG. 7C is a diagram showing a reflection state on an inclined surface inclined downward, and FIG. 7C is a diagram showing a reflection state on an inclined surface inclined upward.

3A and 3B are schematic diagrams illustrating a relationship between the luminance of reflected light and the shape of an inspection surface according to an embodiment of the present invention, where FIG. 3A illustrates a reflection state on a surface higher than a reference surface, and FIG. () Is a diagram showing a reflection state on a surface lower than the reference surface.

FIG. 4 is a luminance distribution diagram for explaining an outline of a defect detection process according to an embodiment of the present invention.

FIG. 5 is a schematic diagram showing a typical example of the shape of a surface irregularity defect.

FIG. 6 is a flowchart showing a flow of a defect detection process according to an embodiment of the present invention.

FIG. 7 is a continuation of the flowchart of FIG. 6, and is a flowchart showing a flow of a coating defect detection process.

8 is a continuation of the flowchart of FIG. 7, and is a flowchart showing the flow of a process for detecting a defect (black defect) on a painted surface.

FIG. 9 is a continuation of the flowchart of FIG. 8, and is a flowchart showing a flow of detection processing of a defect (white defect) on a painted surface.

[Explanation of symbols]

 Reference Signs List 1 inspection surface 2 aperture fluorescent lamp 2a opening 3 line sensor camera (imaging camera) 4 image processing device P viewpoint θ light distribution angle α light receiving angle of camera β optical axis angle of aperture fluorescent lamp γ aperture of fluorescent lamp and viewpoint Φ Angle of actual irradiation to viewpoint W Irradiation range Wp Setting range of viewpoint HL Illumination height Lc Distance between camera and viewpoint

Continued on the front page F-term (reference) 2F065 AA49 AA61 CC31 DD02 DD04 DD09 FF04 FF44 GG03 GG16 HH02 JJ03 JJ05 JJ08 JJ26 QQ06 QQ21 QQ32 QQ43 SS04 SS13 UU03 UU05 UU07 2G051 AB07 AB12 BA20 BB01 CB01 CA04 CB01

Claims (7)

[Claims] 1. An inspection object is illuminated with light from an aperture fluorescent lamp, a reflection image of the inspection object irradiated with light is captured by an imaging camera, and the surface of the inspection object is captured based on image information captured. A method for detecting the defect of the above, wherein the optical axis of the aperture fluorescent lamp is directed from the vertical direction to the horizontal direction with respect to the inspection object within a range where the irradiation light from the aperture fluorescent lamp does not directly enter the imaging camera. And setting the viewpoint of the imaging camera in a predetermined range near the imaging camera end or the end of the irradiation range on the inspection object irradiated by the aperture fluorescent lamp. Characteristic defect detection method. 2. The defect detection method according to claim 1, wherein the predetermined range is a range in which an angle between the opening of the aperture fluorescent lamp and the viewpoint is 6 deg or less. 3. A processing area is extracted from the image information, and the average brightness or mode brightness of the extracted processing area is ± 3.5 to ± 3.5.
A binarization process of the image information is performed using a predetermined threshold value within a range of 4.5σ, and a surface irregularity defect is detected based on a predetermined feature amount of the binarized particles obtained by the binarization process. 3. The defect detection method according to claim 1, wherein: 4. The inspection object is a painted surface, a processing area is extracted from the image information, and a predetermined threshold value within a range of 25 to 35% of the average luminance or the mode luminance of the extracted processing area is used. Performing a binarization process on the image information, and detecting a coating defect due to a paint mass based on a predetermined feature amount of the binarized particles obtained by the binarization process. 13. The defect detection method according to claim 1. 5. The method according to claim 1, wherein the characteristic amount is a size of the binarized particle,
4. The method according to claim 3, wherein the information is any one or a combination of an area, a maximum / minimum luminance, an aspect ratio, an area ratio, and a number.
Or the defect detection method according to 4. 6. An aperture fluorescent lamp for irradiating light to the inspection object, an imaging camera for capturing a reflected image of the inspection object irradiated with light, and an image camera for processing the image information captured and processing the image information. A defect detection device comprising: image processing means for detecting a surface defect, wherein the optical axis of the aperture fluorescent lamp is the inspection target in a range where irradiation light from the aperture fluorescent lamp does not directly enter the imaging camera. The viewpoint is set obliquely from the vertical direction to the horizontal direction with respect to the object, and the viewpoint by the imaging camera is an end of the irradiation range on the inspection object irradiated by the aperture fluorescent lamp on the imaging camera side or A defect detection device, wherein the defect detection device is set in a predetermined range near the end. 7. The defect detection apparatus according to claim 6, wherein the viewpoint is set in a range where an angle between the viewpoint and the opening of the aperture fluorescent lamp is 6 deg or less.