JPH109837A - Surface flaw inspection apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide surface flaw inspection apparatus enabling early prevention of occurrence of a costing defect even when a surface to be inspected is a complicated curved surface. SOLUTION: This equipment has an illuminating means 1 which is shaped like a gate surrounding a body 5 bending an object of inspection and forms a prescribed light-and-shaped pattern on a surface to be inspected, an image pickup device fixing means 2 which is fitted with a plurality of image pickup device CCD cameras 3 forming a received light image on the basis of reflected light from the surface to be inspected, an inspection processing means 4 which detects a flaw on the surface on the basis of the received light image and outputs the result of the detection, a flaw number statistically processing means 11 which executes a statistic processing on the basis of information on the detected flaw from the inspection processing means and information on a surface to be coated and a flaw causing source estimating means 12 which estimates a causing source of the flaw from the data from flaw number statistically processing means. The body is passed through the inside of the gate formed by the illuminating means and the image pickup device fixing means and flaw inspection by the inspection processing means and the estimation of the flaw causing source by the flaw causing source estimating means are conducted.

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 apparatus used for detecting surface defects such as irregularities on a painted surface of an automobile body.

[0002]

2. Description of the Related Art As a conventional surface defect inspection apparatus, for example, there is one described in Japanese Patent Application Laid-Open No. 64-38638.

In the apparatus described in the publication, a band of light is formed on a surface to be inspected of an object to be inspected, and the band of light is imaged by a camera, and the band of light is moved to reflect the image. Is recorded continuously and stepwise, and finally, the recording of these partial images is edited into the whole image, and the defect on the surface to be inspected and the coordinate information of this defect are output.

[0004]

However, in the conventional surface defect inspection apparatus as described above, for example,
When the inspection target has a complicated curved surface such as a car body, the reflection direction of the light band changes according to the curvature at the curved portion, so that the reflected image is always projected on the image center of the camera. Is required, and since the curved surface of the body of the automobile differs for each part or vehicle type, there is a problem that the control becomes more complicated, and it has been a problem to solve such a problem.

The detection of a defect utilizes the fact that it appears as a change in the contour of an image in a dark part or a light band in a reflection image of the light band. A method of automatically and stably detecting the defect is used. There is a problem that no consideration is given to the device and the device, and it has been a problem to solve such a problem.

Further, in the current automatic painting line for automobiles and the like, the detection and the statistical processing of the coating defects are carried out by humans with a lot of man-hours and time. Since it takes a very long time to specify, there is a problem that it is difficult to suppress the occurrence of a coating defect at an early stage, and there is a problem to solve such a problem.

[0007]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned conventional problems. (1) Even if a surface to be inspected is a complicated curved surface, surface defects can be automatically and accurately detected by simple control. High-speed processing and detection can be performed well. For example, a coating defect on a painted surface of an automobile or the like can be detected automatically and accurately in a short time. (2) Various kinds of bodies and parts and a plurality of Even for complex automatic coating lines such as automobiles that use various types of paints and coating machines, quickly process the automatic detection results of coating defects and identify the source of the defects, so that It is an object of the present invention to provide a surface defect inspection apparatus capable of providing accurate information of a defect source.

[0008]

According to a first aspect of the present invention, there is provided a surface defect inspection apparatus for irradiating a surface of an object to be inspected with light and reflecting light from the surface to be inspected. A surface defect inspection apparatus that creates a light-receiving image based on the light-receiving image and detects a defect on the surface to be inspected based on the light-receiving image. Illumination means for forming a pattern; imaging apparatus fixing means having a portal shape surrounding an object to be inspected, and a plurality of imaging apparatuses for producing a light-receiving image based on light reflected from the surface to be inspected; Inspection processing means for detecting defects on the surface to be inspected based on the received light image obtained by the device and outputting the result, and statistics on the number of defects for statistical processing based on the defect detection information from the inspection processing means and the information on the surface to be coated Processing means and statistical processing means for number of defects Defect source estimating means for estimating the source of the defect based on the defect statistical processing result and empirical defect / defect information is provided. It is characterized in that it is configured to perform a defect inspection of a surface and to estimate a defect generation source, and such a configuration is a means for solving the above-described problem.

In the embodiment of the surface defect inspection apparatus according to the present invention, as described in claim 2, the portal shape of the illuminating means has a shape substantially conforming to the front contour in the moving direction of the inspection object. The illuminating means may include a plurality of light sources arranged at substantially equal intervals on a white background plate, and the light source may be inspected. On the surface side, a light diffusion sheet in which light transmitting parts and matte black parts are alternately arranged, and a light-dark pattern is formed on the surface to be inspected by passing light from a light source through the light transmitting part of the light diffusion sheet The light-diffusing sheet of the illuminating means is stretched over a mat-like black sheet guide having a portal shape surrounding the object to be inspected, as described in claim 4, wherein the sheet guide Is movable from the light source and the background plate It can be made of the formation.

Similarly, in the embodiment of the surface defect inspection apparatus according to the present invention, as set forth in claim 5, the gate-shaped shape of the imaging device fixing means is formed in a front contour in the moving direction of the inspection object. The image pickup device may be a CCD camera, and the entire field of view of each CCD camera may be in the moving direction of the object to be inspected. In addition to forming a continuous band along the front contour, the fields of view of adjacent CCD cameras overlap in a predetermined area, and the moving direction of the object to be inspected and the horizontal or vertical direction in the received image of the CCD camera are The light and dark pattern may be of a matched configuration, as described in claim 7.
A configuration in which the interval between the light and dark patterns changes based on the number of light and dark patterns reflected on the CCD camera can be adopted.

Similarly, in the embodiment of the surface defect inspection apparatus according to the present invention, the visual field adjustment, the focus adjustment and the overlap amount adjustment of the CCD camera are performed by moving the inspection object. It can be configured to use a reference model that has a shape that is substantially adapted to the frontal contour in the direction and that has lines (including grid lines or the like) or point figures drawn on the surface thereof. , Claim 9
The reference model may be configured to have a shape that substantially matches the maximum cross-sectional profile of the cross-section in the moving direction of the object to be inspected, as described in (10). As described above, the color of the part other than the line of the figure of the reference model is the color with the highest lightness in the paint color of the surface to be inspected of the object to be inspected. The lens aperture and the shutter speed may be adjusted. As described in claim 11, a test piece painted in the paint color of the surface to be inspected and larger than the camera field of view is attached to the surface of the reference model. The configuration may be such that the lens aperture and shutter speed of the CCD camera are adjusted while imaging the test piece while being in contact with the CCD camera.

[0012] Similarly, in an embodiment of the surface defect inspection apparatus according to the present invention, as described in claim 12, the inspection processing means is configured to perform spatial inspection on image data of a received image obtained by a plurality of imaging devices. An image enhancement processing means for extracting only a component having a level equal to or higher than a predetermined value in a high frequency region of the frequency components, and a movement amount and a movement direction of the object to be inspected are predetermined in image data temporally different from the image enhancement processing means. And a tracking processing unit that detects a target portion that matches under the following condition. As described in claim 13, the image emphasizing processing unit performs the processing in the extracted high frequency region. In addition, a configuration including a defect shape determination unit that determines the size (area) and shape (aspect ratio) of a component whose level is equal to or greater than a predetermined value can be employed.

[0013] Similarly, in an embodiment of the surface defect inspection apparatus according to the present invention, the inspection processing means includes an inspection start / end determination means for determining the start and end of the inspection; A movement amount measuring means for measuring a movement amount of the object to be inspected from the start of the inspection is provided. Based on information obtained from these means, a position of the target portion detected by the tracking processing means on the surface to be inspected is calculated. May be provided on the developed view of the object to be inspected. As described in claim 15, the developed view of the object to be inspected is an imaging device for the object to be inspected. The object to be inspected passes through the inside of the portal of the illuminating means and the imaging device fixing means, as described in claim 16. And inspect the inspected surface It may be of a configuration having a rate matching means for matching the speed of the conveyor moving the object to be inspected and the inspection object when that Tsu.

Similarly, in an embodiment of the surface defect inspection apparatus according to the present invention, as described in claim 17, the number-of-defects statistical processing means includes the type of the object to be inspected, the type of paint (color), and the like. 19. An apparatus according to claim 18, further comprising an inspection object information input means, wherein statistical processing is performed based on defect size, shape, and body position information from the inspection processing means. The source estimating means specifies a defect occurrence part based on defect statistical processing information (defect number) for each type, part, and defect type (size, shape) from the defect number statistical processing means and empirical defect occurrence information. It can be of a configuration.

[0015]

According to the surface defect inspection apparatus according to the first aspect of the present invention, the illuminating means having a portal shape surrounding the inspection object, and the imaging device having the portal shape surrounding the inspection object as well. An imaging device fixing means provided with a device is used to form a light / dark pattern of light on the surface to be inspected, and the object to be inspected is passed by the inspection processing means by passing the object to be inspected through the illuminating means and the portal type of the imaging device fixing means. Since the defect inspection of the surface and the estimation of the defect source by the defect source estimating means are performed, even when the inspected surface has a complicated curved surface, the illumination and imaging conditions for the entire inspected surface are substantially reduced. It is possible to detect surface defects automatically, accurately, and at high speed with simple control without performing any complicated control on the illuminating means, the image pickup device, and the inspected object. With many types Even for complex automatic painting lines such as automobiles that use bodies and parts and multiple painting machines and various types of paints, the results of automatic detection of painting defects should be statistically processed quickly and their sources identified. Accordingly, an extremely excellent effect that it is possible to provide accurate information of the defect source to the upstream line and to suppress the occurrence of the coating defect at an early stage is provided.

According to the surface defect inspection apparatus of the second aspect of the present invention, the portal shape of the illuminating means is substantially adapted to the front contour in the moving direction of the object to be inspected. In addition, the illumination conditions for the entire surface to be inspected can be made more uniform, and surface defects can be detected more easily and more accurately.

According to the surface defect inspection apparatus according to the third aspect of the present invention, in addition to the effects of the first and second aspects, a bright and dark pattern can be formed more accurately, and the detection of surface defects is further facilitated. And contribute to improvement of accuracy.

According to the surface defect inspection apparatus of the fourth aspect of the present invention, in addition to the effect of the third aspect, the maintenance of the sheet guide, the light source and the background plate can be easily dealt with by moving the sheet guide. be able to.

According to the surface defect inspection apparatus according to the fifth aspect of the present invention, the gate-shaped shape of the image pickup device fixing means is substantially adapted to the front contour in the moving direction of the object to be inspected.
In addition to the effects of the first aspect, the imaging conditions for the entire surface to be inspected can be made more uniform, and surface defects can be detected more easily and more accurately.

According to the surface defect inspection apparatus according to the sixth aspect of the present invention, in addition to the effects of the first and fifth aspects, it is possible to inspect the entire surface to be inspected without gaps, thereby further improving the accuracy of the inspection. Can be realized.

According to the surface defect inspection apparatus according to the seventh aspect of the present invention, in addition to the effect of the sixth aspect, the inspection accuracy can be further improved.

According to the surface defect inspection apparatus according to claims 8 to 11 of the present invention, in addition to the effects of claims 6 and 7,
By employing the reference model, the adjustment of the imaging device can be performed more easily and more quickly, and the inspection accuracy can be further improved.

According to the surface defect inspection apparatus of the twelfth aspect of the present invention, in addition to the effect of the first aspect, the surface defect can be detected more quickly and more accurately. According to the surface defect inspection device according to item 13,
The defect shape such as the size (area) and shape (aspect ratio) of the defect can also be determined.

According to the surface defect inspection apparatus of the present invention, in addition to the effect of the twelfth aspect, the surface defect can be easily recognized on the developed view of the inspection object, and the subsequent correction can be made. According to the surface defect inspection apparatus according to the fifteenth aspect of the present invention, in addition to the effect of the fourteenth aspect, in addition to the effect of the fourteenth aspect, the position of the surface defect on the developed view of the inspection object is provided. Can be displayed more accurately.

Further, according to the surface defect inspection apparatus of the present invention, in addition to the effects of the first, twelfth and fourteenth aspects, the speed of the object to be inspected and the speed of the conveyor can be surely matched. In addition, the inspection accuracy and the accuracy of displaying surface defects on a developed view can be further improved.

Further, according to the surface defect inspection apparatus of the present invention, statistical processing is performed based on defect information including size, shape, and position information detected by the inspection processing means and vehicle type, site, and paint information. This makes it possible to perform highly accurate surface defect inspection corresponding to the type of vehicle and the type of paint.

Further, according to the surface defect inspection apparatus according to the eighteenth aspect of the present invention, it is possible to specify a defect generation site, and to provide accurate information of a defect generation source to an upstream line. Can be.

[0028]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a surface defect inspection apparatus according to the present invention will be described below with reference to the drawings. In this example,
The figure shows a case where a defect of a painted surface, which is a surface to be inspected, is inspected in a body of an automobile, which is an object to be inspected.

As shown in FIGS. 1 to 3, the surface defect inspection apparatus according to this embodiment includes an illuminating unit 1, an imaging unit fixing unit 2, and a plurality of imaging units 3 attached to the imaging unit fixing unit 2. Inspection processing means 4, defect number statistical processing means 11
And a defect source estimating means 12.

The illuminating means 1 has a gate-shaped (arch) shape substantially adapted to the front contour of the body 5 in the moving direction, and is configured to project a predetermined light / dark pattern on the surface to be inspected. .

Further, the image pickup device fixing means 2 includes an illumination means 1
And are arranged in parallel with the illumination means 1 in substantially the same shape.

Further, the image pickup device 3 is a CCD camera (hereinafter referred to as "camera 3"), and is attached to a predetermined position of the image pickup device fixing means 2 so as to pick up an image of a surface to be inspected on which a light and dark pattern is reflected. It is fixed.

Further, the inspection processing means 4 mainly comprises an image enhancement processing means 41, a tracking processing means 42, a host computer 43 and the like.

Further, a printer is connected to the host computer 43 as an inspection result output means 44.

Further, the defect number statistical processing means 11
It is composed of vehicle type / paint information input means 111 and statistical processing means 112, and performs statistical processing based on defect information including size, shape, and position information detected by the inspection processing means 4 and vehicle type / part / paint information.

Further, the defect source estimating means 12 comprises:
The source of the defect is specified based on the statistical processing result calculated by the defect number statistical processing means 11 and empirical defect information on the coating line.

Further, the body 5 is placed on the trolley 6 and moves inside the portal of the illuminating means 1 and the image pickup device fixing means 2 by the rail 7 and the conveyor 8 while the predetermined processing is performed by the inspection processing means 4. The processing is performed, and the inspection of the painted surface of the body 5 which is the surface to be inspected is performed.

The illumination means 1 and the image pickup device fixing means 2
It is installed in a direction orthogonal to the transport direction of the body 5 (the direction of arrow A in FIG. 2). The camera 3 is fixed in a state where the camera 3 is adjusted to a predetermined position and angle of the imaging device fixing means 2 by the adjustment fixing jig 3a.

As shown in FIG. 1, the illuminating means 1 has a shape substantially conforming to the front contour in the moving direction of the body 5, so that the distance from the light irradiation surface to the surface of the body 5 is substantially independent of the position. The lighting means 1
Light from the body 5 can be applied to the surface of the body 5 evenly and almost uniformly.

As shown in FIGS. 4 and 5, the illuminating means 1 is colored white or diffusely reflects the light from the light source 101 to irradiate the surface to be inspected with no need and evenness. Background plates 10 surface-treated as described above
2 and a plurality of light sources 101 are attached to the background plate 102 at substantially equal intervals.

The light source 101 is a U-tube type fluorescent lamp.
A power supply 107 (see FIG. 7) for high-frequency lighting of the light source 101 is mounted on the rear side of the power supply. The light source 101, the background plate 102, and the power supply 107 constitute one illumination unit 104, and the illumination unit 1 is configured by attaching the illumination unit 104 to a column-shaped column 103 as shown in FIG. 5 without any gap. ing.

The light diffusing sheet 105 shown in FIG. 6 is made of a flexible and translucent material that can be easily deformed according to the front contour shape of the body 5. By applying a matte black masking tape at regular intervals, the light transmitting portion 105a and the matte black portion 10
5b are alternately arranged. Here, the reason for diffusing the light is to suppress the influence of the metallic glitter material when the body paint color is metallic paint or the like.

When the light diffusion sheet 105 is made of a material which is likely to be wrinkled, and the wrinkles cause shading or uneven illumination, the light diffusion sheet 105 is supported from below by a sheet guide 106 as shown in FIG. At the same time, the light diffusion sheet 105 is stretched in accordance with the front contour shape of the body 5 to suppress wrinkles.

Further, the sheet guide 106 is painted in matte black, and the support 103 is
5 are overlapped with the matte black portion 105b.
The light diffusion sheet 105 and the sheet guide 106 are fixed to the support pillar 103 by matte black bolts, nuts, and washers (not shown).

Further, a caster 106a is attached to the sheet guide 106, and the light diffusing sheet 105
The body 5 can be moved in the front-rear direction while being stretched. Accordingly, by moving the sheet guide 106 in the direction of arrow B as shown in FIG. 8, for example, the light source 101 can be easily replaced.

As shown in FIG. 9, the image pickup device fixing means 2 has a shape substantially conforming to the front contour shape of the body 5 in the moving direction. This is from each camera 3 to body 5
This is to make the distance to the surface substantially constant. As a result, the size of the field of view of all the cameras 3 becomes substantially the same. Fine adjustment of the distance and adjustment of the direction of the camera 3 according to the shape of the body 5 are performed by an adjustment fixture 3a fixed to a predetermined position of the imaging device fixing means 2. In the case where the adjustment cannot be performed even with the adjustment fixture 3a, for example, on a surface having a large difference in height such as a hood (bonnet) and a roof (ceiling) of the body 5, the field of view of the camera 3 is changed by changing the focal length of the lens. May be adjusted.

Further, as shown in FIG. 10, the fields of view of the cameras 3 adjacent to each other are adjusted so as to form an overlapping strip having a predetermined size or more. As described above, the heights of the hood portion and the roof portion are greatly different from each other. However, as is clear from FIG. 10, these portions are not inspected simultaneously by the same camera. As described above, the hood cameras (3) H1 to H8 and the roof cameras (3) R1 to R7 may be switched and adjusted.

As shown in FIG. 21, it is appropriate that the switching position of the camera 3 is the position of the front and rear window portions (indicated by reference numeral C) where no inspection is necessary. At this time, the field of view of the roof cameras (3) R1 to R7 is as shown in FIG. 10, similarly to the left and right side cameras (3) SL1 to SL5, SR1 to SR5, and the hood cameras (3) H1 to H8. It becomes belt-like at the roof.

The overlap of the camera's visual field is an area shown by the hatched portion in FIG. However,
Since the number of cameras 3 increases as the amount of overlap increases, the resolution of the image, that is, the size of the visual field per camera is determined from the minimum size of the defect to be detected, and the size of the visual field Then, the approximate number of cameras necessary for inspecting the entire surface to be inspected may be determined, and the overlap amount may be finally set. Further, as shown in FIG. 10, each camera field of view represented by a rectangle is fixed in a direction such that the body 5 moves in the horizontal or vertical direction in the image received by the camera without being oblique.

FIG. 18 shows an outline of the inspection line. The carriage 6 on which the body 5 is mounted is moved by the conveyor 8. The transport conveyor 8 includes a chain 81,
A driving device 82 and a hook 84 are provided, and the driving device 82 includes a pulse generator 83 for detecting rotation amount information.
Is provided. The hook 84 is hooked on the claw 61 at the lower part of the carriage 6, and the chain 81 is driven by the rotation of the driving device 82 in the direction of the arrow D to convey the body 5 in the direction of the arrow A.

At this time, if there is play between the claw 61 and the hook 84, the movement amount of the chain 81 and the movement amount of the carriage 6 on which the body 5 is mounted do not exactly match, so that
An error may occur at the time of calculating the movement amount of the body 5 using the rotation amount information of the driving device 82 obtained from the pulse generator 83 described later, and the detection accuracy may be reduced. Therefore, as shown in FIG.
1 is sandwiched between the front and rear of the body 5 in the transport direction so as to be always in contact with no gap. In addition, chain 8
By adjusting so that there is no excess slack in 1, the speed of the body 5 and the speed of the conveyor 8 can be made more consistent.

As another example, if the entirety of the conveyor 8 and the rails 7 is uphill in the transport direction, the claws 61 and the hooks 84 are always in contact with the weight of the body 5. Therefore, no rattling occurs. Such a speed matching means is not limited to only this embodiment.

The body 5 mounted on the trolley 6 having the same amount of movement of the conveyor 8 by the above-mentioned speed matching means is attached to the illuminating means 1 and the image pickup device fixing means 2 to which the camera 3 is attached at the aforementioned position. The inside of the gate
, And moves smoothly without vibration at the same time, and at the same time, the inspection processing means 4 automatically inspects for defects on the painted surface of the body in the procedure described below.

FIG. 19 shows an example of an image and a processing flow in the image enhancement processing means 41 constituting the inspection processing means 4.

When the inspection surface on which the light and dark pattern is projected by the illuminating means 1 is imaged by the camera 3, an original image a as shown in FIG. Then, in the original image a input in step S1 shown in FIG.
Since light is diffusely reflected at the uneven defect portion E, the light pattern appears as a dark portion as shown in FIG.
When the defect E exists in a dark pattern, it appears as a bright portion.

An edge detection process such as differentiation is performed on the original image a in step S2, and binarization is performed with a predetermined threshold value in step S3, and the luminance change in the image as shown in FIG. Area, that is, the area with a high spatial frequency was white, and the other areas were black 2
The value image b is obtained. In step S4, labeling (numbering) is performed on the white pixels of the binary image b.
Further, in step S5, the area / centroid is calculated.
In the white pixels of the binary image b, the defect E is an isolated point, and the boundary of the light and dark pattern is a large object that crosses the top and bottom of the screen. Therefore, in step S6, the area is determined using a predetermined determination value. When only isolated points having a small area are extracted, an image as shown in FIG. Here, if there are irregularities on the painted surface that do not become defects such as yuzu skin, these become isolated points (hereinafter referred to as noise N) together with the defects as shown in FIG. There are cases.

The operation of the tracking processing means 42 for extracting only the defect E from such an image will be described with reference to FIGS. When the process of extracting isolated points is continuously performed in time by the image enhancement processing means 41, images of the area determination result are as shown in FIGS. 20A to 20F. (G) of FIG.
It becomes as shown in.

That is, since the camera 3 and the illuminating means 1 are fixed and the body 5 moves, the defect E on the body surface in the camera image changes according to the movement of the body 5 in FIG.
(G) moves in the direction of arrow G, but noise N
Occurs randomly regardless of the movement of the body 5. Accordingly, from the images of the continuous area determination results that differ in time, a defect in which the movement amount and the movement direction of the body 5 match under the predetermined condition can be finally determined as the defect E.

Since the moving direction of the defect E in the image is determined by the direction in which the body 5 passes through the camera 3, the position of the camera 3 is fixed and the body 5 is conveyed as in this embodiment. If the direction is always the same, it can be determined for each camera 3. Further, the field of view of the camera 3 is
As described above, if the defect E is set parallel to the transport direction of the body 5, the defect E moves in the horizontal or vertical direction in the image. In the present embodiment, the camera 3 is described as being fixed in such a direction that the defect E moves in the horizontal direction (direction of arrow G) in the image as shown in FIG.

In the two temporally different successive images obtained in this way, first, the center of gravity coordinates in the y direction (vertical direction of the screen) of each white pixel of each image are compared. As described above, since the defect E moves in the horizontal direction in the image, if there is a white pixel having substantially the same barycenter coordinate in the y direction between the two images, it is determined that the white pixel is likely to be the defect E. Since it is possible, it is stored in a memory as a defect candidate.

Next, regarding the comparison in the x direction, the difference between the barycentric coordinates in the x direction between the two images in the white pixels in the defect candidate indicates the number of moving pixels in the image, and the sign indicates the moving direction. In step S8, the actual number of moving pixels calculated from the moving amount of the body 5 in step S7 shown in FIG. 19D and the moving direction of the body 5 in the image are compared in step S8. Since it can be determined that the white pixel is more likely to be the defect E, the temporally new x, y barycenter coordinates of the white pixel are stored.

When the above series of processing is repeated, and the number of times of comparison in one white pixel is equal to or greater than the predetermined number, the white pixel is determined to be defect E in step S9, and step S10 is performed. Then, the final barycentric coordinates and area are written and stored in the defect list. Then, the tracking processing means 42 performs the above-described processing continuously while the body 5 is in the field of view of the camera, and sends the defect list to the host computer 43 after passing through the body 5.

Here, the actual moving pixel number can be calculated from the equation (1).

Actual number of moving pixels X = (inter-image time t × body moving speed v × image size L) / camera field of view A) (1) Here, the inter-image time t is the two temporal values to be compared. Is a time difference between different images, and in the present embodiment, corresponds to the processing time of the image enhancement processing. This can be measured by counting the intervals at which the tracking unit 42 receives data (area / center-of-gravity coordinate data after area determination) from the image enhancement processing unit 41. (For example, 0.1 [s]) The body moving speed v is calculated by the host computer 43 from the rotation amount information of the driving device 82 by the pulse generator 83, and is sent to the tracking processing unit 42 as needed. (For example, 100 [mm / s]) Further, the image size L is the number of pixels in the body moving direction in the image, and for example, the body 5 moves in the x direction in an image of xy = 512 × 480 pixels. Then L = 5
It becomes 12.

Further, the camera field of view A is the dimension of the camera field of view on the surface to be inspected in the body moving direction. For example, one camera field of view (rectangle in FIG. 10) on the surface to be inspected is x × y = 120 × If the body 5 moves in the x direction at 100 [mm], A = 120 [mm].

Next, the relationship between the inter-image time t, the body moving speed v, and the camera visual field A will be described.

In this embodiment, the inter-image time t corresponds to the image emphasis processing time. However, if the speed v is too fast to pass through the field of view A during the time t, if there is a defect, the same Since the defect does not appear more than once in the image, the above tracking process is not established. Therefore, if the defect has at least two
The time t, the speed v, and the field of view A are set so as to be reflected more than once.

Further, if the above-mentioned image enhancement processing is executed by interrupting the timer at predetermined time intervals, the time t becomes constant, so that various adjustments and calculations become easy. Note that the image enhancement processing unit 41 and the tracking processing unit 42 are not limited to the configuration of the present embodiment.

When inspection of one body is completed by the inspection processing means 4 as described above, based on the defect inspection result,
A mark (for example, a mark ●) is displayed at a position on the body development view corresponding to a defect position on the body surface. This procedure will be described below.

First, the inspection processing means 4 includes an inspection start / end determination means and a movement amount measuring means.

The inspection start / end determination means detects the movement of the body 5 and determines the start and end timings of various inspection processes. For example, the transmission type photoelectric switch is moved in a direction perpendicular to the body transport direction. Further, it can be realized by mounting at a height such that the front and rear ends of the body 5 block the light of the photoelectric switch, and at a position where the body 5 blocks the photoelectric switch immediately before the front end of the body is reflected in the camera field of view. . At this time, if the relative positional relationship between the body 5 and the carriage 6 is known, the photoelectric switch may be mounted in accordance with the carriage 6.

Further, as other examples, the brightness of the image when the body 5 is included in the field of view of the camera and the body 5 is reflected in the camera image, and when only the background is reflected without the body 5. The start and end timings of the inspection may be determined using the difference between the two. Note that the inspection start / end determination means is not limited to the present embodiment.

The moving amount measuring means measures the moving amount of the body 5 based on the inspection start point detected by the inspection start / end determining means. In this embodiment, the movement amount is calculated from the rotation information of the driving device 82 in which the movement amount is obtained from the pulse generator 83 and time information such as an internal clock of the host computer 43. That is, the inspection processing means 4
Can always grasp the inspection position of the body 5 and calculate the position on the body of the defect E detected by the preceding tracking processing means 42. Is stored in the list.

The above series of processing is executed for each camera image, and based on the detected defect list information, for example, a display position in a development view of the body 5 as shown in FIG. 21 is calculated and marked.

In this case, as for the defect display position in the y direction in FIG. 21, the size of each camera field of view and the camera position are known, and the coordinates of the center of gravity of the defect moving in the image in the y direction are almost the same as described above. Since it does not change, it can be easily calculated once the contraction scale of the developed view is determined. Similarly, the defect display position in the x direction can be calculated from the body movement amount with reference to the inspection start point P shown in FIG. 21 as described above. The developed view in which the mark is displayed at the defect position is output to an output device such as a monitor or a printer, for example.

Tables 1 and 2 show examples of processing results calculated by the defect number statistical processing means 11 every hour.

[0077]

[Table 1]

[0078]

[Table 2]

As shown in Tables 1 and 2, the number of defects and defect types (size,
Statistics).

FIG. 26 shows a flow of estimating a defect source. In step 21, vehicle type classification, step 22 paint / color classification, step 23 body part classification, and step 24 defect type classification. It is determined whether each is good or abnormal, and if abnormal, classification of defect size / shape / occurrence frequency or the like is performed in step 26, collation with line defect information is performed in step 27, and step 28 In step 29, the generation source from the coating line configuration is specified, and in step 29, an abnormality display / instruction is performed.

In this way, the source of the defect is specified more accurately based on the result of the statistical processing of the number of defects and the empirical defect information on the line.

FIG. 11 is a schematic front view, plan view and side view of a reference model 91 for adjusting the door surface and the hood / trunk surface.
This is equivalent to 1 f / t. FIG. 12 is a schematic front view, a plan view, and a side view of the reference model 92 for adjusting the pillar surface and the roof surface, and correspond to the pillar surface 92p and the leaf surface 92r.

The shapes of these reference models 91 and 92 are
As is clear from the figure, it substantially matches the front contour of the body 5 in the moving direction. In this embodiment, since the height difference between the hood surface and the roof surface of the body 5 is large and the inspection is performed by different cameras, two types of reference models 91 and 92 are prepared. May be prepared according to the shape of the body (inspection object) 5. Also, grid lines at predetermined intervals are drawn on the planes and side surfaces of the reference models 91 and 92. Each camera 3 captures the grid lines and confirms them on a monitor to view the camera's field of view.
Adjust as shown.

That is, since it is sufficient that the size of the field of view of the camera can be confirmed, the grid lines may be formed of points having predetermined intervals as shown in FIG. 14A, or as shown in FIG. 14B. Alternatively, it may be a figure such as a quadrangle that is almost the same as the size of the field of view determined in advance. At this time, the reference models 91 and 92 displayed on the monitor
The focus adjustment is performed at the same time while looking at the figure.

FIG. 13 shows the illumination means 1 at the time of the above adjustment.
FIG. 8 is a schematic plan view showing a positional relationship between the camera 3 and a reference model 91 (92). In this case, since the body 5 is moved along the rail 7 by the conveyor 8, the reference model 91 (92) is in the same position as the body 5 at the time of movement, that is,
It is installed at a position at 90 ° with respect to the rail so that both ends coincide with the side surfaces of the body 5.

FIG. 16 shows the illumination means 1, the camera 3 and the reference model 9 when adjusting the hood / trunk surface and the door surface.
FIG. 17 is a schematic front view showing the positional relationship with FIG.
FIG. 4 is a schematic front view at the time of adjusting a roof surface and a pillar surface.

Then, in the same manner as described above, in order to match the position of the body at the time of conveyance with the positions of the reference models 91 and 92,
Various adjustments of the camera 3 are performed by mounting the reference models 91 and 92 on the adjustment tables 93 and 94, respectively.

As shown in FIGS.
The shapes of the first and the second 92 have to be substantially adapted to the front contour in the moving direction of the body 5 and to the largest cross-sectional contour among the cross sections in the moving direction of the body 5. The fact that the contour is the largest means that the distance from the camera 3 is the smallest, and if the field of view of the camera is adjusted in this state, the distance from the camera 3 is long in other parts, so that it is adjacent. The overlap between the cameras 3 is always present.

In other words, if the amount of overlap is adjusted in a state where the profile of the cross section is not the maximum, as shown in FIG. 9, the cross section of the camera view is triangular, so that the view becomes smaller as the distance between the camera 3 and the body 5 becomes smaller. Therefore, a predetermined overlapping amount cannot be ensured in a portion having a cross-sectional profile larger than that at the time of adjustment. For this purpose, the reference models 91 and 92 only need to have a shape adapted to the largest cross-sectional profile. For the body 5 of the vehicle, for example, a door portion on the side, a roof center portion on the roof portion, and a hood / trunk In the part, the highest part is suitable for the shape of the reference model.

FIG. 15 shows an example in which the shape of the body is different. Of these, as shown in FIG. 15A, in the case of two types of bodies 5A and 5B, the front contour of the body 5A in the moving direction is larger in all parts than the front contour of the body 5B in the moving direction. , Reference models 91 and 92 are as shown by thick lines in the figure corresponding to the front contour of body 5A. Further, as shown in FIG. 15 (b), when the size of the contour of the bodies 5A and 5B differs depending on the region, the larger one of the bodies 5A and 5B having the larger front contour is selected to smoothly connect them. In the shape of a bold line.

Next, a method of adjusting the camera 3 will be described.

The distance from the camera 3 to the surface to be inspected, that is, the photographing distance varies depending on the shape and type of the body 5.
In order to achieve focus in all cases, it is necessary to set the depth of field (focusing range) of the camera 3 sufficiently large with respect to the change in the shooting distance. Since this depth of field can be calculated from the lens aperture value, the lens focal length and the shooting distance, if the lens specifications and the shooting distance, and the necessary depth of field are determined in advance from the size of the camera field of view, An approximate lens aperture value can be determined.

Further, the shutter speed of the camera 3 needs to be a value that does not blur the image received by the camera when the body 5 is moving as in the present embodiment. An approximate aperture and shutter speed may be finely adjusted and determined.

Next, the lens aperture and the shutter speed are adjusted using an oscilloscope or the like so that the output signal level of the camera 3 does not saturate when the reflectance is the highest in the light reflection characteristic of the surface to be inspected. Then, in the case of the vehicle body 5 as in this embodiment, the adjustment may be performed with the paint color having the highest brightness such as white or silver metallic. In addition, the amount of reflection of the light depends on the illumination and the camera 3.
And the closer the distance to the surface to be inspected, the greater, for example,
If the illumination and camera positions are fixed as in the present embodiment, adjustment may be performed on the surfaces of the reference models 91 and 92 shown above, which are created from the maximum front contour of the body 5.

As a guide for the above adjustment, brightness (luminance)
In order to use the dynamic range in the direction without waste, it is sufficient that the output signal level of the camera 3 is slightly before the white level is exceeded. Further, in order to perform the adjustment efficiently, the background parts other than the figure on the surface of the reference models 91 and 92 on which the figure for the field of view adjustment is drawn have the highest reflectance such as white and silver metallic as described above. In the state, the adjustment can be performed together with the visual field adjustment of the camera 3.

Also, as described above, the reference models 91 and 92
If it is difficult to paint the adjustment surface, prepare a test piece larger than the camera field of view painted in white or silver metallic, place it in contact with the reference model surface corresponding to the camera field you want to adjust, and image it with the camera 3 You can make adjustments while doing so.

Next, the light / dark pattern will be described.

When edge detection by differentiation is used in the image enhancement processing means 41 described above, a luminance change in an image is extracted, and therefore, as shown in FIG. Is extracted as Therefore, if the defect E is near the boundary between the light and dark patterns, the defect E and the boundary may be integrated, and the defect E may not appear as an isolated point.

Further, as the number of boundary lines reflected on one screen increases, the frequency at which the defect E disappears increases, so that the accuracy of defect detection decreases. For example, since the tip of the front fender of the body 5 has a convex curved surface, and acts as a convex lens, if the pitch of the light and dark pattern is constant regardless of the part, the camera image is more visible than the flat part. In many cases, the above boundary line is reflected, and the accuracy of defect detection is reduced.

An example of solving the above problem is shown in FIG.
4 and FIG.

In FIGS. 24 and 25, the dotted lines show light and dark patterns which pass through the light diffusion sheet 105, are reflected on the surface of the body 5, and are reflected in the field of view of the camera 3.
The drawings show examples of each image. For example, in order to make four light-dark pattern boundary lines appear in an image, the pitch of the light-dark pattern may be designed in consideration of the shape of the body 5.
Here, the number of boundaries of the light and dark patterns is a problem,
There is no limitation on how the light and dark patterns appear in the image, that is, the order of light / dark, since it is unrelated to the principle of detection using diffuse reflection at a defect.

As shown in FIGS. 24 and 25,
When the pitch of the light and dark pattern is different between the side surface of the body 5 and the horizontal plane, matting is performed so that the light and dark pattern is not discontinuous at a position where the side of the light and dark pattern sheet transitions to the horizontal plane, for example, at a gate-shaped R portion. If you attach black tape of the hood /
The light and dark pattern can be projected without extreme distortion even in the R portion between the trunk surface.

In the examples shown in FIGS. 24 and 25, the number of boundaries is four in the image.
Need not be the same. In addition, the frequency of defects appearing on the screen increases as the number of boundaries decreases, but if the pitch of the light and dark patterns is too wide in order to reduce the boundaries, the defect is detected using irregular reflection due to unevenness in the defects. Since the accuracy of detecting small defects is reduced, the light and dark patterns may be experimentally designed in consideration of the above.

Next, a development of the object to be inspected will be described.

In this embodiment, regardless of the shape of the body 5, the light / dark pattern of the illuminating means 1 is formed on the surface of the body reflected in the field of view of the camera 3. That is, as shown in FIG. 23, the camera 3 is configured to capture an image obliquely from the front of the body 5, and the camera mounting angle at this time is θ2 and the camera angle of view is θ1.

When the above inspection processing is performed at such a camera position and the defect position is displayed in a normal development view as shown in FIG. 21, the defect position may not coincide with the actual defect position on the body. This is because the camera 3 captures an image from a diagonally forward viewpoint as shown in FIG. 23, whereas the developed view is viewed from the side of the body 5 (side view) and directly above (horizontal plane). This is because the views are different from each other. In order to prevent the display shift of such a defect, as shown in FIG. 22, the state without rotation shown in FIG.
22 (b), that is, a development view in which the body 5 is viewed obliquely from the front like the actual camera 3 may be used. The rotation angles of the developed view at this time may be determined experimentally by the angles θ1 and θ2.

Next, another embodiment shown in FIG. 27 will be described. In FIG. 27, reference numeral 114 denotes a body color detecting means. In the embodiment shown in FIG.
Although the vehicle type, the paint type and the color information are input to the paint information input unit 111, in the embodiment shown in FIG.
In this case, vehicle type and paint type information are input to the vehicle type / paint information input unit 111, and color information is detected and input from the body 5 itself by the body color detection unit 114.

Further, still another embodiment shown in FIG. 28 will be described. In the embodiment shown in FIG. 28, a case is shown in which vehicle type and color information are detected and input from the body 5 itself by the vehicle type detecting means 113 and the body color detecting means 114, respectively.

[Brief description of the drawings]

FIG. 1 is a schematic front view showing a surface defect inspection apparatus according to an embodiment of the present invention, together with a functional block diagram.

FIG. 2 is a schematic plan view of a surface defect inspection apparatus according to one embodiment of the present invention.

FIG. 3 is a schematic side view illustrating a surface defect inspection apparatus according to an embodiment of the present invention.

FIG. 4 is a schematic explanatory view showing a part of a lighting unit in one embodiment of the present invention.

FIG. 5 is an explanatory front view showing a positional relationship between a lighting unit and a body in the embodiment of FIG. 1;

FIG. 6 is a schematic perspective explanatory view showing an example of a light diffusion sheet and a sheet guide of the illumination means.

FIG. 7 is a schematic plan view illustrating a positional relationship among a lighting unit, an imaging device, and a body.

FIG. 8 is an explanatory plan view showing movement of the light diffusion sheet.

FIG. 9 is a schematic front explanatory view showing an example of a mounting position of the camera.

FIG. 10 is an explanatory diagram of a camera field of view.

FIG. 11 is a front view (a), a plan view (b), and a side view (c) showing an example of a reference model for a hood / trunk surface and a door surface.

FIG. 12A is a front view, FIG. 12B is a plan view, and FIG. 12C is a side view of an example of a reference model for a roof surface and a pillar surface.
It is.

FIG. 13 is a schematic plan view showing the positional relationship between a lighting unit, a camera, and a reference model.

FIGS. 14A and 14B are explanatory diagrams illustrating two examples (a) and (b) of a figure for camera view adjustment on the surface of a reference model.

FIG. 15 shows two examples (a) and (b) of different shapes of the reference model.
It is a schematic front explanatory view for each explaining.

FIG. 16 is a schematic front explanatory view showing an example of camera field-of-view adjustment using the reference model shown in FIG. 11;

FIG. 17 is a schematic front explanatory view showing an example of camera field-of-view adjustment using the reference model shown in FIG. 12;

FIG. 18 is a schematic side view illustrating a conveyor partially enlarged.

FIG. 19 is an explanatory diagram showing an example of images (a) to (c) and a processing flow (d) in the image enhancement processing means and the tracking processing means.

FIG. 20 is an explanatory diagram showing the movement of a defect in an image different in time, divided into (a) to (g).

FIG. 21 is an explanatory diagram showing a camera field of view and a development view.

FIG. 22 is an explanatory diagram illustrating a state in which the body is not rotated (a) and a state in which the body is rotated (b).

FIG. 23 is a schematic plan view showing the camera mounting angle and the angle of view with respect to the body.

FIG. 24 is a schematic plan view showing a light-dark pattern on a side surface of a body.

FIG. 25 is a schematic side view showing a light-dark pattern on a horizontal plane of the body.

FIG. 26 is an explanatory diagram showing a flow of estimating a defect source.

FIG. 27 is a schematic front view showing a surface defect inspection apparatus according to another embodiment of the present invention together with a functional block diagram.

FIG. 28 is a schematic front explanatory view showing a surface defect inspection apparatus according to still another embodiment of the present invention together with a functional block diagram.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Illumination means 2 Imaging device fixing means 3 CCD camera (imaging device) 4 Inspection processing means 5 Body (inspection object) 11 Defect number statistical processing means 12 Defect source estimating means 41 Image enhancement processing means 42 Tracking processing means 43 Host computer 44 Detection result output means (printer) 91, 92 Reference model 101 Light source 102 Background plate 105 Light diffusion sheet 105a Light transmission part 105b Matte black part 106 Sheet guide 111 Vehicle type / paint information input means 112 Statistical processing means 113 Vehicle type detection means 114 Body color detection means

[Procedure amendment]

[Submission date] August 19, 1996

[Procedure amendment 1]

[Document name to be amended] Drawing

[Correction target item name] FIG.

[Correction method] Change

[Correction contents]

FIG. 11

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Tadashi Watanabe Nissan Motor Co., Ltd., 2 Takara-cho, Kanagawa-ku, Yokohama, Kanagawa

Claims (18)

[Claims] 1. A surface for irradiating a surface to be inspected of an object to be inspected with light, forming a light-receiving image based on reflected light from the surface to be inspected, and detecting a defect on the surface to be inspected based on the light-receiving image. In the defect inspection apparatus, a lighting means having a gate shape surrounding the object to be inspected and forming a predetermined light and dark pattern on the surface to be inspected, and a gate shape surrounding the object to be inspected and extending from the surface to be inspected. An image pickup device fixing unit to which a plurality of image pickup devices for creating a light reception image based on reflected light are attached, and a defect on a surface to be inspected is detected based on the light reception image obtained by the image pickup device and the result is output. Inspection processing means, defect number statistical processing means for performing statistical processing based on defect detection information and coated surface information from the inspection processing means, and defect statistical processing results and empirical defect / defect information from the defect number statistical processing means Defect generation to estimate the source of defects A surface defect inspection apparatus comprising a source estimation unit, wherein a defect is inspected on a surface to be inspected and a defect generation source is estimated by passing the inspection object through a gate shape of an illumination unit and an imaging device fixing unit. 2. The surface defect inspection apparatus according to claim 1, wherein the portal shape of the illuminating means is a shape substantially adapted to a front contour in a moving direction of the inspection object. 3. The illuminating means includes a plurality of light sources arranged at substantially equal intervals on a white background plate, and light transmitting portions and matte black portions are alternately arranged on the inspection surface side of the light sources. The surface defect inspection apparatus according to claim 1, further comprising a light diffusion sheet, wherein a light-dark pattern is formed on the surface to be inspected by passing light from a light source through a light transmission portion of the light diffusion sheet. 4. A light-diffusing sheet of the illuminating means is stretched over a matte black sheet guide having a gate shape surrounding an object to be inspected, and the sheet guide is movable from a light source and a background plate. The surface defect inspection apparatus according to claim 3, wherein 5. The surface defect inspection device according to claim 1, wherein the gate-shaped shape of the image pickup device fixing means is a shape substantially adapted to a front contour in a moving direction of the object to be inspected. 6. The imaging device is a CCD camera, and each C
The entire field of view of the CD camera forms a continuous band along the front contour in the moving direction of the test object, and the adjacent C
6. The method according to claim 1, wherein the fields of view of the CD cameras overlap each other in a predetermined area, and the moving direction of the object to be inspected coincides with the horizontal or vertical direction in the received image of the CCD camera. The surface defect inspection device according to the above. 7. The surface defect inspection apparatus according to claim 6, wherein the intervals between the light and dark patterns of the light and dark patterns change based on the number of the light and dark patterns reflected on the CCD camera. 8. The field of view adjustment, focus adjustment and overlap amount adjustment of the CCD camera are performed in such a manner that a shape substantially conforming to the front contour in the moving direction of the object to be inspected is formed and a figure of a line or point at a predetermined interval is formed on the surface thereof. The surface defect inspection apparatus according to claim 6, wherein the inspection is performed using a drawn reference model. 9. The surface defect inspection apparatus according to claim 8, wherein the reference model has a shape substantially adapted to a maximum cross-sectional profile of a cross-section in a moving direction of the inspection object. 10. A color of a portion other than a line of a figure of a reference model is a color having the highest lightness in a paint color of a surface to be inspected of an object to be inspected, and a lens of a CCD camera while imaging the surface of the reference model. The surface defect inspection apparatus according to claim 8, wherein an aperture and a shutter speed are adjusted. 11. A test piece, which is painted in a paint color of a surface to be inspected and is larger than a camera field of view, is provided in contact with the surface of the reference model, and the lens aperture and shutter speed of the CCD camera are adjusted while imaging the test piece. The surface defect inspection apparatus according to any one of claims 8 to 10, wherein 12. An inspection processing means, comprising: an image enhancement processing means for extracting only a component having a high frequency region and a level equal to or higher than a predetermined value among spatial frequency components in image data of a received light image obtained by a plurality of imaging devices; And a tracking processing means for detecting a target portion in which the moving amount and the moving direction of the object to be inspected coincide with each other under predetermined conditions in temporally different image data from the image enhancing processing means. A surface defect inspection device according to item 1. 13. The image emphasis processing means includes a defect shape judging means for judging a size (area) and a shape (aspect ratio) of a component having a level higher than a predetermined value in the extracted high frequency region. The surface defect inspection apparatus according to claim 12, wherein 14. The inspection processing means includes an inspection start / end determination means for determining the start and end of the inspection, and a movement amount measuring means for measuring a movement amount of the object from the start of the inspection, and obtained from these means. 13. The apparatus according to claim 12, further comprising: means for calculating a position of the target portion detected by the tracking processing means on the surface to be inspected based on the information, and displaying the result on a developed view of the object to be inspected. The surface defect inspection device according to the above. 15. The surface defect inspection apparatus according to claim 14, wherein a development view of the inspection object is drawn based on an attachment angle and an angle of view of the imaging device with respect to the inspection object. 16. The speed of a test object and a conveyor for moving the test object when the test object passes through the inside of the gate of the illuminating means and the image pickup device fixing means and inspects the test surface. 15. The surface defect inspection apparatus according to claim 1, further comprising a speed matching unit that matches the speed of the surface defect. 17. The defect number statistical processing means includes inspection object information input means such as a type of an inspection object and a paint type (color), and is based on defect size, shape, and body position information from the inspection processing means. The surface defect inspection device according to claim 1, wherein the surface defect inspection device performs statistical processing. 18. The defect generation source estimating means is provided for each type, each part, and the defect type (size, shape) from the defect number statistical processing means.
2. The surface defect inspection apparatus according to claim 1, wherein a defect occurrence site is specified by another defect statistical processing information (defect number) and empirical defect occurrence information.