JP2000321039A - Paint defect inspection apparatus and method - Google Patents

Abstract

(57) [Problem] To provide a coating defect inspection apparatus and method capable of accurately measuring the size of a coating defect. SOLUTION: Two rangefinders 20a and 20b are set according to the angle of view of the CCD camera 6, and the CCD cameras 6 and C
The distance between two points on the surface to be inspected in the direction of the angle of view of the CD camera 6 is measured. Then, based on the measured distance data, a predetermined calculation formula is used to calculate the field of view of the CCD camera 6 due to the change in the shooting distance and the inclination with respect to the surface to be inspected, and thus the change in the area of each imaged pixel per pixel. Is corrected, and the size (area) of the detected paint defect is corrected.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and method for inspecting a paint defect used for optically inspecting the state of a paint surface of a vehicle body (paint defect) in an automobile factory, for example.

[0002]

2. Description of the Related Art Conventional paint defect inspection apparatuses are disclosed, for example, in JP-A-9-184713 and JP-A-10-96.
No. 615, for example.

In the apparatus described in the publication, a painted surface of an automobile body flowing through a line is imaged using a fixed camera and stripe illumination, and a plurality of stripe images (light and dark pattern images) taken in time series are taken. Processing is performed to detect a defect on the painted surface, and output the position and size of the defect to a monitor or a printer.

[0004]

However, in such a conventional paint defect inspection apparatus, since the inspection camera is of a fixed type, the shooting distance and inclination of the vehicle body flowing along the line with respect to the inspection surface are reduced. As a result, since the field of view of the camera and, consequently, the size of the area per pixel of the captured image also change sequentially, an error may occur in the determination of the size of the defect. In other words, when the shooting distance and the inclination change, even if a defect of the same size is photographed, it does not become the same size, and the defect is photographed larger as the photographing distance is shorter, and smaller as the photographing distance is longer. Therefore, the size of the defect cannot be accurately measured, and there is a possibility that an error may occur in the determination.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems in a conventional paint defect inspection apparatus, and is capable of accurately measuring the size of a defect in consideration of a change in the field of view of a camera. An object of the present invention is to provide a defect inspection apparatus and method.

[0006]

The above object of the present invention is achieved by the following means.

(1) The coating defect inspection apparatus according to the present invention comprises:
In a coating defect inspection device that optically detects a coating defect on a surface to be inspected of an object to be inspected, an illumination unit that forms a predetermined light and dark pattern on the surface to be inspected, while relatively moving between the object to be inspected, An imager for imaging a light-dark pattern formed on the surface to be inspected by the illuminator; and a distance between the imager and two or more points on the surface to be inspected in the direction of the angle of view of the imager. Distance measuring means, processing means for processing image data of a plurality of light and dark patterns imaged in time series by the imaging means, and detecting a coating defect on a surface to be inspected; Based on the distance data, a predetermined calculation formula, a correction unit that corrects the size of the coating defect detected by the processing unit, and an output unit that outputs a result of the processing unit and the correction unit. Features and That.

(2) The coating defect inspection apparatus according to the present invention
In a coating defect inspection device that optically detects a coating defect on a surface to be inspected of an object to be inspected, an illumination unit that forms a predetermined light and dark pattern on the surface to be inspected, while relatively moving between the object to be inspected, An imager for imaging a light-dark pattern formed on the surface to be inspected by the illuminator; and three-dimensional CAD data of the object to be inspected and dynamic relative position information of the imager with respect to the object to be inspected, which are calculated in advance. A storage unit for storing data of a distance between the imaging unit and each of two or more points on a surface to be inspected in a direction of an angle of view of the imaging unit for each photographing time; Processing means for processing image data of a plurality of imaged light and dark patterns to detect a coating defect on a surface to be inspected; and a processing method using a predetermined calculation formula based on distance data stored in the storage means. means Therefore a correction means for correcting the magnitude of the detected painted defect, and having an output means for outputting the results of said processing means and said correcting means.

(3) The coating defect inspection method according to the present invention comprises:
In a coating defect inspection method for optically detecting a coating defect on a surface to be inspected of an object to be inspected, a step of forming a predetermined light and dark pattern on the surface to be inspected and imaging while moving relative to the object to be inspected. Imaging the light and dark pattern formed on the surface to be inspected by the means, and measuring a distance between the imager and two or more points on the surface to be inspected in the angle of view direction of the imager; Processing the image data of a plurality of light-dark patterns taken in chronological order to detect a coating defect on the surface to be inspected; and, based on the measured distance data, a predetermined calculation formula to detect the coating. The method includes a step of correcting the size of the defect and a step of outputting the detected coating defect after the correction.

(4) The coating defect inspection method according to the present invention comprises:
In a coating defect inspection method for optically detecting a coating defect on a surface to be inspected of an object to be inspected, a step of forming a predetermined light and dark pattern on the surface to be inspected and imaging while moving relative to the object to be inspected. Means for imaging a light and dark pattern formed on the surface to be inspected by the means; processing image data of a plurality of light and dark patterns imaged in time series to detect a coating defect on the surface to be inspected; The imaging means for each photographing time calculated in advance based on the three-dimensional CAD data of the inspection object and the dynamic relative position information of the imaging means with respect to the inspection object, and the inspection surface in the direction of the angle of view of the imaging means Based on the data of the distance between the two or more points above, by a predetermined formula, a step of correcting the size of the detected paint defect, and a step of outputting the corrected paint defect detected Characterized by having .

[0011]

According to the present invention, based on data of the distance between the imaging means and two or more points on the surface to be inspected in the direction of the field of view of the imaging means, the object to be inspected is calculated by a predetermined calculation formula. The size of the detected paint defect is corrected by correcting a change in the field of view of the camera due to a change in the photographing distance and the inclination with respect to the inspection surface, and a change in the size of the area per pixel of the captured image. Therefore, the size of the coating defect can be accurately measured. Therefore, the defect correction level can be determined by the inspection device, and there is no variation in determination by a person, so that the coating quality can be improved.

[0012]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a configuration diagram of an optical system of a coating defect inspection apparatus according to an embodiment of the present invention, and FIG.
An example of installation of a stripe light source is shown, and FIG.
An example of camera installation is shown. FIG. 2 is a diagram showing a positional relationship between the stripe light source and the CCD camera.

The vehicle body 1 of the vehicle to be measured is a bogie 2
While being placed on the table, it is conveyed from the painting booth to the inspection stage by a conveyor. In the inspection stage, a plurality of lighting devices 4 are arranged side by side on a tunnel-shaped mounting stand 3 so as to surround the transported vehicle body 1, and inside each of the lighting devices 4, a tunnel-shaped striped sheet stand is also provided. 5 are arranged. The number of the illuminating devices 4 attached to the stand 3 is set to an appropriate value such that a light and dark pattern is projected all over the entire outer peripheral surface 1a of the vehicle body 1. The illumination device 4 and the stripe sheet stand 5 constitute a stripe light source (illumination means).

Each lighting device 4 is provided with, for example, a plurality of straight tube type fluorescent lamps. The stripe sheet stand 5 is configured such that, for example, a diffusion plate that scatters and diffuses the light of a fluorescent lamp to create a light source similar to a surface light source, and a stripe sheet that forms a stripe (light and dark stripes) pattern are attached to the stand. ing. The stripe sheet is formed by applying black stripes in the vertical direction at predetermined intervals in the horizontal direction (see FIG. 2).

Light from each of the lighting devices 4 is diffused by the stripe sheet stand 5 to become light (stripe light) having a planar stripe (bright and dark) pattern. Surface) 1a.
As a result, on the outer peripheral surface 1a of the vehicle body 1, a stripe-like light-dark pattern (light-dark pattern) similar to the stripe sheet of the stripe sheet stand 5 is projected.

The light and dark pattern formed on the outer peripheral surface 1a of the vehicle body 1 includes a plurality (here, N) of color CCD cameras (imaging means) (hereinafter simply referred to as "CCD cameras") 6
Is imaged. The CCD camera 6 is also attached to a tunnel-shaped camera stand 7 so as to surround the vehicle body 1. The number and position of the CCD cameras 6 are appropriately set so that the entire outer peripheral surface 1a of the vehicle body 1 can be imaged throughout. That is, each CCD camera 6 has a predetermined imaging area (part) A as a visual field, and the adjacent C
The imaging area A of the CD camera 6 is continuous.
Therefore, all the outer peripheral surfaces 1a of the vehicle body 1 are completely projected by all the installed CCD cameras 6.

Each CCD camera 6 (camera stand 7)
Are arranged, for example, in front of the stripe seat stand 5 in the moving direction B of the vehicle body 1 (see FIG. 2).

In this embodiment, one set (two) of non-contact measurement type rangefinders (distance measuring means) 20a and 20b are attached to each CCD camera 6. The two distance meters 20a and 20b are set in accordance with the angle of view of the corresponding CCD camera 6, and measure the distance between the surface 1a to be inspected and the CCD camera 6 in the angle of view. More specifically,
The distances L1 and L2 between the corresponding CCD camera 6 and the two points P1 and P2 on the inspected surface 1a in the direction of the angle of view of the CCD camera 6 are measured (see FIGS. 4 and 6). Here, P1 is the center position of the horizontal field of view at the foremost side of the CCD camera 6, P2 is the center position of the horizontal field of view at the farthest side of the CCD camera 6, and L1 is the position on the CCD camera 6 and the inspection surface 1a. L2 is the measurement distance between the CCD camera 6 and the point P2 on the inspection surface 1a. That is, the distance L1 on the near side is measured by the distance meter 20a, and the distance L2 on the far side is measured by the distance meter 20b. The angle of view of the CCD camera 6 is geometrically determined in advance by the size of the lens and the CCD element. For example, when a 35 mm lens and a 1/3 inch CCD element are used, the angle of view of the CCD camera is 7.8 degrees horizontally and 5.9 degrees vertically.
Degree.

Each CCD camera 6 transmits data of a stripe image (also referred to as a light / dark pattern image) of the image pickup area A at a predetermined time interval (for example, 1/30 second) via a camera control unit to be described later. To the image processing device. At this time, since the vehicle body 1 is moving in a certain direction at a certain line speed, CC
The stripe image captured by the D camera 6 at regular intervals is an image in which the imaging portion A on the outer peripheral surface 1a of the vehicle body 1 is shifted at regular intervals. Further, the two rangefinders 20a and 20b attached to each CCD camera 6 measure the distances L1 and L2 to the surface 1a to be inspected at the same timing in conjunction with the imaging operation of the corresponding CCD camera 6, respectively. The result is sent to the image processing apparatus. That is, the data of the stripe image and the data of the measurement distances L1 and L2 are simultaneously taken into the image processing device at each photographing time.

The width C of the lighting device mounting stand 3
Can be adjusted within a certain range so as to be able to cope with various types of vehicles having different vehicle sizes (the same applies to the stripe seat stand 5 and the camera stand 7).

Further, here, the form in which the vehicle body 1 is moved has been described, but the present invention is not limited to this. For example,
The stripe light source (illumination device 4 (mounting stand 3) and stripe seat stand 5) and the CCD camera 6 (camera stand 7) may be moved to change the imaging region A on the outer peripheral surface 1a of the vehicle body 1 with time. .

FIG. 3 is a block diagram showing the overall configuration of the signal processing system of the present apparatus.

Each of the N CCD cameras 6 is connected to a corresponding camera control unit 8, and each camera control unit 8 is connected to a corresponding image processing device 9. Further, the two distance meters 20a and 20b attached to each CCD camera 6 correspond to the corresponding CCs.
The camera control unit 8 of the D camera 6 and the image processing device 9 are connected. CCD camera 6 and rangefinder 20
The operations of a and 20b are both controlled by the corresponding camera control unit 8. As described above,
Data of a stripe image (light and dark pattern image) on the outer peripheral surface 1a of the vehicle body taken at regular intervals by the CCD camera 6 is sent as a video signal to the image processing device 9 via the camera control unit 8, and the distance meter 2
The data of the distances L1 and L2 measured by 0a and 20b are sent directly to the image processing device 9 at the same time and at the same time as the data of the captured stripe image. The data of the stripe image at the same time and the data of the measurement distances L1 and L2 taken into the image processing device 9 are temporarily stored in association with each other and processed.

Each of the image processing devices 9 functions as a processing unit and a correction unit, and processes a plurality of time-series still stripe images captured by the CCD camera 6.
In addition to detecting candidate points for paint defects, the distance meter 20a,
It has a function of correcting the size of the detected defect candidate point based on the data of the distances L1 and L2 measured by 20b. Data such as the position and size of the detected defect candidate point (the size after correction) is stored in a built-in memory (RA
M).

A host computer 10 is connected to each image processing device 9. The host computer 10
It functions as a processing means and processes the data of the defect candidate points sent from each of the image processing devices 9 to finally recognize (detect) the coating defect on the inspected surface 1a and to finally recognize the coating defect. It has a function of totalizing the paint defects and editing the paint defects into a predetermined display format (for example, displaying the positions and sizes of the detected defects on a development view of the corresponding vehicle type). The processing result of the host computer 10 is output to a color monitor 11 and a color printer 12 as output means.

The host computer 10 is connected to a pulse generator 14 for outputting a number of pulses corresponding to the number of rotations of a conveyor motor 13 for driving the conveyor. The number of pulses output from the pulse generator 14 corresponds to the line speed (moving speed of the vehicle body 1), whereby the moving amount of the vehicle body 1 can be recognized.

Further, the host computer 10 detects a limit switch 15 for detecting that the vehicle body 1 has entered the inspection stage (vehicle IN) and detects that the vehicle body 1 has left the inspection stage (vehicle OUT). Switch 16 is connected. The host computer 10 sends the vehicle body IN from each of the limit switches 15 and 16.
When the information or the vehicle body OUT information is input, an ON / OFF signal (a start command or an end command) for image processing is output to each image processing device 9.

Here, the limit switch 16 is provided to detect that the vehicle body 1 has left the inspection stage, but the invention is not limited to this. For example, based on image information from each CCD camera 6,
May come out of the inspection stage.

The defect detection processing in the present apparatus configured as described above is performed as follows.

First, the principle of detecting a coating defect is as follows.

When a light-dark pattern is formed on a painted surface, light is irregularly reflected at a defective portion and appears as an isolated point. Therefore, the image is captured by the CCD camera 6 and subjected to high-speed image processing to detect a defect from an isolated point (a defect candidate point) and its movement.

Next, the processing procedure will be described.
An image of the imaged portion A of the surface 1a to be inspected is captured from the CD camera 6, and a predetermined emphasis process (for example, area determination) is performed to facilitate detection of a defective portion, thereby extracting a defect candidate point. This is performed for all of the plurality of light and dark pattern images having different imaging times. When this image processing is performed at regular time intervals, the actual defect moves together with the vehicle body 1 but the erroneously detected point does not move. Therefore, the defect candidate points extracted from each image are tracked to determine whether the defect is a defect. Specifically, the defect candidate point is not subject to the inspection of the vehicle body 1 only when the movement amount of the extracted defect candidate point is equal to the actual movement amount of the vehicle body 1 (imaging part A) obtained from the output of the pulse generator 14. It is determined that the defect actually exists on the surface 1a. By tracking the movement of the defect candidate point in this way,
Coating defects can be detected with high accuracy. Here, the above-described enhancement processing and defect candidate point extraction processing for each still image are performed by the image processing device 9, and the above-described final defect determination processing for a plurality of time-series images is
This is performed by the host computer 10.

The correction of the size of the detected paint defect is performed as follows. Here, the correction of the size of the coating defect is performed at the stage when the defect candidate point is extracted.

When the square lattice is photographed by the CCD camera 6, if the CCD camera 6 is tilted with respect to the inspection surface 1a, as shown in FIG. The field of view is small because the distance L1 to the point P1 is short, and the field of view is wide on the far side because the distance L2 between the CCD camera 6 and the point P2 on the inspected surface 1a is long. As a result, the area per pixel of the image is smaller on the near side than on the far side. The image captured by the CCD camera 6 is shown in FIG.
As shown in (1), in image processing, it is handled as data for each fine pixel (pixel) of the number of horizontal pixels × the number of vertical pixels. At this time, since the size (area) of the defect is obtained by multiplying the number of pixels included in the defect image by the area per pixel, even if the defect has the same size, the defect on the front side is more A large image is taken, and a defect on the far side is taken smaller. Therefore, in order to accurately measure the size of the defect, it is necessary to correct the detected defect area in consideration of the change in the field of view of the CCD camera 6.

FIGS. 4 and 5 are diagrams for explaining the principle of area conversion (correction) when two rangefinders are used according to the present embodiment.

The distance between the surface 1a to be inspected and the CCD camera 6 in the direction of the angle of view (front side: L1, rear side: L2) by two rangefinders 20a, 20b attached to the CCD camera 6.
Is measured. Here, the horizontal field of view is represented by X, and the vertical field of view is represented by Y.

First, the visual field (X1, Y1) on the near side is obtained from the measured distance L1 by the following equation.

Y1 = L1.times.sin (vertical angle of view / 2) .times.2 X1 = L1.times.sin (horizontal angle of view / 2) .times.2 If the measured distance L1 is other than the specified value (for example, if the field of view is In this case, the calculation is not performed, and the area of the defect (defect candidate point) is calculated only in S2 described later.

From the measured distance L2, the visual field (X2, Y2) on the far side is obtained by the following equation.

Y2 = L2 × sin (vertical angle of view / 2) × 2 X2 = L2 × sin (horizontal angle of view / 2) × 2 If the measured distance L2 is other than the specified value (for example, if the field of view is a window, etc.) ), The calculation is not performed, and the area of the defect (defect candidate point) is calculated only in S1 described later.

Next, the values X 1, Y 1,
Using X2 and Y2, S1 = (X1 / number of horizontal pixels) * (Y1 / number of vertical pixels) S2 = (X2 / number of horizontal pixels) * (Y2 / number of vertical pixels) Size (area) per pixel on the side S1, S2
Respectively.

The barycentric coordinates (pixel position) (x, y) (see FIG. 7) where the defect (defect candidate point) exists, the detection value Q of the number of pixels included in the defect (defect candidate point), 1
From the areas S1 and S2 per pixel, the true area of the defect = Q × {(S1 × (number of vertical pixels−y) / number of vertical pixels + S2 × (y / number of vertical pixels)} defect (defect) The size of the candidate point) is converted from the number of pixels to the area to obtain a corrected true area.

In the above formula, Y1 and Y2
Is based on the assumption that the imaging region (the surface 1a to be inspected) is perpendicular to the horizontal field of view with respect to the CCD camera 6, which is different from the actual field of view having an inclination.
It becomes an approximate expression. Therefore, the size (area) of the defect (defect candidate point) can be obtained more accurately by performing correction based on the inclination.

In this case, two rangefinders 20a and 20b are used for each CCD camera 6, but the number of rangefinders used is not limited to this. By increasing the number of rangefinders and performing the same processing, it is possible to perform detailed distance measurement and shape prediction of the surface to be inspected, and even if the surface to be inspected has a complicated shape, the per-pixel The area can be determined accurately. For example, in FIG. 6, the size (area) of the defect can be obtained more accurately by measuring the distance between the four corner points of the field of view and the CCD camera using four rangefinders.

Further, here, the distance between the CCD camera 6 and the surface 1a to be inspected is measured at each photographing time using the distance meter 20, but the present invention is not limited to this. For example, based on three-dimensional CAD data of the vehicle body 1 and relative position information between the CCD camera 6 and the inspection surface 1a at each imaging time, the distance between the CCD camera 6 and the inspection surface 1a at each imaging time is determined in advance. Is obtained, the same effect as when the distance meter 20 is used can be obtained. In this case, the distance data obtained in advance is stored in a built-in memory (such as a ROM) (storage means) of the corresponding image processing device 9.

Further, three-dimensional CAD data of the vehicle body 1 and
The distance between the CCD camera 6 and the inspection surface 1a may be calculated in real time for each imaging time based on the relative position information between the CCD camera 6 and the inspection surface 1a at each imaging time.

FIGS. 8 and 9 are flowcharts showing a specific procedure of the above-described defect detection processing.

First, in step S11, each image processing apparatus 9 inspects the corresponding CCD camera 6 in order from the earliest photographing time at a certain time T _N (N is a set value of a comparison number to be described later). Stripe image data of the imaging portion A on the surface 1a is fetched, and the stripe image data is stored as color image data in a built-in memory. In addition,
At the first photographing time, the count value of the number of pulses from the pulse generator 14 indicating the position of the imaging part A of the CCD camera 6 is reset to zero. This reset processing is performed by the host computer 10 operating in synchronization with the image processing apparatus 9.

Then, in step S12, each image processing device 9 sends two corresponding distance meters 20a, 20b
At the same time T _N , data of the measurement distance (front side: L1, rear side: L2) between the inspection surface 1a and the CCD camera 6 at the same time T _N is fetched, and the measured distance data is stored in the built-in memory.

As described above, the distance meters 20a, 20a
0b instead of actually measuring the distance,
The distance between the inspection surface 1a and the CCD camera 6 may be calculated based on the AD data and stored in the built-in memory, and the distance data may be read in step S12. Alternatively, the distance between the inspection surface 1a and the CCD camera 6 may be calculated in real time based on CAD data or the like.

In step S13, each of the image processing devices 9 measures the measurement distances L1 and L1 acquired in step S12.
Based on L2, the corresponding fields of view X1, Y1, X2, Y2 of the CCD camera 6 are calculated by the above-mentioned predetermined formula, and furthermore, the size (area) S1, Calculate S2.

In step S14, each of the image processing devices 9 determines whether or not the striped image fetched in step S11, that is, the original image has a defective portion (a portion considered to be a defect included in the original image). In this case, a predetermined emphasis process is performed to make it easier to extract a state at a previous stage that is a defect candidate point. By this emphasis processing,
Only a portion (defect portion) considered to be a defect included in the original image is extracted. Extraction of a defective portion by such enhancement processing is performed by a known method, for example, area determination, smoothing, or the like. Then, the position, area (size), and color (RGB of the defect and the background) of the extracted defect portion
Value) and xy size (horizontal width and vertical length).

Further, in particular, the area (size) of the defect portion in the above-described defect extraction data is corrected based on the camera view calculation result in step S13. Specifically, as described above, the coordinates (x, y) of the center of gravity at which the defect exists, the detection value Q of the number of pixels included in the defect, and the areas S1 and S2 per pixel obtained in step S13. From the above, a corrected true area is obtained by the above-mentioned predetermined formula (proportional formula).

Therefore, as a result of the processing in step S14, a defective portion is extracted, and the position, area (true area corrected in size in the visual field), color (RGB values of the defect and the background), and xy size of the defective portion are extracted. The data (defect extraction data) of the height (width and height) is obtained. These data are
It is stored in a built-in memory of each image processing device 9.

In step S15, the value i of a counter for counting the number of captured original images to be compared is set to one.

Then, in step S16, step S
The defect extracted from the original image at the time T _{Ni immediately} before the i-th image which has already been subjected to the processing in step S 14 and the defect newly extracted in step S 1
Then, a comparison is made with the defective portion extracted from the original image at the time T _N that has undergone the processing of Step 4. Note that at the very first stage where only one original image has been captured, this comparison process cannot be actually performed. However, at the stage where two or more original images have been captured, the image taken immediately before is captured. The comparison with the original image will be performed.

In step S17, the host computer 10 counts the pulses output from the pulse generator 14, calculates the current position of the imaging region A based on the counted number of pulses, and compares the current position with the time to be compared. The movement amount D of the imaging part A between the original image captured at the time T _N and the original image captured at the time T _Ni is calculated by comparing with the position of T _Ni .

In step S18, the host computer 10 compares one of the defective portions extracted from the original image in S14 with the position of the defective portion extracted from the original image at time T _Ni , and The movement amount d of the unit is calculated.

Then, in step S19, step S
The movement amount D of the imaging region A calculated in step 17 and the step S1
By comparing the movement amount d of the defective part calculated in step 8, it is determined whether the movement amount D of the imaging region A is substantially equal to the movement amount d of the defective part. When the moving amount d of the defective portion is substantially equal to the moving amount D of the imaging region A (S19: YES), the process proceeds to step S20, and the moving amount d of the defective portion is substantially equal to the moving amount D of the imaging region A. If not (S19: NO), the process immediately proceeds to step S21.

Then, in step S20, the imaging region A
Step S1 presenting a movement amount d substantially equal to the movement amount D of
The defect extracted in step 4 is registered as a defect candidate point indicating the next step which is more certain as a possibility. When this registration is performed, the defect extraction data obtained in step S14 is read out and registered. Therefore, the defect candidate point recording data at each time also includes the position, area (true area corrected in size in the visual field), color (RGB values of the defect and the background), and xy size (width and height) of the defect candidate point. Data).

Then, in step S21, step S
The defective portion processed in step 18 to step S20 is
Whether it is the last defective part extracted in step S14, in other words, whether there is any other defective part that is not the target of processing in steps S18 to S20 to decide. If it is the last defective portion (S21: YES), the process proceeds to step S22. If it is not the last defective portion (S21: NO), the process returns to step S18, and the other defective portion extracted from the original image in step S14. The movement amount d is calculated for one of them. That is, the processing in steps S18 to S20 is performed in step S14.
Is performed on all of the defective portions extracted from the original image.

Then, in step S22, the counter
It is determined whether the value i is N. If the counter value i is N
(S22: YES), the process proceeds to step S24 and the counter
If the value i is not N (S22: NO), step S23
Proceed to. Then, in step S23, the counter value i is
The value is incremented by one, and the process returns to step S16. I
Therefore, the processing of steps S16 to S21 is
Time T newly taken in_N N images before
All the original images at are compared. example
For example, if N = 5, the time T_N smell
All the defects extracted from the original image
Time T_N-1 , T_N-2 , T_N-3 , T _N-4 , T_N-5 To
Defects extracted from the original images captured
Are checked against each other, and
The moving amount d of the defective portion set as the supplementary point is determined by the imaging unit at each time.
It is determined whether the movement amount D of the position A is substantially equal to the movement amount D.

Then, in step S24, step S
The number of times each defect candidate point is registered is counted for all the defect candidate points registered in 20.

Then, in step S25, the number counted in step S24 is set to a preset count number M (where N ≧ N) for all the defect candidate points.
M) It is determined whether it is not less than. If there is any one whose count number is equal to or larger than the set value M (S25: YES), in step S26, a defect candidate point whose count number is equal to or larger than the set value M is finally determined as a defect existing on the inspection surface 1a. Is determined, and the process proceeds to step S27. If there is no value equal to or greater than the set value M, that is, if all of the values are less than the set value M (S25: NO), the process returns to step S11 to fetch the next original image. Perform a similar series of processing.

Then, in a step S27, it is determined whether or not an end command has been issued. If there is an end command (S
27: YES), end the above series of processes, and if there is no end command (S27: NO), return to step S11,
The next original image is fetched and the above series of processing is performed. That is, the above series of processing is continued until an end command is issued.

The defect is identified based on the defect candidate point record data at each time corresponding to the defect determined in step S26, the color, size (area), size change degree, length and width of the defect. This is performed by calculating the ratio and estimating the type of the defect based on the information.

Therefore, according to the present embodiment, CC
Based on the measurement data of the distances L1 and L2 between the D camera 6 and the two points P1 and P2 on the inspection surface 1a in the direction of the field of view of the CCD camera 6, the inspection surface 1
CCD camera 6 based on changes in shooting distance and tilt with respect to a
The size of the detected paint defect is corrected by correcting the change in the size of the area per pixel of the captured image, and thus the captured image, so that the size of the paint defect can be accurately measured. can do. As a result, the defect correction level can be determined by the inspection device, and there is no variation in determination by a person, so that the coating quality can be improved.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an optical system of a paint defect inspection apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a positional relationship between a stripe light source and a CCD camera.

FIG. 3 is a block diagram showing an overall configuration of a signal processing system of the apparatus shown in FIG.

FIG. 4 is a drawing for explaining the principle of area conversion (correction) when two distance meters are used according to the present embodiment.

FIG. 5 is an explanatory diagram similar to FIG. 4;

FIG. 6 is an explanatory diagram showing a field of view when a square grid is photographed.

FIG. 7 is an explanatory diagram of image processing data.

8 is a flowchart showing a specific procedure of a defect detection process in the signal processing system of FIG.

FIG. 9 is a flowchart following FIG. 8;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Car body (inspected object), 4 ... Illumination device (illumination means), 5 ... Stripe sheet stand (illumination means), 6 ... Color CCD camera (imaging means), 9 ... Image processing apparatus (processing means, correction means) Reference numeral 10 host computer (processing means) 11 monitor (output means) 12 printer (output means) 20a, 20b distance meter (ranging means)

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Kiyoshi Yoshida 2nd Takaracho, Kanagawa-ku, Yokohama-shi, Kanagawa F-term (reference) 2F065 AA02 AA22 AA58 AA61 BB05 BB24 CC11 DD03 EE11 FF04 GG03 GG15 GG17 HH02 JJ03 JJ16 JJ16 JJ26 MM03 PP22 QQ23 QQ31 SS02 SS06 SS13 2G051 AA89 AB12 BA20 CA03 CA04 DA01 DA06 EA14 EA16 ED07 FA10

Claims (4)

[Claims] A coating defect inspection apparatus for optically detecting a coating defect on a surface to be inspected of an object to be inspected, comprising: a lighting unit for forming a predetermined light and dark pattern on the surface to be inspected; Imaging means for imaging a light and dark pattern formed on the surface to be inspected by the illumination means while relatively moving; and two or more points on the surface to be inspected in the direction of the angle of view of the imaging means. A distance measuring unit that measures a distance between the plurality of light and dark patterns imaged in time series by the imaging unit, and a processing unit that detects a coating defect on a surface to be inspected; Correction means for correcting the size of the coating defect detected by the processing means by a predetermined calculation formula based on the distance data measured by the means, and output means for outputting the results of the processing means and the correction means When, Coating defect inspection apparatus characterized by having. 2. A coating defect inspection apparatus for optically detecting a coating defect on a surface to be inspected of an object to be inspected, comprising: a lighting unit for forming a predetermined light and dark pattern on the surface to be inspected; Imaging means for imaging the light and dark pattern formed on the surface to be inspected by the illumination means while moving relatively; three-dimensional CAD data of the object to be inspected; and dynamic relative position information of the imaging means with respect to the object to be inspected. Storage means for storing data of the distance between the imaging means and two or more points on the surface to be inspected in the direction of the field of view of the imaging means at each imaging time point calculated in advance based on the imaging means; Processing means for processing image data of a plurality of light and dark patterns imaged in chronological order by a processing means for detecting a coating defect on a surface to be inspected, based on distance data stored in the storage means,
A correction means for correcting the size of the coating defect detected by the processing means by a predetermined calculation formula; and an output means for outputting a result of the processing means and the correction means. Inspection equipment. 3. A coating defect inspection method for optically detecting a coating defect on a surface to be inspected of an object to be inspected, the method comprising: forming a predetermined light and dark pattern on the surface to be inspected; While moving, the light and dark pattern formed on the surface to be inspected is imaged by the imaging means, and the distance between the imaging means and two or more points on the surface to be inspected in the direction of the angle of view of the imaging means is increased. Measuring, processing image data of a plurality of light and dark patterns taken in time series to detect a coating defect on the surface to be inspected, based on the measured distance data, by a predetermined calculation formula A step of correcting the size of the detected coating defect, and a step of outputting the corrected detected coating defect. 4. A coating defect inspection method for optically detecting a coating defect on a surface to be inspected of an object to be inspected, the method comprising: forming a predetermined light and dark pattern on the surface to be inspected; A step of imaging a light and dark pattern formed on the surface to be inspected by the imaging means while moving, and processing image data of a plurality of light and dark patterns imaged in time series to detect a coating defect on the surface to be inspected The imaging means and the angle of view of the imaging means for each imaging point calculated in advance based on three-dimensional CAD data of the inspection object and dynamic relative position information of the imaging means with respect to the inspection object Correcting the size of the detected paint defect by a predetermined formula based on the data of the distance between the two or more points on the surface to be inspected, and the detected paint defect after the correction. And the process of outputting A coating defect inspection method.