JP2795013B2 - Painted surface evaluation device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a painted surface evaluation apparatus for evaluating the quality of a painted surface.

[0002]

2. Description of the Related Art Conventionally, as this kind of paint surface evaluation apparatus, there is an "coating film smoothness automatic inspection apparatus" disclosed in, for example, Japanese Patent Application Laid-Open No. 1-210806. This device is
As shown in FIG. 9, for example, light of a light and dark stripe pattern is projected on a painted surface of an automobile by a light and dark pattern projection means M2. Then, the reflected light of the projected light and dark pattern light from the painted surface is imaged by an image pickup means M3 such as a television camera, and the intensity level signal of the image of the reflected light picked up by the image pickup means is converted to a fine line excess processing means M5. Then, the captured light / dark pattern image is converted into a light / dark stripe pattern of a binary signal in comparison with a specific reference value, and the binarized light or dark portion is thinned to a predetermined line width. Then, in the smoothness determining means M6, the direction of each thinned line is extracted in a predetermined unit, and the smoothness of the painted surface is determined based on the difference in the direction of the line.

[0003]

In the above-described conventional apparatus, it is necessary to convert the reflected light into binary signals before thinning processing of the intensity level of the reflected light. Therefore, when the light is bright, the black portion of the line width after the thinning becomes thinner, and when the light is darker, the black portion becomes thicker, which causes a problem in the smoothness determination.

[0004] The present invention has been made in view of the above,
It is an object of the present invention to provide a paint surface evaluation apparatus capable of removing binarization processing from a captured image and performing evaluation without variation.

[0005]

In order to achieve the above object, a coating surface evaluation apparatus according to a first aspect of the present invention comprises: an irradiating means for irradiating a coating surface with light having a predetermined light and dark stripe pattern; Imaging means for imaging reflected light of the pattern light from the painted surface; density bandwidth calculating means for obtaining a density bandwidth from an image of the reflected light imaged by the imaging means; and reflection band light reflected by the imaging means. A density gradient distribution calculating means for obtaining a density gradient distribution from an image; and an evaluation means for evaluating a painted surface based on the density bandwidth and the density gradient distribution obtained by the density bandwidth calculating means and the density gradient distribution calculating means, respectively. It is the gist to have.

According to a second aspect of the present invention, there is provided an apparatus for evaluating a painted surface, comprising: an irradiating means for irradiating the painted surface with light having a predetermined light and dark stripe pattern; and a reflected light of the light having the light and dark stripe pattern from the painted surface. Imaging density imaging apparatus, density bandwidth calculation means for obtaining a density bandwidth from an image of reflected light captured by the imaging means, and density gradient distribution calculation for obtaining a density gradient distribution from an image of reflected light captured by the imaging means Means, a light and dark fringe number index calculating means for obtaining an index of the number of light and dark fringes from the image of the reflected light imaged by the imaging means, the density bandwidth calculating means, the density gradient distribution calculating means and the light and dark fringe number index calculating means. The gist of the present invention is to have evaluation means for evaluating a painted surface based on the obtained density bandwidth, density gradient distribution, and number index of light and dark stripes.

[0007]

According to the first aspect of the present invention, there is provided a painted surface evaluation apparatus,
The painted surface is irradiated with light of a predetermined light and dark stripe pattern, the reflected light from the painted surface is imaged, the density bandwidth and the density gradient distribution are obtained from the image of the reflected light, and based on the density bandwidth and the density gradient distribution. To evaluate the painted surface.

According to a second aspect of the present invention, there is provided an apparatus for evaluating a painted surface, wherein the painted surface is irradiated with light having a predetermined light and dark stripe pattern, the reflected light from the painted surface is imaged, and the density of the reflected light is calculated from the image of the reflected light. The bandwidth, the density gradient distribution, and the index of the number of light and dark stripes are obtained,
The coated surface is evaluated based on the density bandwidth, the density gradient distribution, and the number index of light and dark stripes.

[0009]

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

FIG. 1 is a block diagram showing a configuration of a painted surface evaluation apparatus according to one embodiment of the present invention. The painted surface evaluation apparatus shown in FIG. 1 includes a light-dark stripe pattern projecting device 3 that irradiates a test object, for example, a painted surface 1a of an automobile 1 with a predetermined light-dark stripe pattern light, and the light-dark stripe pattern projecting device 3 uses the light-dark stripe pattern projecting device 3. An image pickup device 5 that picks up an image of a light and dark stripe pattern projected on the painted surface 1a irradiated with the light of the stripe pattern as a level signal of light, and the direction of each pixel from the data of the image picked up by the image pickup device 5; That is, the density gradient of each pixel and the intensity of each pixel, that is, the edge intensity, and the density bandwidth of the entire image are obtained, and the standard deviation σ and the density of the density gradient distribution of the pixel having the edge intensity of 濃度 or more of the density bandwidth are obtained. It has a detecting device 7 for outputting the bandwidth W, and an evaluation judging device 9 for evaluating the quality of the painted surface 1a from the density bandwidth W and the standard deviation σ of the density gradient distribution outputted from the detecting device 7. In addition,
In general, as the smoothness of the painted surface deteriorates, the light and dark stripes applied to that part will be distorted and disturbed,
The painted surface evaluation apparatus of the present embodiment uses a density gradient distribution as an index for measuring the degree of distortion, and does not use binarization processing of an original image as in the related art.

FIG. 2 is a block diagram showing a detailed configuration of the detection device 7 shown in FIG. As shown in the drawing, the detection device 7 includes a density bandwidth detection unit 71 and an edge intensity calculation unit 73 to which the original image data 50 of the painted surface 1a captured by the imaging device 5 is supplied. FIG. 3 shows an example in which the original image data 50 is binarized. FIG. 3A shows an original image of a good painted surface, and FIG. 3B shows an original image of a bad painted surface. I have. On a good painted surface, the slit light is reflected almost as it is because the coating is uniform, and nearly parallel stripes as shown in Fig. (A) can be obtained. And the stripe pattern is greatly disturbed as shown in FIG.

The density bandwidth detecting means 71 checks the density distribution of the image as shown in FIG. 4A and obtains the density bandwidth from this distribution. To reduce the effect of noise
As shown in FIG. 4B, 10% of the cumulative density histogram
The range of ~ 90% is determined as the density bandwidth W.

The edge strength calculating means 73 calculates primary differential coefficients dx and dy of each pixel, and calculates the sum of squares of the primary differential coefficients as edge strength. Then, the density bandwidth obtained by the density bandwidth detecting means 71 and the edge strength calculated by the edge strength calculating means 73 are supplied to a density gradient distribution calculating means 75, and the density gradient distribution calculating means 75 converts the primary differential coefficient dx , Dy, a density gradient at each pixel, that is, an edge intensity weighted density gradient distribution g, and a standard deviation σ of the density gradient distribution are calculated. At this time, if the edge intensity of the pixel is less than half of the density bandwidth, it is prevented from participating in the density gradient histogram.
That is, the edge intensity of each pixel is subjected to threshold processing with the density bandwidth, and only the density gradient of the pixel whose edge intensity exceeds 濃度 of the density bandwidth is examined. You have solved the problem. FIG. 5 shows a histogram of the density gradient distribution. FIG. 5A shows a case of a good painted surface, and FIG. 5B shows a case of a bad painted surface. On a good painted surface, the direction of the density gradient is uniform, so the histogram has a small variance as shown in FIG. On the other hand, on the bad painted surface, the direction of the density gradient is not uniform, and the histogram has a large variance as shown in FIG.

Returning to FIG. 1, the evaluation judging device 9 evaluates the quality of the painted surface 1a based on the density bandwidth W from the detecting device 7 and the standard deviation σ of the density gradient distribution. The quality is evaluated, for example, by the following equation.

Evaluation value = A × W + B × σ + C W: density bandwidth σ: standard deviation of density gradient distribution A,
B, C: constants The relationship between the evaluation value, the concentration bandwidth W, and the standard deviation σ is obtained in advance by experiments, and A, B, and
C is determined.

In the above embodiment, without performing the binarization processing,
Since the density gradient of the image is obtained and the distribution is examined, the threshold processing based on the edge intensity based on the density bandwidth is performed. The effect on the dispersion of the evaluation points can be eliminated, and the density gradient distribution is effective for any of various coating colors having different density bands, for example, silver metallic color.

FIG. 6 is a block diagram showing the configuration of another embodiment of the present invention. In the embodiment shown in the figure, the function of the detection device 7 of the embodiment shown in FIG. 1, that is, the density gradient and edge intensity of each pixel and the density bandwidth of the entire image are obtained from the image data, and the density bandwidth of 1 is obtained. The detection device 70 having a function of obtaining an index of the number of bright and dark stripes in addition to the function of obtaining the standard deviation σ and the density bandwidth W of the density gradient distribution of the pixel having the edge strength of / 2 or more is the same as the detection device 7 of FIG. 1 and has been learned based on the density bandwidth W, the standard deviation σ of the density gradient distribution, and the index information of the number of bright and dark stripes output from the detecting device 70 instead of the evaluation judging device 9 in the embodiment of FIG. 1 in that it has an evaluation determining device 90 for evaluating the quality of the painted surface 1a using the neural network described above, and the other configuration is the same as that of the painted surface evaluating device in FIG.

More specifically, the detecting device 70 obtains the density gradient and edge intensity of each pixel, the frequency analysis of the accumulated pixel value in the stripe direction, and the density bandwidth of the entire image from the image data from the imaging device 5, The standard deviation σ of the density gradient distribution of a pixel having an edge intensity equal to or more than の of the bandwidth, the density bandwidth W, and the index N of the number of bright and dark stripes, which evaluate the frequency at which the peak power is applied by frequency analysis of the cumulative pixel value in the stripe direction. It is supplied to the determination device 90.

The evaluation determining device 90 makes the neural network learn from the standard deviation σ of the density gradient distribution, the density bandwidth W, and the index N of the number of light and dark stripes, and evaluates the painted surface 1a using the learned neural network.

FIG. 7 is a block diagram showing a detailed configuration of the detection device 70. The difference is that a bright and dark stripe number calculation means 77 is provided in addition to the configuration of the detection device 7 shown in FIG. There is only. The number of light and dark stripes calculating means 7
7 is to add a pixel value in the direction of the light and dark stripes, perform frequency analysis on a change in the accumulated pixel value in a direction orthogonal to the light and dark stripe direction, and determine the frequency at which the peak appears as an index N of the number of light and dark stripes.
And

FIG. 8 is a diagram showing the configuration of a neural network 91 included in the evaluation / judgment device 90. The neural network 91 includes a density bandwidth W from the detection device 70, a standard deviation σ of a density gradient distribution, It is composed of an input layer 92 to which the index N of the number of stripes is input, an intermediate layer 93, and an output layer 94 to output five-level evaluation points. Based on the input density bandwidth W, the standard deviation σ of the density gradient distribution, and the index N of the number of light and dark stripes, the output of the neural network 91 is the sum of squares of the error with the sensory evaluation value of the painted surface evaluated by a human. Learn to minimize.
Then, based on this learning, the neural network 91 realizes a mapping table between the density bandwidth W, the standard deviation σ of the density gradient distribution, the index N of the number of light and dark stripes, and the sensory value, and evaluates the sensory value in five steps. Output the corresponding evaluation point.

[0022]

As described above, according to the present invention,
The painted surface is irradiated with light of a predetermined light and dark stripe pattern, the reflected light from the painted surface is imaged, and the density bandwidth, the density gradient distribution and the index of the number of light and dark stripes are obtained from the image of the reflected light, and the density bandwidth is determined. Since the coating surface is evaluated based on the density gradient distribution and the number index of the number of light and dark stripes, a binarization process as in the related art is not required, so that evaluation can be performed with high accuracy without variation in evaluation points.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a painted surface evaluation apparatus according to one embodiment of the present invention.

FIG. 2 is a block diagram showing a detailed configuration of a detection device used in the painted surface evaluation device of FIG.

FIG. 3 is a diagram showing an original image captured by an imaging device of the painted surface evaluation device of FIG. 1;

FIG. 4 is a diagram showing a density gradient distribution.

FIG. 5 is a diagram showing a histogram of a density gradient distribution.

FIG. 6 is a block diagram showing a configuration of another embodiment of the present invention.

FIG. 7 is a block diagram showing a detailed configuration of a detection device used in the embodiment of FIG.

FIG. 8 is a configuration diagram of a neural network used in the evaluation determination device of the embodiment in FIG. 6;

FIG. 9 is a configuration diagram of a conventional painted surface evaluation device.

[Explanation of symbols]

 1a Painted surface 3 Light and dark stripe pattern projection device 5 Imaging device 7, 70 Detection device 9 Evaluation judgment device

──────────────────────────────────────────────────続 き Continued on the front page (58) Fields surveyed (Int. Cl. ⁶ , DB name) G01B 11/00-11/30 G01N 21/00-21/01 G01N 21/17-21/61

Claims (2)

(57) [Claims] 1. An irradiating means for irradiating a painted surface with light of a predetermined light and dark stripe pattern, an image pickup means for imaging reflected light of the light of the light and dark stripe pattern from the painted surface, and a reflection image picked up by the image pickup means Density bandwidth calculation means for obtaining a density bandwidth from an image of light; density gradient distribution calculation means for obtaining a density gradient distribution from an image of reflected light captured by the imaging means; the density bandwidth calculation means and the density gradient distribution An evaluation means for evaluating a coating surface based on the density bandwidth and the density gradient distribution obtained by the calculation means. 2. An irradiating means for irradiating a light of a predetermined light and dark stripe pattern onto a painted surface, an image pickup means for picking up reflected light of the light of the light and dark stripe pattern from the painted surface, and a reflection image picked up by the image pickup means. Density bandwidth calculating means for obtaining a density bandwidth from an image of light; density gradient distribution calculating means for obtaining a density gradient distribution from an image of reflected light captured by the imaging means; and a reflected light image captured by the imaging means. Brightness / darkness fringe number index calculating means for obtaining an index of the number of bright / dark fringes; density bandwidth, density gradient distribution and the number of bright / dark fringes obtained by the density bandwidth calculating means, the density gradient distribution calculating means and the bright / dark fringe number index calculating means, respectively. An evaluation means for evaluating the painted surface based on the index.