JP2013210227A - Visual inspection device for light transmitting molding - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an appearance inspection apparatus capable of distinguishing a color difference even in an image having a low saturation such as a light transmissive molded body.
In an appearance inspection apparatus for a light-transmitting molded product,
The image data is composed of a color space data group 1 composed of light intensity values of RBG color components, and the color space data group 1 is converted into a color space data group 2 composed of color components including hue and saturation. The first conversion means, the second conversion means for rotating the hue of the color space data group 2 to convert it to the color space data group 3, and the color space data group 3 with the light quantity values of the color components of the RBG. A third conversion means for converting the color space data group 4 into
An appearance inspection apparatus that selects one of the color components constituting the color space data group 4 and corrects the light amount value of the selected color component using the light amount values of the remaining color components is provided.
[Selection] Figure 2

Description

  The present invention relates to a visual inspection apparatus for a light-transmitting molded article that detects partial discoloration such as slight color unevenness that occurs when glass or resin having light transparency is molded using image processing.

  When parts are manufactured using transparent glass or resin as a material (hereinafter referred to as a light-transmitting molded product), discoloration and color unevenness may occur due to adhesion of a small amount of foreign matter or heat deterioration. Such discoloration and color unevenness not only impair the appearance of the appearance, but also cause a decrease in mechanical strength and deterioration with time of the parts. For this reason, in the manufacturing process of a light-transmitting molded product, an appearance inspection mainly including a visual inspection is generally performed. However, since such a visual inspection is a sensory inspection, it is known that there are individual differences in the inspection results due to factors such as differences in the physical condition and proficiency of the inspector, and the inspection results are not stable. It was difficult to ensure a certain level of quality. Further, the visual inspection generally has a problem that the inspection time is longer than the automatic inspection performed by the machine.

  In order to solve such problems, an attempt is made to improve inspection stability or shorten inspection time by taking an image using an imaging device such as a CCD camera and performing an appearance inspection using image processing. Has been done.

For example, the technique used in the color unevenness appearance inspection apparatus disclosed in Patent Document 1 will be described with reference to FIG. 8. From the photographed primary image (S61), a step of photographing a test object with a CCD color camera. A step (S62) of generating, for each pixel, three primary color data corresponding to the light quantity values of the three primary color components of R (red), G (green), and B (blue), and using the image processing apparatus, Steps (S63, S64) for calculating color unevenness distribution data by performing calculation processing for each pixel, and color unevenness of the primary image based on the color unevenness distribution data using the conversion table for pseudo colorization processing shown in FIG. A step (S65) of reproducing a pseudo-colored secondary image representing a distribution for a secondary image monitor is performed, and a color unevenness appearance inspection (S66) of the test object is performed.

As disclosed in Patent Document 2, a method shown in FIG. 10 is known as a method for clarifying a slight color difference of a test object. A step of inputting a color image (S81), a step of performing negative-positive inversion conversion as necessary (S82), and RGB data of each pixel constituting the color image are converted into hue (H), saturation (S), HSI conversion step (S83) represented by luminance (I), information acquisition means for obtaining color information including at least saturation from each pixel of the color image, and a predetermined amount of saturation of each pixel of the color image The step of shifting (S84), the step of obtaining the mode value of hue (S85), the step of rotating the hue of each pixel from the obtained mode value of hue (S86), and the step of extracting the target pixel (S87) And a step of converting the saturation / luminance (S88) and a step of converting the gradation of the obtained image data (S89), and an image in which a slight color difference is clarified is output (S90).

JP-A-5-332838 JP 2009-152868 A

A material that is remarkably soiled or discolored due to dirt or color unevenness generated on the light-transmitting molded product can be easily identified by visual inspection, and can also be inspected by a conventional visual inspection apparatus. However, the appearance inspection of the light-transmitting molded body has the following inherent problems because the material constituting the molded body transmits light. The first is that ambient light and light that illuminates the test object are observed overlapping with dirt and color unevenness, etc., so the unique color due to dirt and color unevenness becomes inconspicuous and is distinguished by the color characteristics The problem is that it becomes difficult. That is, as shown in FIG. 12, even if dirt or color unevenness indicates a specific color (FIG. 12A), the surrounding light color (FIG. 12B) is observed by overlapping. In this state, there are slight inherent color changes in the surrounding colors, making it difficult to visually distinguish the colors. (Fig. 12 (c))
The second is the intensity of light that is reflected or scattered by the influence of the shape of the light-transmitting molded body when ambient light or light that irradiates the test object is reflected or scattered inside the light-transmitting molded body. However, when the light-transmitting molded product is observed, there appears to be a black portion with respect to the surroundings, which may cause an illusion of dirt or color unevenness. Since colors such as dirt and color unevenness are inconspicuous, it has been a problem that a shadow caused by reflection and scattering of light is mistaken as dirt and color unevenness.

  In the method of Patent Document 1, as shown in S63 and S64 of FIG. 8, a difference between arbitrary two components of RGB (red (R) and blue (B) in S63 of FIG. 8) is taken. Based on the size, the color unevenness distribution F is obtained. When the light-transmitting molded product is observed, the color tends to be inconspicuous, so that the RGB color components are close to each other, and the color unevenness distribution F obtained by taking the difference between any two components is small. Become. Therefore, even if the light-transmitting molded article is observed by the method of Patent Document 1, it is not possible to distinguish the problem of dirt, color unevenness, and the like by color characteristics.

The method of patent document 2 is demonstrated using FIG. In FIG. 11, when color information is represented by hue (H), color (S), and luminance (I), the circumferential direction is the hue axis (H), the radial direction is the saturation axis (S), and it is perpendicular to the paper surface. The HSI color space with the direction as the luminance axis (I) is shown. As shown in FIG. 11 (1), the color information of each pixel of the image data is plotted on the HSI color space, and the center value of the pixel distribution (displayed as “Δ” in FIG. 11 (1)). The saturation is shifted so that the center value of the pixel distribution becomes the center of the HSI color space as shown in FIG. By shifting the color information of each pixel in this way, as shown in FIG. 11 (1), the hue (H) is close and the saturation is distributed (the same color has a difference in shading). Can be recognized as a difference in hue (H), and it becomes easy to distinguish adjacent images of color differences. When the light-transmitting molded body is observed, the color is not noticeable. Therefore, the distribution of the color information of each pixel is distributed in the central portion when plotted on the HSI color space. Even if the center value of the pixel distribution is obtained as shown in FIG. 11 (1), the color information distribution of the underlying pixel is distributed at the center. Even if shifting is performed, the amount that can be shifted is small. The method of Patent Document 2 cannot solve the problem of distinguishing a color difference image having a small saturation.
Accordingly, an object of the present invention is to provide an appearance inspection apparatus capable of distinguishing between different colors even in an image having a low saturation such as a light-transmitting molded body.

In order to solve the above-mentioned problem, the invention described in claim 1 is a color imaging device that receives light transmitted through the inspection object or reflected from the inspection object, and an image captured by the color imaging device. In a visual inspection device for a light-transmitting molded article, comprising a recording means for recording data,
The image data is composed of a color space data group 1 composed of light quantity values of red (R), blue (B), and green (G) color components, and the color space data group 1 is represented by hue (H). And a color space data group 3 by rotating the hue (H) of the color space data group 2 and the first conversion means for converting the color space data group 2 composed of color components including color saturation (S). A second conversion means for converting, and a third color space data group 3 for converting the color space data group 3 into a color space data group 4 composed of light quantity values of red (R), blue (B), and green (G) color components. Conversion means,
Appearance of a light-transmitting molded article, wherein one of the color components constituting the color space data group 4 is selected, and the light quantity value of the selected color component is corrected using the light quantity values of the remaining color components Inspection equipment. (A “color space data group” used in this specification refers to a collection of pixel data expressed by color components such as an HSI color space and an RGB color space constituting image data.)

  Next, the invention described in claim 2 is characterized in that the correction subtracts a value of a component having a small light amount value from the remaining color components from which the light amount value of the selected color component is not selected. It is an external appearance inspection apparatus of the light transmissive molded article of Claim 1.

The invention described in claim 3 is characterized in that the correction subtracts a value of a component having a large light amount value from the remaining color components from which the light amount value of the selected color component is not selected. The light-transmitting molded product appearance inspection apparatus according to claim 1.

  Furthermore, the invention described in claim 4 is characterized in that the correction obtains and subtracts the average value of the light amount values of the remaining color components not selected from the light amount value of the selected color component. 1. An appearance inspection apparatus for a light-transmitting molded article according to 1.

  As described below, according to the present invention, it becomes possible to detect partial discoloration such as slight color unevenness generated in a light-transmitting molded product obtained by molding glass or resin having light transparency. In addition, the inspection sensitivity in the appearance inspection is stable and can contribute to the improvement of quality. In addition, since it is possible to perform automatic inspection using a computer, the time required for visual inspection can be shortened, so that the appearance of a light-transmitting molded product that can improve production efficiency such as reduction in production cost An inspection device can be provided.

Configuration diagram showing the overall configuration of the appearance inspection device Flow chart showing image processing procedure of visual inspection apparatus Sensitivity distribution when the present invention is applied to colors with low saturation The figure explaining the change of the color component of 1st Example of this invention The figure explaining the change of the color component of 2nd Example of this invention Schematic diagram of the measuring part of the visual inspection device with a light-shielding wall Schematic of the measuring part of the visual inspection device with reflection illumination Flow chart for explaining the operation of a conventional color unevenness appearance inspection apparatus Schematic diagram showing a conversion table for pseudo color processing provided in the image processing apparatus A flowchart showing a color image processing procedure in a conventional image processing apparatus. The figure explaining the color image processing procedure in the conventional image processing apparatus by plotting in the HSI color space A diagram explaining that the color of spots such as color spots is obscured by ambient light

A mode for carrying out the present invention (hereinafter referred to as an embodiment) will be described.
FIG. 1 is a configuration diagram showing the overall configuration of an appearance inspection apparatus according to the present invention. In order to inspect the appearance of the inspection object S formed using glass / resin having light transmittance as a raw material, it is installed on the diffusion plate 3 which also serves as an inspection table. The light source 2 is arranged below the diffusion plate 3 and is configured to radiate light toward the diffusion plate 3, and the light L emitted from the light source passes through the diffusion plate 3 to become diffused light L ′. The object S is illuminated from a random direction. A color imaging device 1 is installed above the inspection object S, and images the inspection object S illuminated by the light emitted from the light source 2.

Image data photographed by the color imaging device 1 is transferred to the image processing device 4 for performing image processing of the present invention. The image processing apparatus 4 includes an input device such as a computer 40 for performing image processing, a printer 41 for outputting data, a monitor 42 for displaying data, and a keyboard / mouse for operating the image processing apparatus 4. 43. Further, the computer device 40 is installed with an interface unit 40d which is a connection unit for inputting image data captured by the color imaging device 1 to the image processing device 4, a CPU 40a which performs image processing calculation, and an image processing program of the present invention. The main memory 40b and an auxiliary storage device 40c for storing data and the like are included.

Next, the flow of processing of image data according to the present invention will be described.
FIG. 2 is a flowchart showing a procedure of image processing of the appearance inspection apparatus according to the present invention.
The image input S <b> 1 is a process of inputting image data captured by the color imaging device 1 to the image processing device 4. In many cases, the image data is expressed as an image that is divided into light intensity values of three color components of red (R), green (G), and blue (B) for each pixel. There are several known representation methods using hue (H) and saturation (S) as color representation methods that are different from the color representations of the three primary colors of RGB. HSI (Hue, Saturation, Intensity) color space and HSV (HSV) There are what are called Hue, Saturation, Value) color space, HLS (Hue, Lightness, Saturation) color space, and the like.
The HSI conversion S2 is a step of converting image data expressed by RGB color components into a color space expressed by hue (H) and saturation (S). Note that the HSI conversion S2 is an equivalent conversion that only converts the expression method of the color components constituting the image data, and does not change the color of the image that the human observes through the monitor 42 of the visual inspection apparatus.

The extracted color designation S3 is a step of obtaining color information by designating an area in which a noticed color spot or the like is confirmed. When a color sample such as a color spot exists in advance, a color may be designated from the color sample.
The selection color designation S4 is a step of designating whether the color designated in the extraction color designation S3 is to be converted into a color designated, for example, R, G, or B.
Hue rotation S5 is a process of changing the hue of the entire image data so that one of R (red), G (green), and B (blue) appears to be highlighted. The amount of change in hue is the difference between the hue (H) of the color designated in the extracted color designation S3 and the color designated in the selected color designation S4.
The RGB conversion S6 is a step of converting the image data after the hue rotation S5 into image data expressed in the RBG color space.

Steps S7 to S12 are steps (gradation processing) for generating a single color grayscale image from the image data obtained by the RGB conversion S6.
In step S7, the image data expressed in the RGB color space is stored in the memory of the computer device 40 by distinguishing the selected color data from the remaining two unselected data (non-selected color 1 and non-selected color 2). It is a process to do. In the description, it is described that the selected color is G (green) and the remaining non-selected colors have a magnitude relationship of R (red)> B (blue).
In the image processing according to the present invention, three types of correction methods can be implemented as gradation processing methods.

The first method is the method described in Steps S8 and S9, in which the difference between the light amount value of the color component of the selected color and the light amount value of the color component of the color that takes a large value among the non-selected colors is taken. It is.
The second method is a method described in step S10 and step S11, in which the difference between the light amount value of the color component of the selected color and the light amount value of the color component of the color that takes a small value among the non-selected colors is taken. It is.
The third method is a method described in step S12 and step S13, in which the average value of the light amount value of the color component of the selected color and the light amount value of the color component of the non-selected color is obtained and a difference is obtained. .
By performing the above gradation processing on the image data, a single color grayscale image is generated from the relationship between the light quantity values of the color components of the selected color and the non-selected color, and is displayed on the screen S13.
In addition, when the obtained result becomes smaller than 0 in the process S9, the process S11, and the process S12, the obtained value is assumed to be 0.

When the inspection apparatus of the present invention is used, FIG. 3 shows in what gradation other colors are expressed with respect to the selected color by the image processing. In FIG. 3, the selected color is G (green) [corresponding to a hue of 120 degrees], and the intensity of the gradation obtained by the image processing is expressed as sensitivity. Note that the sensitivity is a unitless amount because of the difference between the light amount values of the two color components to be compared.
Since the present invention is intended for the appearance inspection of the light-transmitting molded body, as a state of low saturation, which is a characteristic when the light-transmitting molded body is observed, with respect to the light amount value of the color component of the non-selected color, This is a value obtained by taking as an example a case where the light amount value of the color component of the selected color shows a value 10% higher.
When the selected color is corrected with the maximum value of the non-selected color (step S9), only the color close to the selected color appears as a grayscale image. This is effective when only the target color is cut out when the colors of the parts to be compared are similar. (Fig. 3III))
On the other hand, when the selected color is corrected with the minimum value of the non-selected color (step S11), colors in a wide hue range around the selected color appear as a grayscale image. This is effective in grasping a color similar to a case where an image with low saturation can recognize a slight hue in a gray background and the human eye easily misidentifies the hue. (Fig. 3 (I))
When the average value of non-selected colors is determined and corrected for the third selected color, it has an intermediate property between the above two methods, and the hue difference from the selected color is also expressed as a gray value. Is done. (Fig. 3 (II))
The above three types of methods can be selected according to the state of the object to be inspected.

FIG. 4 is a diagram showing changes in hue and gradation when the first embodiment of the present invention is applied. The left diagram in FIG. 4A is a diagram in which the color components are expressed by hue and saturation, and the hue is expressed in the circumferential direction and the color is expressed in the radial direction. It is a figure generally called a hue circle.
The example of FIG. 4 shows a process of distinguishing by applying the present invention to three kinds of colors (a, b, c) having the same saturation value and close hue values. The second, third, and fourth diagrams from the left in FIG. 4A show the three colors (a, b, and c) of R (red), G (green), and B (blue). The figure at the time of expressing with the color component of a color is shown. 4 (1), for example, if the R, G, and B color components take values of 90, 100, and 110, respectively, the three color components are relatively close to each other. The degree becomes smaller and visually, it is recognized as a color in which the color tone of a single color is slightly colored. If such a color is expressed by a hue circle, it is distributed in the central portion. In the hue circle shown in FIGS. 4 (1) and 4 (2), for convenience of explanation, the color value is greatly emphasized and displayed so that the difference in hue between the colors can be seen.

4 will be described with reference to the flowchart of FIG. From the acquired image, three locations with close color tones are selected, and the color information of each point (a, b, c) is acquired. (FIG. 4 (1) (a) to (c)) (Step S1)
The color information of each point is converted into hue and color and displayed on the hue circle. (FIG. 4 (1) left) (Step S2)
Focusing on the color of point b (step S3), G (green) is selected as the color for hue conversion of the color of point b. (Process S4)
The difference (θ) between the hue value of point b and the hue of G (green) is obtained, and the hue of the entire image is shifted by θ so that the hue of point b is green. (FIG. 4 (2)) (Step S5)
The color information of each point after hue conversion is represented by R, G, and B. (Step S6) (FIG. 4 (2) (a ′) to (c ′))

Here, when correcting the selected color with the maximum value of the non-selected color, the non-selected colors R (red) and B (blue) are selected for the selected color G (green) component. The light quantity values of the color components are compared, and the light quantity value of the large color component is subtracted from the light quantity value of the color component of the selected color G (green) (FIG. 4 (3)) (step S9).
When correcting the selected color with the average value of the non-selected colors, the R (red) and B (blue) color components of the non-selected color are compared with the G (green) component of the selected color. The average value of the light quantity values is obtained and subtracted from the light quantity value of the color component of the selected color G (green) (FIG. 4 (4)) (step S12).
The process S9 or S12 is repeated for the entire image, and the obtained image is displayed. (Process S13)
As described above, by performing image processing according to the procedure according to the present invention, the matching color is maximized with respect to the focused color, and even if the hue is close, it is distinguished as a difference in shade of gradation. be able to.
Note that, as described with reference to FIG. 3, the gradation of light and shade obtained from the difference between adjacent hues can be adjusted by selecting a correction method. As shown in FIG. 4, when hues are close to each other, it is suitable to correct the color component of the non-selected color using a maximum value or an average value.

Although the method of distinguishing when the hues are close to each other has been described, colors with low saturation are recognized as a single grayscale image even if the hue values are not close. It is difficult to find a feature, but rather, it creates a visual illusion in relation to the surrounding colors, which may be recognized as a color different from the original color.
Even in such a case, it is effective to perform the image processing of the invention.
If it demonstrates using FIG. 5, three places will be selected from the acquired image and the color information of each point (a, b, c) will be acquired. (FIG. 5 (1) (a) to (c)) (Step S1)
The color information of each point is converted into hue and color and displayed on the hue circle. (FIG. 5 (1) left) (Step S2)
Focusing on the color of point b (step S3), G (green) is selected as the color for hue conversion of the color of point b. (Process S4)
The difference (θ) between the hue value of point b and the hue of G (green) is obtained, and the hue of the entire image is shifted by θ so that the hue of point b is green. (FIG. 5 (2)) (Step S5)
The color information of each point after hue conversion is represented by R, G, and B. (Step S6) (FIG. 5 (2) (a ′) to (c ′))
Here, with respect to the G (green) component that is the selected color, the light amount values of the R (red) and B (blue) color components that are the non-selected colors are compared, and the light amount value of the small color component is Subtract from the light quantity value of the color component of the selected color G (green) (FIG. 5 (3)) (step S11).
The above-described step S11 is repeated for the entire image, and the obtained image is displayed. (Process S13)
In this way, by performing image processing according to the procedure according to the present invention, it is possible to remove complementary colors from the focused color, so even if the saturation is low and it is difficult to determine the color characteristics, It is possible to prevent visual illusion.

Although the inspection method and the inspection apparatus of the present invention have been described using the overall configuration diagram of FIG. 1, as described above, the inspection method and apparatus aim to efficiently determine a subtle difference in color tone. Therefore, various lights such as movements of people and objects around the inspection machine and indoor lighting are detected by the color imaging device 1 as stray light, which may cause erroneous detection.
Thus, in order to prevent the influence of the surrounding environment, it is effective to use the light shielding wall 5 as shown in FIG. 6 so as to cover the inspection object S, the color imaging device 1 and the like.
By providing the light shielding wall 5, only the light emitted from the light source 2 is captured by the color imaging device 1, so that erroneous detection due to stray light can be prevented.
In addition to preventing false detection, it is expected that detection sensitivity is improved by eliminating stray light. For this reason, it becomes possible to determine a finer change in color tone with high accuracy.

The light source 2 that irradiates the inspection object S is disposed below the inspection object S in FIGS. 1 and 6, and is disposed so as to be transmissive illumination with respect to the color imaging device 1. In the present invention, the position of the light source 2 is not limited to transmitted illumination. As shown in FIG. 7, it may be installed on the side of the color imaging device 1 to provide reflected illumination.
Further, if the diffusion plate 3 has the function of a light guide plate, the diffused light L from the lower side of the inspection object S can be obtained by installing the light source 2 on the side wall of the diffusion plate 3 and introducing light from the side surface of the diffusion plate. 'Can be irradiated, and irradiation can be performed in the same manner as in the case of the transmitted illumination in FIG.

S: Inspected object L: Light emitted from a light source L ′: Diffused light 1: Color imaging device 2: Light source 3: Diffusion plate 4: Image processing device 40: Computer device 40a: CPU
40b: main memory 40c: auxiliary storage device 40d: interface unit 41: printer 42: monitor 43: input device 5: light shielding wall

Claims (4)

A light source for irradiating the inspection object with light;
A color imaging device that receives light that is transmitted through the inspection object or reflected from the inspection object; and
Recording means for recording image data captured by the color imaging device;
In the visual inspection apparatus for light-transmitting molded products,
The image data is composed of a color space data group 1 composed of light quantity values of red (R), blue (B), and green (G) color components,
First conversion means for converting the color space data group 1 into a color space data group 2 composed of color components including hue (H) and saturation (S);
Second conversion means for rotating the hue (H) of the color space data group 2 to convert it into the color space data group 3;
A third conversion means for converting the color space data group 3 into a color space data group 4 composed of light quantity values of red (R), blue (B), and green (G) color components;
Appearance of a light-transmitting molded article, wherein one of the color components constituting the color space data group 4 is selected, and the light quantity value of the selected color component is corrected using the light quantity values of the remaining color components Inspection device. 2. The light-transmitting molded article according to claim 1, wherein the correction subtracts a value of a component having a small light amount value from the remaining color components not selected from the light amount value of the selected color component. Visual inspection equipment. 2. The light-transmitting molded article according to claim 1, wherein the correction subtracts a value of a component having a large light amount value from the remaining color components not selected from the light amount value of the selected color component. Visual inspection equipment. The visual inspection of the light-transmitting molded product according to claim 1, wherein the correction is performed by obtaining and subtracting an average value of the light amount values of the remaining color components not selected from the light amount value of the selected color component. apparatus.

Images