JP2009115512A - Article inspection method - Google Patents

Abstract

A determination means 6 creates a three-dimensional basic image 8 for the entire inner surface 2A of a tire 2 based on 4000 times of image data taken by a camera 5.
The first calculation unit 6B of the determination unit 6 obtains an image (G1) of the contour line 11 of the inner surface 2 of the tire 2 from which noise is removed based on the basic image 8.
Next, the second calculation unit of the determination unit 6 obtains an image (G2) of only the contour line 11 from which noise and defective portions are removed, based on the basic image 8.
Next, the comparison unit 6D compares G1 and G2 to obtain an image (G3) in which the defective portion is emphasized. The determination unit 6F determines whether the appearance of the inner surface 2A of the tire 2 is good based on the image (G3) in which the defective portion is emphasized.
[Effect] The quality of the appearance of the inner surface 2A of the tire 2 can be accurately determined.
[Selection] Figure 1

Description

  The present invention relates to an article inspection method, and more particularly to an article inspection method suitable for inspecting the appearance of an article having a curved portion on a surface such as a tire.

Conventionally, an inspection method for inspecting the appearance of a tire is known (for example, Patent Document 1 and Patent Document 2).
In such a conventional inspection method, first, a line-shaped laser beam is irradiated toward the side surface of the tire and the reflected light of the laser beam is photographed by a camera while rotating the tire. Next, the shape of the tire as the object to be inspected is detected from the image of the laser beam obtained by the camera, and the tire shape is compared with legitimate reference data stored in advance in the judging means, so that the appearance of the tire is good or bad Is determined.
Japanese Patent Laid-Open No. 11-138654 Japanese Patent Laid-Open No. 2003-240521

By the way, in the conventional inspection apparatus, it is necessary to prepare the determination reference data as the reference for pass / fail determination in advance for the entire area of the tire to be inspected, and therefore it is very complicated to create such determination reference data. There were drawbacks.
Further, in the conventional inspection method, since the reference data prepared in advance and the actually measured data are compared for all the pixels constituting the image data, the position of the inspection target region of the tire is determined when obtaining the actually measured data. You must know exactly. For this reason, there is a drawback that the positioning accuracy when moving the tire needs to be strict, and the configuration of the inspection apparatus becomes complicated by the amount that requires positioning means with high positioning accuracy.

In view of the above-described circumstances, the present invention moves the object to be inspected and the line-shaped light while irradiating the object to be inspected with a line-shaped light, so that the place where the object is irradiated with the light. In the article inspection method in which the photographing means is photographed a plurality of times and the quality of the appearance of the inspection object is determined based on the data of the image photographed by the photographing means.
A three-dimensional basic image of the inspection object is created on the basis of the data of the inspection object image captured by the imaging means a plurality of times, and then each pixel constituting the three-dimensional basic image For each first predetermined number of pixels including the positions of the pixels, a first average value that is an average height is calculated, and all the pixels constituting the basic image are obtained for them. The first average image data from which noise has been removed is obtained by replacing the first average value, and then, for each pixel constituting the three-dimensional basic image or the first average image data, The second average value obtained by calculating a second average value that is an average value of heights for a second predetermined number of surrounding pixels including the position, and obtaining all the pixels constituting the basic image for them. Replaced the defective part and noise. The average image data is obtained, and then the height difference between the first average image data and the second average image data is obtained to create enhanced image data relating to the height difference, and the height in the enhanced image data The quality of the inspection object is determined based on the difference.

  According to such a configuration, since it is not necessary to create the determination reference data before the inspection is performed, the preparatory work before the inspection can be simplified as compared with the conventional case, and the appearance of the inspection object can be simplified. Pass / fail can be accurately determined.

The present invention will be described below with reference to the illustrated embodiments. In FIG. 1, reference numeral 1 denotes a tire inspection apparatus for inspecting the appearance of the inner surface 2A of a tire 2 as an object to be inspected. The tire inspection apparatus 1 includes a rotary table 3 that rotates while holding a tire 2 to be inspected, and a linear laser beam L that is arranged at a predetermined position toward an inner surface 2A of the tire 2 on the rotary table 3. Based on the light projecting means 4 for irradiating, the camera 5 attached to a robot hand (not shown) and photographing the inner surface 2A of the tire 2 where the laser light L is radiated, and the image data taken by the camera 5 And determining means 6 for determining whether the appearance of the inner surface 2A of the tire 2 is good or bad.
The rotary table 3 is supported horizontally and can be rotated clockwise in conjunction with the motor 7. When the tire 2 to be inspected is supplied in a rollover state on the rotary table 3 with the rotary table 3 stopped, the axis of the tire 2 is rotated by a fixing means (not shown) provided on the rotary table 3. The tire 2 is positioned so as to substantially coincide with the rotation center of the table 3 and is held at that position.
When the tire 2 is fixed on the turntable 3 as described above, the laser light L is irradiated from the light projecting means 4 toward the inner surface 2A of the tire 2 and the motor 7 is operated to turn the turntable. 3 is rotated once at a constant speed in the clockwise direction.

When the line-shaped laser light L is irradiated from the light projecting means 4 toward the inner surface 2A of the tire 2, the irradiated portion of the laser light L from a direction that forms an angle of 45 degrees with respect to the irradiation direction of the laser light L The reflected light from the camera is photographed by the camera 5. Thereby, the irradiation part (inner surface 2A of the tire 2) of the laser beam L photographed by the camera 5 has a generally arcuate shape. FIG. 3 shows the image data of the inner surface 2A of the tire 2 obtained by one photographing with the camera 5 as an image.
Thus, the data of the image of the inner surface 2A taken by the camera 5 is stored in the primary image storage unit 6A of the determination means 6. In the present embodiment, while the tire 2 is rotated once in conjunction with the motor 7, the laser light L of the inner surface 2A of the tire 2 is irradiated every 0.5 mm of the relative movement distance between the tire 2 and the laser light L. The camera 5 takes a total of 4000 shots of the area where the image is present.
The primary image storage unit 6A of the determination unit 6 stores data of 4000 shots sequentially transmitted from the camera 5, and uses the principle of triangulation based on the stored image data. A three-dimensional basic image 8 shown as an image is created (see S1 in FIG. 2). Then, the determination means 6 recognizes the basic image 8 and determines whether the appearance of the inner surface 2 </ b> A of the tire 2 is good or not after performing an averaging process twice described later.

By the way, there is a case where a bulging portion (bulging) having a diameter of about 2 mm, which is difficult to distinguish with the naked eye, is formed on the inner surface 2A of the tire 2 to be inspected, and such a bulging portion is a defective portion. Therefore, when there is such a defective portion, it is necessary to determine that the tire 2 is defective. On the other hand, the inner surface 2A of the tire 2 has “whiskers” that are small protrusions when the tire 2 is molded, and the “inner surface” 2A of the tire 2 is also attached with “powder” such as a material. is there. Such “beard” and “powder” do not become defective parts, but the image data of the inner surface 2A of each tire 2 photographed by the camera 5 and the above basic image 8 show such “beard” and “powder”. Will be included as noise.
When it is attempted to determine whether or not there is a bulging portion which is the defective portion on the inner surface 2A of the tire 2 based on the image data including such noise and the three-dimensional basic image 8, such noise is generated. There is a risk of erroneously determining that it is a defective part.
Therefore, the determination means 6 of the present embodiment first performs an averaging process based on the image data stored in the primary image storage unit 6A and the three-dimensional basic image 8 to each camera 5. Noise (whiskers and dust) is removed from the image data and the three-dimensional basic image 8 described above.

That is, FIG. 5 is a conceptual diagram showing image data obtained by the first photographing by the camera 5 stored in the primary image storage unit 6A. Here, a curve that continues in a mountain shape shows the contour of the inner surface 2A of the tire 2. Line 11. The left-right direction in FIG. 5 corresponds to the width direction of the tire 2, and the up-down direction indicates the height direction.
In FIG. 5, thin lines extending in the height direction from a plurality of locations of the contour line 11 indicate the noise 12 that becomes the above-mentioned “beard” and “powder” locations, and the height direction of the contour line 11. A long and narrow triangle extending in a circle indicates the bulging portion 13 which is a defective portion. As described above, the image at each photographing stored in the primary image storage unit 6A is in a state where the bulging portion 13 which is a defective portion and the noise 12 which is not a defective portion are mixed.
Therefore, in this embodiment, the noise 12 such as “beard” or “powder” causes an erroneous determination when determining the bulging portion 13, and therefore, averaging processing is performed to remove the noise 12.

That is, FIG. 9 represents each time image stored in the primary image storage unit 6A as a horizontal row of pixels, and represents each time image sequentially captured by the camera 5 in the vertical direction. That is, the three-dimensional basic image 8 shown in FIG. 4 is expressed as an array of pixels. In the present embodiment, when the tire 2 is rotated once, a total of 4000 images are taken by the camera 5, so that the first and 4000th images are adjacent to each other in the vertical direction. Actually, the number of pixels in each time in the horizontal direction is 1280, and the number of pixels in the vertical direction is 4000. For convenience, only a part is shown in FIG.
By the way, the size of the bulging portion 13 which is a defective portion is generally larger than the total number of pixels in the vertical and horizontal directions as indicated by the imaginary line circle 14 in FIG. Therefore, in the first image data stored in the primary image storage unit 6A, the first calculation unit 6B of the determination unit 6 firstly determines the position of the pixel a1 and the surrounding nine pixels including the position. A first average value that is an average value of the heights is obtained and stored, and the height is replaced with the first average value obtained as described above for the position of the pixel a1 in the basic image 8. (See FIGS. 1 and 2).
Next, in the first-time image data shown in FIG. 9, with respect to the position of the pixel a2, the first height that is the average value of the heights of a total of nine pixel positions in the vertical and horizontal directions with the position as the center. The average value is obtained and stored, and the height is replaced with the first average value obtained as described above for the position of the pixel a2 in the basic image 8.
In this way, the first calculation unit 6B obtains the first average value that is the average value of the heights of the nine positions around the positions of the pixels a1 to a1280 in the first image, and a1 The position of the pixel of ~ a1280 is replaced with the obtained first average value.
Thereby, the first average image data G1 can be obtained for the first captured image (see FIG. 6). In the first average image data G1, the noise 12 is removed from the first image data shown in FIG. 5, and the contour 11 and the bulging portion 13 extending in the height direction remain. The process for obtaining the first average image data G1 in this way is the first averaging process S2 (FIG. 2).

Next, when the first averaging process S2 is performed on the first image of the camera 5 as described above, the second calculation unit 6C of the determination unit 6 uses the image data stored in the primary image storage unit 6A. Based on the first image, the second average value, which is the average value of the heights of a total of 81 pixels for each of the nine vertical and horizontal positions of the adjacent position including the position, for the position of the pixel a1 described above. It is obtained and saved, and the height of the position of the pixel a1 in the basic image 8 is replaced with the second average value obtained as described above (see FIGS. 1, 2, and 9).
Next, the second calculation unit 6C obtains the second average value that is the average value of the heights of a total of 81 pixels in the vertical and horizontal positions of the adjacent position including the position for the pixel of a2. This is stored, and the height of the pixel a2 in the basic image 8 is replaced with the second average value obtained as described above.
In this way, the second calculation unit 6C obtains the second average value that is the average value of the heights of the 81 pixels around the pixels a1 to a1280 in the first image. , A1 to a1280 pixel positions are replaced with the second average values obtained as described above.
Thereby, the second average image data G2 can be obtained for the first captured image (see FIG. 7). In the second average image data G2, the noise 12 and the bulging portion 13, which is a defective portion, are removed from the first image data shown in FIG. 5, and only the outline 11 remains. The process for obtaining the second average image data G2 in this way is the second averaging process S3 (FIG. 2).

The reason for obtaining the second average value for a total of 81 pixels in the second averaging process S3 is that the size of the bulging portion 13 that is a defective portion is as shown by an imaginary line circle 14 in FIG. The size is slightly larger than the size of several pixels. Therefore, the second average value is obtained as described above for a total of 81 pixels for nine vertical and horizontal dimensions that can completely include the circle 14 as the bulging portion 13.
In the first average image data G1 shown in FIG. 6, the noise 12 is removed, and the data of the contour line 11 of the tire 2 including only the bulging portion 13 of the defective portion is obtained, while the data shown in FIG. In the second average image data G2, only data of the contour line 11 of the tire 2 from which the noise 12 and the bulging portion 13 which is a defective portion are removed is obtained.

Therefore, next, the comparison unit 6D of the determination unit 6 uses the first average image data G1 shown in FIG. 6 and the second average image data shown in FIG. 7 for the positions of the pixels a1 to a1280 for the first shooting. A difference in height of the contour line 11 obtained by subtracting G2 from G1 is obtained by comparing with G2 (see FIGS. 1 and 2). Then, when the comparison unit 6D obtains the height difference between the contour lines 11 of the image data G1 and G2 for each of the pixels a1 to a1280, the comparison unit 6D creates and saves the emphasized image data G3 based on the difference. (FIGS. 2 and 8). The process of creating and saving the enhanced image data G3 is an image enhancement process S4. In the emphasized image data G3, the height of the bulging portion 13 which is a defective portion is emphasized.
The determination means 6 of the present embodiment performs the averaging processes S2 and S3 twice and the subsequent image enhancement process S4 on the captured image of the first camera 5 as described above, and thereafter the determination means. 6, for the remaining 2 to 4000 shot image data, the same two averaging processes S2 and S3 and subsequent image enhancement as in the case of the first shot image data described above. Processing S4 is performed (see FIG. 2).
Thereby, as shown as an image in FIG. 10, it is possible to obtain data in a state where 4000 pieces of enhanced image data G3 are arranged at the same pitch and at the same height. In FIG. 10, for example, the four arcuate portions indicate the bulging portions 13 that are defective portions.
In the present embodiment, the averaging process S2 and S3 and the image enhancement process S4 are performed twice at the position of each pixel in the basic image 8 based on the image data obtained by 4000 times of shooting. Thus, 4000 pieces of enhanced image data G3, which is three-dimensional data about the inner surface 2A of the tire 2 to be inspected, are obtained.

The 4000 enhanced image data G3 obtained as described above is three-dimensional data. Therefore, next, the image processing unit 6E of the determination unit 6 replaces the 4000 enhanced image data G3, which is 3D data, with 2D image data (S5 in FIGS. 1 and 2).
At this time, as shown in FIG. 11, the emphasized image data G3 includes XY directions orthogonal to each other on the horizontal plane and data in the height direction (Z direction) orthogonal to these, Replaces the magnitude in the height direction (Z direction) of the enhanced image data G3 for each time with the magnitude of the luminance (FIG. 12).
In each of the enhanced image data G3, the bulging portion 13 that is a defective portion has a larger value in the height direction (Z direction) than other normal portions. By performing the conversion processing S5, the luminance of the bulging portion 13 which is a defective portion is recognized by the determination unit 6F of the determination means 6 as a larger luminance than other normal portions (FIG. 1). , FIG. 2).
Thereafter, the determination unit 6F of the determination unit 6 replaces a portion with high luminance (the bulging portion 13 which is a defective portion) in the two-dimensionalized enhanced image data G3 each time with white and also a portion with low luminance (normal) A binarization process for replacing (part) with black is performed (FIGS. 1 and 2).
Thereafter, when the determination unit 6F of the determination unit 6 recognizes that there is a white spot in each of the emphasized image data G3 after the binarization process, the tire 2 that is the inspection target is defective. It is determined that there is a bulging portion 13 that is a portion, and the fact that the tire 2 is defective is displayed on a display device (not shown). On the other hand, the determination means 6 determines that the tire 2 to be inspected is a non-defective product when there is no white portion and the whole is black for the emphasized image data G3 each time after the binarization process. This is displayed on the display device.

According to the present embodiment described above, it is not necessary to create determination reference data that is a determination criterion for pass / fail for the entire inner surface 2A of the tire 2 to be inspected before the inspection. Therefore, according to the present embodiment, it is possible to simplify the preparatory work before performing the inspection, as compared with the conventional inspection method described above.
Further, since the judgment reference data stored in advance and the measured data taken by the camera are not compared to judge the quality of the inner surface 2A of the tire 2, positioning means in the rotational direction when the tire 2 is rotated is required. do not do. Therefore, the structure of the whole tire inspection apparatus 1 can be simplified compared with the conventional case which requires such positioning means.
Further, in this embodiment, since the weighted image data G3 is obtained after performing the averaging processes S2 and S3 twice, the quality of the appearance of the inner surface 2A of the tire 2 is determined based on that. The quality of the appearance of the inner surface 2A can be accurately determined.
In addition, according to the present embodiment, even when a gentle bulging portion 13 (bulging) that cannot be accurately detected by two-dimensional inspection (2D inspection) occurs on the inner surface 2A of the tire 2, It can be accurately determined that there is a bulging portion 13 as a defective portion.

In the above embodiment, the binarization process is performed after the enhanced image data G3 is two-dimensionalized. However, a gray scale process or a pattern matching process may be performed instead of the binarization process. That is, the gray scale process recognizes the difference in luminance of the two-dimensional image with the brightness of 256 gradations, and determines the brightness of the image after the above-described enhanced image data G3 is two-dimensionalized. The quality of the tire can be determined by comparing with the threshold value. When performing pattern matching processing, a basic image of a two-dimensional image is stored in advance, and the above-described two-dimensional image of the emphasized image data G3 is compared with a basic image stored in advance. The quality of the tire can be determined.
Further, in the above embodiment, after obtaining the emphasized image data G3, the two-dimensional processing S5 and the binarization processing S6 are sequentially performed on the emphasized image data G3. However, these processing may be omitted. That is, as shown by an imaginary line in FIG. 2, when the emphasized image data G3 is obtained in the image enhancement processing S4, as shown in FIG. 13, a predetermined threshold value δ is exceeded in the emphasized image data G3. It may be determined whether or not there is a place where the tire 2 is present, and if there is a place where the threshold value is exceeded, the tire 2 may be determined to be defective.
Further, in the above-described embodiment, the first average image data G1 and the second average image data G2 are obtained based on the image data of each time stored in the primary image storage unit 6A and the basic image 8, but the second average The image data G2 may be obtained based on the first average image data G1 obtained in advance, instead of the image data of each time and the basic image 8. That is, the second average image data G2 may be obtained by replacing the first average value obtained in the first averaging process with each pixel and performing the second averaging process based on the first average value.

BRIEF DESCRIPTION OF THE DRAWINGS The whole block diagram which shows one Example of this invention. The figure which shows the process process by the determination means 6 shown in FIG. The conceptual diagram which shows the image by one imaging | photography with the camera 5 shown in FIG. The conceptual diagram which shows the basic image 8 produced by the determination means 6 shown in FIG. The conceptual diagram which shows the image by one imaging | photography with the camera 5 shown in FIG. The conceptual diagram which shows the image after the 1st averaging process by the determination means 6 shown in FIG. The conceptual diagram which shows the image after the 2nd averaging process by the determination means 6 shown in FIG. The conceptual diagram which shows the image after the image enhancement process by the determination means 6 shown in FIG. The figure which showed the picked-up image by the camera 5 shown in FIG. 1 as an arrangement | sequence of a pixel. The conceptual diagram which shows the continuous state of the image after the image enhancement process by the determination means 6 shown in FIG. The figure which expressed the required location in the image after the image enhancement process by the determination means 6 shown in FIG. 1 by the three-dimensional position. The figure which shows the state after carrying out the two-dimensional process of the image shown in FIG. The conceptual diagram which shows the principal part in the other Example of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 2 ... Tire 2A ... Inner surface of tire 2 5 ... Camera 6 ... Determination means 8 ... Basic image G1 ... 1st average image data G2 ... 2nd average image data G3 ... Enhanced image data

Claims (6)

While irradiating the inspected object with line-shaped light, relatively moving the inspected object and the line-shaped light, and photographing the portion where the light is irradiated on the inspected object multiple times by the imaging means, In the article inspection method in which the quality of the appearance of the inspection object is determined based on the data of the image captured by the imaging unit,
Based on the data of the image of the inspected object taken multiple times by the imaging means, create a three-dimensional basic image about the inspected object,
Next, for each pixel constituting the three-dimensional basic image, a first average value, which is an average height value, is calculated for a first predetermined number of pixels surrounding the pixel position. The first average image data from which noise has been removed is obtained by replacing all the pixels constituting the basic image with the first average value obtained for them.
Next, for each pixel constituting the three-dimensional basic image or the first average image data, a second predetermined number of pixels including the position of the pixel is an average height value. 2 average value is calculated, each pixel constituting the basic image is replaced with the second average value obtained for them, and second average image data from which defective portions and noise are removed is obtained,
Next, a height difference between the first average image data and the second average image data is obtained, and enhanced image data relating to the height difference is created, and the height difference is determined based on the height difference in the enhanced image data. An article inspection method characterized by determining the quality of an inspection object.   2. The article inspection method according to claim 1, wherein the quality of the inspection object is determined by comparing the height difference of the emphasized image data with a predetermined threshold value.   The difference in height in the emphasized image data is replaced with a luminance level, and the emphasized image data is converted into data of a two-dimensional image. Then, based on the data of the two-dimensional image, the inspection object The article inspection method according to claim 1, wherein quality is determined.   4. The article inspection method according to claim 3, wherein the quality of the inspection object is determined by binarizing the data of the two-dimensional image.   4. The article inspection method according to claim 3, wherein the two-dimensional image data is subjected to gray scale processing to determine whether the inspection object is good or bad.   The article inspection method according to claim 3, wherein the basic data in the two-dimensional image is stored, and the quality of the inspection object is determined by comparing the data of the two-dimensional image with the basic data. .

Images