JP2004220244A - Tooth contact extraction method and device therefor - Google Patents

Abstract

An object of the present invention is to provide a tooth contact extraction method and device capable of detecting a tooth contact surface with high accuracy using Komeitan.
Kind Code: A1 An image capturing apparatus captures a tooth contact tooth surface on which Komunetan has been applied, and converts each of R, G, and B components of the captured RGB color image into a grayscale image by an image processing apparatus to obtain a G component. The grayscale image obtained by adding the luminance of the B component and adding the luminance of the inverted image obtained by inverting the density of the grayscale image and the luminance of the R component is binarized by removing parts other than the tooth contact surface based on the histogram. A method for converting the R, G, and B components into a grayscale image by adding the luminance of the G component and the B component, and calculating the inverted luminance of the grayscale image and the method described above. An image processing apparatus for removing a portion other than the tooth contact surface from the grayscale image to which the luminance of the R component has been added based on the histogram to obtain a binarized image.
[Selection diagram] Fig. 1

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a tooth contact extraction method for extracting a tooth contact surface of a gear such as a differential gear, and an apparatus therefor.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, gears such as hypoid gears and spiral bevel gears including a ring gear and a pinion gear incorporated in a differential of an automobile have been subjected to tooth contact inspection to prevent generation of noise and abnormal noise. In this tooth contact inspection, after applying Kometan to the ring gear, the pinion gear was engaged and the contact surface of the ring gear from which Kometan was scraped off was visually inspected for the center of gravity and area per tooth. However, manual inspection requires time and effort and is inefficient. Therefore, a color image of the tooth surface where a tooth contact surface is formed by applying Kometan to a gear, and then meshing with the gear, is taken. There is a tooth contact measuring device that decomposes into three components (Red) B (Blue) G (Green), binarizes the components, and performs an inspection (for example, see Patent Document 1). In addition, a yellow paint different from RBG is applied to the gears, and the gears are meshed with each other to take a color photograph of the tooth surface having a tooth contact surface, and the tooth surface of the photographed RGB color image is subjected to image processing to reduce the surface contact. There is a method for extracting a surface state to be inspected (for example, see Patent Document 2).
However, according to Patent Document 1, since the RBG color image is directly binarized, there is a problem that a subtle hue is lost and the tooth contact surface cannot be accurately detected. This is because the color of Gwangmyeong Tan varies slightly depending on the lot, the melting condition, the coating condition, etc., and if the RBG color image is directly binarized, the color and the light change depending on the difference, the melting condition, the coating condition, etc. of the Komitsutan lot. This was caused by the inability to accurately discriminate the difference between the color of the tooth contact surface where the red was scraped off. In other words, both Komeitan rubbed and Komeitan with thin coating are the same in that the ground of the gear can be seen through. Further, according to Patent Literature 2, since the RGB components of the captured RGB color image are converted into a histogram, even if the paint to be applied is a specific color (yellow) that does not include the RGB components, the color changes depending on the degree of application and the like. It was difficult to discriminate subtle differences from the shade of the tooth contact surface where the yellow paint was scraped off. In addition, when demonstrating the inspection ability of the equipment, it must be performed based on the judgment of the tooth contact surface by a skilled person.However, since a special yellow paint different from Kometan is used, the expert There is a problem that it is necessary to be skilled in the inspection of the test.
[0003]
[Patent Document 1] JP-A-3-115829 [Patent Document 2] JP-A-9-89533
[Problems to be solved by the invention]
SUMMARY OF THE INVENTION It is an object of the present invention to provide a tooth contact extraction method and device capable of detecting a tooth contact surface with high accuracy by using Komeitan.
[0005]
[Means for Solving the Problems]
In order to achieve the above-described object, the present invention provides a method of photographing a tooth surface on which a tooth contact surface generated by meshing with a gear coated with Komitsutan with a photographing device and photographing each of RGB color images of RGB. The components are converted into a grayscale image by an image processing apparatus, and the brightness of the grayscale image is inverted after adding the brightness of the G component and the brightness of the B component, and the brightness of the inverted image and the brightness of the R component are added. A method for extracting a tooth contact from a grayscale image obtained by removing a portion other than a tooth contact surface based on a histogram to obtain a binarized image according to the first aspect of the present invention. A second aspect of the present invention is a tooth contact extraction method for averaging an RGB color image of a surface for each of R, G, and B components. The present invention provides a photographing apparatus for photographing a tooth surface on which a tooth contact surface generated by meshing a tooth surface coated with Komitsutan, and an RGB component of a photographed RGB color image. A grayscale image obtained by converting the grayscale image into a grayscale image, adding the luminances of the G component and the B component, inverting the density of the grayscale image, and adding the luminance of the inverted image and the luminance of the R component. A tooth contact extraction device comprising an image processing device for removing a portion other than the tooth contact surface based on the histogram and forming a binarized image according to the invention of claim 4, wherein the illumination device for illuminating light is attached. A tooth contact extraction device according to the fifth aspect of the present invention is the invention of claim 5.
[0006]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a preferred embodiment of the present invention will be described in detail based on the apparatus shown in FIG.
Reference numeral 1 denotes an apparatus main body. The apparatus main body 1 includes an image processing apparatus 3 controlled by a computer 2, an imaging device 4, an illumination device 5, and a mounting table 6 on which a gear to be inspected is mounted.
[0007]
The computer 2 instructs image processing by the image processing device 3, controls photographing of the photographing device 4, and controls the tilting operation of the mounting table 6 according to photographing.
[0008]
Further, as shown in FIGS. 2, 4, 5, and 8, the image processing device 3 averages the R, B, and G components of the RGB color image of the tooth surface photographed a plurality of times by the photographing device 4. Processing ((1) in FIG. 5), conversion processing for converting each of R, B, and G components into a grayscale image ((2) in FIG. 5), addition processing for adding the luminance of the G component and the B component of the grayscale (FIG. 2 (2)), an inversion process (FIG. 2 (3)) for inverting the shading of the gray scale image in which the luminance of the G component and the B component is added, and Addition processing for adding the luminance of the R component ((4) in FIG. 2), tooth-contact surface extraction processing for removing a luminance distribution pattern other than the tooth-contact surface based on the grayscale image obtained by the addition based on the histogram, After extracting the contact surface, the grayscale image In addition to the tooth contact extraction processing for performing a binarization processing (FIG. 5B) as a binarized image, and after extracting a tooth contact surface, a black area having a small area is removed from the binarized image (FIG. 5 (a) and expansion / contraction processing (FIG. 5 (b) b) prevent the end of the tooth contact surface from being separated, and delete the area other than the tooth contact surface which is the maximum area (FIG. 5 (3)). c), a false tooth contact surface removal process by a logical product of the first binarized image (FIG. 5B) and the binarized image after noise processing (FIG. 5C) (FIG. 5C). Noise removal processing such as 5.3D), angle conversion processing for converting a perspective binary image to a projected image (FIG. 5D), and calculating the area of the tooth contact surface based on the angle-converted projected image. The area calculation processing (not shown), the oblique RGB color image and the binarized image were subjected to angle conversion processing to the projected image, and the angle conversion was performed. Centroid calculation processing for obtaining the center of gravity of the surface tooth performs a logical AND operation between the luminance and the binarized image of the color image is presumed to be capable of performing an inspection process (FIG. 8) or the like.
[0009]
The photographing device 4 uses a 3CCD color camera that separately captures R, G, and B components of a photographed image into R, G, and B image sensors, and each of the R, G, and B image sensors of the 3CCD color camera. The R, G, and B components of the RGB color image taken into the image processing device 3 can be directly output to the image processing device 3.
[0010]
The illuminating device 5 illuminates the gear to be inspected. Illumination light is incident on the tooth surface in parallel, and the tooth surface is illuminated with uniform brightness. Specifically, as shown in FIG. 1, the illumination device 5 uses an annular white fluorescent lamp so that the lens of the imaging device 4 is located at the center of the ring and the optical axis of the imaging device 4 and the illumination device 5. The deviation from the irradiation optical axis is reduced so that there is no shadow.
[0011]
Further, as shown in FIG. 1, the mounting table 6 mounts the gear to be inspected at an angle of 20 degrees. By tilting the mounting table 6 by 20 degrees, the optical axis of the photographing device 4 can be arranged so that the optical axis of the photographing device 4 is orthogonal to the inclination of the bevel gear-shaped ring gear, and photographing distortion can be reduced. The mounting table 6 is tilted by 45 degrees when photographing the forward surface side of the ring gear so that the tooth surface on the forward surface faces the photographing device 4, and is inclined by 15 degrees when photographing the retreating surface side of the ring gear and retracts. The surface side faces the photographing device 4. This is because the inclination angle of the tooth surface is different between the advance surface side and the retreat surface side.
[0012]
The mounting table 6 is tiltable by an electric motor (not shown), and the tilt angle is controlled by the computer 2. However, the tilt angle may be manually adjusted. Further, a drive motor (not shown) is provided on the mounting table 6 so that the mounted ring gear can be driven. The contact of the ring gear is inspected while being housed in a differential case (not shown) of the differential device. The ring gear to be photographed is inspected through an inspection opening of the differential case.
[0013]
Next, the tooth contact surface extraction processing will be described. First, Komeitan is applied to the ring gear of the differential gear. Thereafter, the differential case incorporating the differential gear is mounted on the mounting table 6 inclined at 20 degrees. Then, a driving motor for rotating and driving the differential gear is connected to the ring gear in order to scrape off the Komeitan, and then the driving motor is driven. Next, the inspection opening of the differential case is arranged so as to face the photographing device 4, and the tooth upper surface of the ring gear for inspecting the tooth contact surface is arranged so as to be parallel to the photographing device. Further, depending on whether the tooth surface to be photographed is on the advancing surface side or the retreating surface side, the mounting table 6 is turned by turning the 20-degree inclined surface 45 degrees to the left or 15 degrees to the right, and the photographing surface is turned to the photographing device 4 and the photographed image Should be as oblique as possible.
[0014]
During photographing, the tooth surface is illuminated by the white light illuminating device 5. The illuminating device 5 is white illumination using an annular fluorescent lamp, and the lens of the photographing device 4 is arranged at the center of the annular hole of the fluorescent lamp. The direction substantially coincides with the direction, and the shadow of the adjacent gear is not projected on the tooth surface to be photographed. In addition, the white light causes the illumination to be indirect light, so that the tooth surface is illuminated with uniform brightness. As a result, the load of image processing can be greatly reduced.
[0015]
In this way, the tooth surface is illuminated without shadow, and the same tooth surface is photographed a plurality of times from the same position by the photographing device 4 composed of a 3CCD color camera. Since the captured RGB color image is input to each of the R, G, and B image sensors, the R, G, and B components are extracted from each of the R, G, and B image sensors, and the R, G, and B components are extracted by the image processing device 3. Each time, the luminance of each photographed image on the same pixel is added, and the luminance is averaged by dividing the added luminance by the number of photographed images.
[0016]
Next, the image processing device 3 converts the image for each of the R, G, B components into a gray scale of 256 gradations. When grayscaled, the G component and the B component have a luminance of approximately 100 or less, and the R component has a luminance of approximately 100 or more. This is because Kometan applied to the ring gear that makes the tooth contact surface visible can be red.
[0017]
Next, the image processing device 3 adds the luminances of the G component and the B component of the gray scale. As a result, by adding the G component and the B component that do not include the red component of Gwangmyeongtan, the state of the tooth surface that has been thinned due to the rubbing of the Gumyeongtan due to tooth contact can be more clearly expressed. Next, the image processing device 3 inverts the shading of the grayscale image to which the G component and the B component have been added. This is a filter process for converting a positive image into a negative image.
[0018]
Next, the image processing device 3 adds the luminance of the grayscale inverted image to the luminance of the grayscale R component. This makes it possible to further clarify the subtle adhesion state of Komyotan, for example, the luminance difference between the lap step, and to clearly express the tooth contact surface and the non-tooth contact surface portion as gray scale luminance. Lapping step is that when a gear is hardened and a tooth surface is wrapped by a combination of a ring gear and a pinion gear by rocking and rotation by a lapping machine, a wrapping residue is generated around the periphery as shown in FIG. This portion refers to a portion where the Komitsutan appears to be darkened as if it had been scraped off in the same manner as the tooth contact surface because the Komitsutan was difficult to adhere to.
[0019]
After the tooth contact surface is clarified in this way, the image processing device 3 converts the grayscale image into a binarized image after removing portions other than the tooth contact surface. The removal of the tooth contact surface is performed based on the histogram shown in FIG. In the histogram, the tooth contact surface has a luminance distribution pattern in which similar luminances are gathered. Therefore, the threshold was set so that the luminance distribution pattern other than the tooth contact surface was removed, and the luminance distribution pattern other than the tooth contact surface was removed. Thereafter, the grayscale image is binarized. The threshold may be provided in a valley where the luminance of the tooth contact surface is lowered. However, as shown in FIG. 3, it is difficult to determine the luminance distribution between the tooth contact surface and the non-tooth contact surface. A slightly higher non-tooth contact surface side luminance is set as a threshold. Thereby, the distinction between the tooth contact surface and the non-tooth contact surface can be clarified. At this time, if the shift amount from the valley portion is too large, portions other than the tooth contact surface become large. Therefore, by setting the luminance shifted from the valley portion to the five non-tooth contact surface side, noise entering the tooth contact surface can be reduced. Note that the luminance is 256 gradations.
[0020]
After the binarization in this way, the image processing device 3 performs a noise removal process in order to increase the inspection accuracy of the tooth contact surface. In this noise elimination, first, a black area scattered in a small area is removed. Next, noise removal processing is performed by expanding / contracting pixels that are slightly interrupted from the pixels on the tooth contact surface. This processing is intended to prevent truncation of the pixel of the tooth contact surface, the pixel of the interference line, and the like which are being interrupted. Specifically, as shown in FIG. 7, each of the pixels that are being interrupted is surrounded by eight pixels and expanded to nine pixels. As a result, as shown in FIG. 7, the seven U-shaped black pixels are expanded into a square composed of 25 pixels. Next, 16 adjacent pixels surrounding the square of 25 pixels are removed and contracted into 6 square pixels.
[0021]
Due to the expansion and contraction, the end of the tooth contact surface is prevented from being divided, so that the pixel is divided into a series of pixels on the tooth contact surface and pixels on the non-tooth contact surface. At this time, since the tooth contact surface has the maximum area, the image processing device 3 performs a noise removal process of extracting a maximum area portion from the image and deleting pixels other than the maximum area.
[0022]
Subsequently, the image processing device 3 performs an AND operation of the binarized image shown in FIG. 5B and the noise-removed binary image shown in FIG. By removing the false tooth contact surface generated by the above, a binarized image shown in FIG.
[0023]
Next, the binarized image (FIG. 5 (2) d) obtained in this way is subjected to angle conversion as shown in FIG. 5 (4), and the perspective image is changed to a projection image. This is because the photographed tooth surface is formed in an arc shape and becomes a perspective image, which is necessary when obtaining the area of the tooth contact surface and the center of gravity. Further, since the arc surfaces of the tooth surface on the advancing surface side and the retreating surface side are reversed, the angle conversion is performed in accordance with the respective arc surfaces of the irregularities.
[0024]
The area of the tooth contact surface is calculated by counting the number of pixels in the projection image subjected to the angle conversion as described above. The center of gravity of the tooth contact surface is determined from the luminance of the RGB color image shown in FIG. 5A converted to the projected image and the binarized image shown in FIG. 5D converted to the projected image. . At this time, since the RGB color image is a perspective image, it is converted into a projection image by performing angle conversion, and then logically ANDed with the projection image obtained by performing angle conversion of the binarized image shown in FIG. It is assumed that a calculation is performed to find a portion having the highest luminance in the tooth contact surface and to determine the center of gravity of the tooth contact surface.
[0025]
Note that, in the preferred embodiment, the influence of fluctuations in external brightness and the like is reduced by performing the photographing by the photographing device 4 a plurality of times and calculating the average value, but the photographing is performed only once. Of course, an image may be used. In the preferred embodiment, an image is taken by a three-CCD color camera, R, G, and B components are extracted from each of the R, G, and B image sensors and input to the image processing device 3. When used, the image processing apparatus 3 may of course decompose the RGB color image into R, G, and B components.
[0026]
【The invention's effect】
As is apparent from the above description, the present invention captures a tooth surface on which a tooth contact surface generated by meshing gears coated with Kometan with a photographing device and captures each of RGB color images of the captured RGB color image. The components are converted into a grayscale image by an image processing apparatus, and the brightness of the grayscale image is inverted after adding the brightness of the G component and the brightness of the B component, and the brightness of the inverted image and the brightness of the R component are added. By removing the non-tooth contact surface based on the histogram from the grayscale image obtained as described above and forming a binary image, the tooth contact surface can be inspected reliably and accurately, which is extremely effective for labor saving. Become. Moreover, since the grayscale image before binarization is the same image as when viewed visually, it is possible to visually check the image to check whether the device has accurately detected the tooth contact surface, and to tune and set the device. It's easy.
[0027]
As described in claim 2, the RGB color image obtained by photographing the same tooth surface a plurality of times is averaged for each of R, G, and B components, so that the influence of fluctuations in external brightness and the like can be reduced. The detection of the surface is assured.
[0028]
As described in claim 3, by removing noise from the binarized image, there are various advantages such as a highly accurate inspection of the tooth contact surface.
Therefore, the present invention greatly contributes to the development of the industry as a tooth contact extraction method and an apparatus for solving the conventional problems.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram showing a configuration of a preferred embodiment of the present invention.
FIG. 2 is an explanatory diagram showing a process of image processing and enhancing a tooth contact surface in a preferred embodiment of the present invention.
FIG. 3 is a histogram showing luminance of a tooth surface according to the preferred embodiment of the present invention.
FIG. 4 is a flowchart showing an inspection process according to a preferred embodiment of the present invention.
FIG. 5 is an explanatory diagram showing an inspection process of a tooth contact surface in a preferred embodiment of the present invention.
FIG. 6 is a perspective view showing a lap step portion generated on a tooth surface according to a preferred embodiment of the present invention.
FIG. 7 is an explanatory diagram showing noise removal processing due to expansion / contraction in a preferred embodiment of the present invention.
FIG. 8 is an explanatory view showing the position of the center of gravity on the tooth contact surface on the advance surface side and the retreat surface side in the preferred embodiment of the present invention.
[Explanation of symbols]
3 image processing device 4 photographing device 5 lighting device 6 mounting table

Claims (5)

The tooth surface, which shows the tooth contact surface generated by meshing the gears coated with Kometan, is photographed with a photographing device, and each component of RGB of the photographed RGB color image is converted into a grayscale image by an image processing device. In addition to the conversion, the luminance of the G component and the B component are added, and the density of the grayscale image is inverted. The grayscale image obtained by adding the luminance of the inverted image and the luminance of the R component is calculated based on the histogram. A tooth contact extraction method characterized in that a portion other than the tooth contact surface is removed and a binarized image is obtained. The method according to claim 1, wherein an RGB color image of the same tooth surface photographed a plurality of times is averaged for each of R, G, and B components. 3. The method according to claim 1, wherein noise is removed from the binarized image. A photographing device for photographing a tooth surface showing a tooth contact surface generated by meshing a tooth surface coated with Komyotan, an R, G, B component of a photographed RGB color image being converted into a gray scale image, and The brightness of the grayscale image is inverted after adding the brightness of the component and the B component, and the grayscale image obtained by adding the brightness of the inverted image and the brightness of the R component is removed based on the histogram, excluding the tooth contact surface. A tooth contact extraction device, comprising: an image processing device for converting the image into a binary image. The tooth contact extraction device according to claim 4, further comprising an illuminating device that illuminates the tooth surface with white light without shadow.

Images