KR200292152Y1 - Dimension Inspection Unit of Injected E.V.A Midsole - Google Patents

Abstract

The present invention relates to a dimensional inspection system of a shoe midsole, and more particularly, it is possible to measure the injected EVA midsole through an image processing system, thereby minimizing the measurement error in the longitudinal direction of the midsole, reducing defect rate and productivity. To contribute to improvement,

Screw rod (2) for forward and reverse rotation in conjunction with a motor (step motor) (7) on the worktable (1), and coupled to the screw rod (2) moves in the longitudinal direction in accordance with the forward and reverse rotation of the screw rod (2) A moving table 3, a color CCD camera 4 mounted on the moving table 3 and connected to the computer 9, and an object (middle sole 10) to measure dimensions on one side of the work table 1 is placed. The paper is composed of a surface plate (5) to contribute to reducing the defective rate, increase productivity, quality improvement, and the like.

Description

Injection Inspection Unit of Injected E.V.A Midsole

The present invention relates to a dimensional inspection system of a shoe midsole, and more particularly, it is possible to measure the injected EVA midsole through an image processing system, thereby minimizing the measurement error in the longitudinal direction of the midsole, reducing defect rate and productivity. To contribute to the improvement.

Recently, midsole manufacturing technology using injection E.V.A [Injected E.V.A (Ethylene Vinyl Acetate copolymer)] has been developed, and the manufacturing method of the midsole is gradually introduced by introducing its manufacturing method.

However, a disadvantage in the manufacture of this part is that deformation occurs due to shrinkage after molding. Therefore, the inspection of injection preforms is an essential item in the manufacturing process of injection E.V.A midsole.

However, its inspection method relies on the whole inspection by sampling and visual inspection, and such inspection by the naked eye is not performed optimally due to many problems such as the fatigue of the inspector, the consistency and cost of the inspection.

Therefore, the development of automated inspection methods for this situation is very necessary.

For the automatic inspection of such injection molding (preform) from early at home and abroad interested in computer image recognition and image processing that can replace the human eye and began to study for its application.

However, these technologies were not practical because they required expensive and high-speed computers until the early 1990s. However, with the rapid development of computer hardware technology and the development of image processing technology, various products and images have been used at home and abroad. Research on automatic inspection of injection products has been actively conducted, and some have been put into practical use.

However, the 2D vision inspection machine, which has been developed and put into practical use, has a shape such as the recognition speed of the injected EVA midsole and the inspection speed for processing dimensions, the measurement limit and measurement error correction of the two-dimensional image, and the bending of the midsole. Due to the inconsistency of the automatic inspection system for full inspection has a lot of problems.

In order to remedy this problem and to automatically inspect the object in real time, it is urgently needed to establish a new method of inspection.

Therefore, in this design, we construct an image processing system for inspecting the injected EVA midsole, measure the length of the gauge block using it, examine the measurement error according to the length, and correct the midsole. In order to minimize the measurement error in the longitudinal direction of the midsole by providing an inspection method for the midsole having a somewhat inconsistent shape, and to improve productivity from this.

1 is an overall configuration showing an embodiment of the present invention

2 is a side view showing the configuration of the present invention

3 is a flow chart of the inventive image processing algorithm

Figure 4 is a schematic table of the interval length measurement method of the present invention

☞ Explanation of symbols used in the main part of the drawing ☜

1; work table

2; screw rod

3; moving table

4; camera

5; surface plate

7; motor

9; computer

10; midsole

The present invention is largely equipped with a camera that adjusts the distance to the measurement object (midsole, hereinafter midsole) and the image processing unit for inputting the midsole's dimensions (standard) as an image to the image processor and through the input image It can be divided into the image processing process that inspects and verifies dimensions.

Looking at the configuration of the present invention in more detail as follows.

1 is a perspective view schematically showing the configuration of one embodiment of the present invention, Figure 2 is a side view showing the configuration of the present invention, the image processing unit is linked to the motor (step motor) 7 on the work table (1) Screw rod (2) for forward and reverse rotation, a moving table (3) coupled to the screw rod (2) and moved in the longitudinal direction according to the forward and reverse rotation of the screw rod (2), and installed in the moving table (3) It consists of a color CCD camera 4 connected to (9), and a surface plate 5 on which an object (middle sole) 10 whose dimension is to be measured is placed on one side of the work table 1.

The movable table 3 is provided with a movable table 3 which is movable in the longitudinal direction on the worktable 1 so as to adjust the distance to the measurement object, and a color connected to the computer 9 on the movable table 3. CCD cameras (eg 1/2 inch, Toshiba) are combined.

As the camera 4 on the moving cable 3 moves, the image obtained through the correct focus of the midsole 10, which is a subject, is framed with 8-bit resolution of 640 × 480 resolution and RGB, respectively, mounted in the connected computer 9. Obtained using a lever (frame grabber).

In the present invention, in order to obtain the overall shape of the midsole 10, the distance between the center of the camera 4 and the midsole 10 is fixed at approximately 380 mm, and the lens of the camera 4 uses a lens having a focal length of 6 mm.

In the drawings, reference numeral 6 denotes a support, and 8 illustrates a coupling.

A flowchart of an image processing algorithm for extracting the shape of the midsole 10 from the image of the midsole 10 obtained from the image processing system and measuring and inspecting the length thereof is shown in FIG. 3.

An image of the midsole 10 is obtained from the CCD camera 4 for inspection of the injected EVA midsole, and smoothing is performed to remove relatively high frequency components such as noise and isolation point. The contours are then extracted using the first order differential method.

The end points of the left and right contours are searched for the boundary thus extracted, and the length of the midsole 10 is measured using the section length measurement.

Hereinafter, the image processing process of inspecting and confirming the size of the midsole through the image input from the camera 4 will be described.

1) Image smoothing and contour extraction

In general, the image obtained from the camera 4 has a lot of noise components.

Therefore, when the contour of the object is extracted, the noise component may be included. Therefore, before the contour extraction process, it is necessary to remove the noise component through smoothing.

The smoothing process generally uses a filter mask having a size of 3 × 3, 5 × 5, 7 × 7, etc. in order to process at high speed. Such a filter is called a lowpass filter. In the present invention, the following 3 × 3 filter mask was used for the smoothing treatment.

(Lowpass filter of 3 × 3 size)

In order to extract the contour from the image of the object from which the noise component is removed, the first- or second-order differential method of the image is generally used.

The first differential method emphasizes only the boundary part by the slope of the concentration difference of the object boundary as follows, and is mainly used to extract the contour, and the second differential method is used to distinguish the dark and light parts with respect to the boundary.

Edge detection by derivative operator

The first-order differential method used in the present invention obtains the magnitude of the gradient of pixel density at an arbitrary point of the image f (x, y), and the slope of f at an arbitrary point with respect to the function of the image f (x, y). A vector is represented as in Equation 1 below, and the magnitude of the slope may be expressed in Equation 2.

However, in order to perform image processing at high speed, Equation 1 may be expressed as Equation 3 when z is a density value of neighboring pixels at a position of (x, y) as shown in (a) below.

In addition, (b) and (c) below show the Prewitt operator derived as described above, and Equation 3 is expressed in the form of a 3 × 3 mask.

2). Measurement of length of standard specimen and correction of error

Since the length is expressed in the number of pixels from the image of the object, it is necessary to obtain the ratio of the length to one pixel of the image in order to obtain the value of the actual length. To this end, in the present design, the length ratio was obtained by using a 100 mm gauge block as a reference specimen from the gauge block set of Mitutoyo.

The 100mm gauge block used as the reference specimen was image-processed to detect the contour and measured in the x-axis direction. The x-axis length of the reference specimen was 192 pixels, and the length per pixel was 0.52083 mm. And it was found.

In addition, the lens used in the present invention is a precision lens for general CCTV with a focal length of 6mm, so that light refraction and distortion appear, measurement errors may occur depending on the size of the object.

Table 1 compares the errors by measuring the lengths of 50, 100, 150, and 175 mm gauge blocks to compensate for this. As the size of the specimen increases, the length measured through image processing appears to be smaller than the actual length. Able to know. This is considered to be due to the refraction and distortion of the light in the lens, and from this result, the measurement error correction equation as shown in Equation 4 was obtained to reduce the measurement error.

Table 2 shows that the measurement error is reduced to within ± 0.6 mm as a result of measuring the length by the equation (4). This is considered to be within an appropriate error range considering that the dimensional accuracy of the midsole is ± 2 to ± 3mm.

Comparison of real and measured length of gauge block before errcompensation Length of Gauge Block (mm) Measured length (mm) Error (mm) 50 53.6457 3.6457 100 99.9994 0.0006 150 147.3957 -2.6043 175 169.7916 -5.2084

L '= (L-7.009) / 0.9322

(L ': calibrated length, L: actual measured length)

Comparison of real and measured length of gauge block after error compensation Length of Gauge Block (mm) Measured length (mm) Error (mm) 50 50.0286 0.0286 100 99.7542 -0.2458 150 150.5971 0.5971 175 174.6219 -0.3781

3). Length measurement of midsole

Injected E.V.A midsole is a type of sponge product that may cause some bending deformation during shrinkage during cooling after removal from the mold. Such bending deformation is not treated as bad if it is unfolded during the visual inspection and falls within the error range, but when measuring the linear length of both ends by general image processing, the midsole 10 is bent or embossed with respect to the reference product. Since the length error is large, a lot of cases are judged to be defective. To this end, the present invention intends to use the section length measurement method for measuring the length of the midsole 10.

4 is a diagram illustrating a section length measurement method, in which equal parts are drawn by dividing the left and right ends at equal intervals at regular intervals. Here, the midpoint between the upper and lower points where the contour line of the equal line and the midsole 10 meet is obtained. In this manner, the midpoints between the contour lines are obtained for each equal line, and the lengths of the midpoints connecting the midpoints are calculated using Equations 5 and 6 to determine the length of the midsole 10 to be measured. . In the present design, the interval length was measured by dividing the two ends by 40 equal parts.

(l_i: length in one section, x_i, y_i, x_i + 1, y_i + 1: location of intermediate point, L_ {i n t}: total section length)

Table 3 shows the comparison of the midsole's measurement length for bending deformation when the linear length measurement method and the section length measurement method are used.In the case of the linear length measurement method, the general error limit of the midsole according to the degree of bending with respect to the reference product is ± 2. It can be seen that the measurement error appears to be larger than ± 3mm. However, when the interval length measurement method is used, the measurement error is very reduced and it can be seen that it is within the error limit.

Therefore, these results suggest that the section length measurement is a very effective method for measuring the length of the automated E.V.A midsole inspection by image processing, which can increase the reliability of defect identification.

Measured length of injected E.V.A midsole for bending strain Angle of deflection (。) Horizontal linear (mm) Error (mm) Interval length (mm) Error (mm) 15 284.6885 5.0282 292.8674 1.1941 0 289.7169 - 294.0615 - -15 291.3931 1.6764 294.5771 0.5156

As described above, the dimensional measurement of the injection EVA midsole 10 according to the present invention corrects the measurement error by measuring the length of the gauge block, and applies the mid-length measurement method to the midsole having a somewhat inconsistent shape. As a result of measuring (10), the following conclusions were obtained.

1) When the 100mm gauge block is image-processed and the contour is detected and the length in the x-axis direction is measured, the length in the x-axis direction is 192 pixels and the length per pixel is 0.52083 mm.

2) As the size of the gauge block increases, it can be seen that the length measured through image processing is smaller than the actual length, and the measurement error can be reduced to within ± 0.6 mm.

3) As a result of measuring the length of the midsole for bending deformation, it was found that the measurement error was very small and within the margin of error when the section length measurement method was used.

[references]

R. Jain. R. Kasturi and B. G. Schunck Machine Vision. McGraw-Hill 1992

2. R. C Gozalez. R. E. Woods, Disital Image Processing. Addison-Wesley. 1992

3. Lee Sang-ryong, Kim Chae-sang. "Development of axisymmetric measurement system using visual sensor." Korean Society of Precision Engineering, Vol. 14, No. 5, pp.34 ~ 41, 1997.

4. Jaeyeol Kim, Boseok Oh, Yoshin Shin. "Development of visual inspection system for measuring length of injection products using vision system." Korean Society of Precision Engineering, Vol. 14, No. 11 pp.126-134. 1997.

5. Kim Young Il. "Development of Automatic Thickness Measurement Device Using Digital Image Processing Technique." Korean Society of Precision Engineering, Vol. 12, No. 6, pp.72-79. 1995.

6. Chae-Yul Kim, Yong-Hun Cha, Sung-Woon Yoon, Young-Suk Kim, Sang-Hwa Jung, Shin-Shin, Song-Suk Song, Kyung-Ok Park. "Development of Dimensional Inspection Program for Axle Casing Nut Weldment Using Digital Image Processing." Korean Society of Machine Tool Engineers, Vol. 9, No. 5, pp.135-143. 2000.

7. Jang Young-hoon, Kwon Tae-jong, Han Chang-soo, Moon Young-sik. "A Study on Ball-Stud Inspection System Using Vision.", Journal of the Korean Society of Precision Engineering, Vol. 15, No. 12, pp7-13. 1998.

As discussed above, the present invention enables more efficient and accurate measurement and inspection in the dimensional inspection of the injected EVA midsole, thereby minimizing defect rate in production, increasing productivity, and improving quality. Benefits and effects can be planned.

Claims (1)

In constructing the injection E-V shoe sole inspection system, Screw rod (2) for forward and reverse rotation in conjunction with a motor (step motor) (7) on the worktable (1), and coupled to the screw rod (2) moves in the longitudinal direction in accordance with the forward and reverse rotation of the screw rod (2) A moving table 3, a color CCD camera 4 mounted on the moving table 3 and connected to the computer 9, and an object (middle sole 10) to measure dimensions on one side of the work table 1 is placed. Injection E. V sole midsole inspection system, characterized in that consisting of the ground plate (5).