JPH0542327Y2 - - Google Patents

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention relates to an inspection device that uses image processing to determine whether the assembled state of parts is good or bad.

(Conventional technology) Recently, with the development of image processing technology, labor saving and inspection have become possible, not only in the appearance inspection of individual electronic elements and various parts, but also in the inspection of scratches and uneven paint on the painted surface of automobiles. Automatic inspection equipment using visual systems has been introduced with the aim of improving reliability.

Similarly, in the automobile industry, automation of the snap spring assembly inspection process for serration fitting in constant velocity joints, for example, is being promoted, and this inspection is carried out as shown in FIG. 5a. The outline of this test will be explained below based on the same figure.

Step 1 First, an image of the serration fitting portion 2 after the snap ring 1 has been assembled as shown in FIG. 5B is inputted by a television camera installed in an inspection device (not shown).

Step 2 The image input in Step 1 is binarized using a predetermined threshold.

Step 3 Analyze the binarized image data obtained by the processing in Step 2.

Step 4: Extract the image of only the serration site 2 from the images captured by the television camera.

Step 5 The extracted image of the serration region 2 is binarized using the optimal threshold value, and the data after this processing is analyzed.

Step 6 Output the result of this analysis (OK, NG).

By performing such processing, it is determined whether or not the snap spring assembly work of serration fitting in the constant velocity joint has been carried out as specified.

(Problem to be solved by the invention) However, in the case of the conventional inspection equipment as described above, the image input from the television camera is binarized using a set threshold value, and the subsequent processing is For example, if there is a change in lighting conditions, non-uniformity of the material of the workpiece to be copied, or change in the position of the workpiece, masking processing is performed based on this image. There is a risk that the image necessary for specifying the processing area to be extracted may not be obtained, and the image of only the serration site 2 cannot be extracted, making subsequent processing, that is, judgment of the test results, difficult. Will it be?
Alternatively, a misunderstanding may occur in the judgment.

The present invention was developed in view of these conventional problems, and it is possible to always perform the masking process regardless of changes in lighting conditions, non-uniformity of the material of the work to be copied, changes in the position of the work, etc. The processing area to be extracted can be reliably extracted when
The purpose is to make it possible to obtain highly reliable test results.

(Means for Solving the Problems) In order to achieve the above object, the present invention includes an image input means for inputting an image of a workpiece having a circular outline at least in part, on which predetermined parts are assembled;
a grayscale image processing means for quantizing the image; an image storage means for storing the quantized image; and a center point calculation means for calculating the center point of the circular contour based on the quantized image. , an area storage means for storing a processing area in the image input by the image input means, and a processing target area setting for setting a subsequent image processing area based on the center point of the circular outline and the processing area. and a binary value that quantizes only the image within the area set by the processing target area setting means among the images stored in the image storage means and calculates a histogram pattern of density value - number of pixels. storage means for storing a density value-pixel count schitogram pattern when the assembled state of the parts of the workpiece is normal; and a density value calculated by the binarization processing means. - a first determination unit that compares the histogram pattern of the number of pixels and the density value stored in the storage unit and the histogram pattern of the number of pixels, and determines whether the assembled state of the component is good or bad; and the binarization process. Binarization for creating a binarized image by binarizing the image within the image processing area set by the processing target area setting unit based on the density value-pixel count histogram calculated by the unit. an image creation means, a pattern storage means for storing a reference pattern after binarization processing when the assembled state of the parts of the workpiece is normal, and an image creation means for storing an image created by the binarization image creation means. a second determining means for comparing the pattern with a reference pattern stored in the pattern storage means and determining whether the assembled state of the parts is good or bad; and outputting a result determined by the second determining means. It is characterized by comprising an output means.

(Operation) The operation of the present invention will be explained below based on FIG. 1.

The image of the workpiece taken in by the image input means 3 is converted into digital data according to its density value by the grayscale image processing means 4, and is stored in the image storage means 5.
is memorized.

The center point calculation means 6 then uses the image storage means 5
Based on the data stored in the image, the center point of the member having a circular outline included in the image is calculated, and the processing target area setting means 7 calculates the center point of the member having a circular outline included in the image, More area storage means 8
Set the image within a certain area stored in .

Next, the binarization processing means 9 extracts the image of the area set by the binarization processing means 9 from among the images stored in the image storage means 5 and performs the binarization processing. 1 determination means 10 compares the histogram pattern of number of images-density value during this binarization process with the histogram pattern of number of pixels-density value when the workpiece is normal, which is stored in storage means 11. Determine the quality of the assembly condition etc. on the workpiece.

Furthermore, the binarized image creation means 12 binarizes the extracted image based on the threshold value calculated by the binarization processing means 9 to create a binarized image.
This binarized image is determined by the second determining means 13 as follows:
The pattern of the extracted portion and the reference pattern stored in the pattern storage means 14 are compared, and the determination result is outputted to the output means 15.

Therefore, highly reliable inspection work can be performed.

(Example) Below, an example of the present invention will be described in detail based on the drawings.

FIG. 2 shows a schematic configuration diagram of an inspection device for assembling a serration fitting snap spring in a constant velocity joint according to the present invention. As shown in the figure, this inspection device is composed of an image input section and an image processing section, and the image input section is
It consists of an ITV camera 3 as an image input means for inputting an image of the surrounding area of the snap spring, and an A/D converter 16 for converting an analog signal output from the ITV camera 3 into a digital signal.

The image processing section includes an input interface 17 for inputting the digital signal output from the A/D converter 16, an image memory 5 as an image storage means, and a storage means for storing the data output from the input interface 17. and output to RAM 19 as a pattern storage means.
It consists of a CPU 20 that processes the data based on a program stored in the ROM 18, and an output interface 15 that serves as an output means for outputting the processing results to an external device.

The inspection device having such a configuration has a third
It operates as follows based on the operation flowchart shown in the figure. This flowchart will be explained with reference to FIG.

Step 10 First, the ITV camera 3 sends an image around the serration-fitting snap ring to the A/D converter 1.
6 and the CPU via input interface 17
20 and temporarily stored in the image memory 5.

Step 11 The CPU 11 subjects the image stored in the image memory 5 to grayscale image processing based on the program stored in the ROM 18. (Data corresponding to the density value of the pixel stored in the image memory 5 is created.) Step 12 The image processed in step 11 is stored in the image memory 5 again.

For example, the image in this case is as shown in FIG. 4A, when the snap spring 1 is assembled normally.
In the case where the snap spring is not installed, in the case of a so-called single ride where only one side is attached to the groove where the snap spring 1 is attached, in the case of a so-called double ride where both sides are not attached to the groove where the snap spring 1 is attached, furthermore, the snap spring 1 is attached. is 2
Each case differs depending on the so-called double assembly in which two parts are installed.

Step 13 The CPU 20 retrieves the image after the grayscale image processing from the image memory 5, and extracts the outline of the image as shown in FIG. 4B, for example, by applying first-order differential processing, Laplacian, etc. to this image.

Step 14 The CPU 20 processes the image obtained in step 13 and calculates the center point of the work including the snap ring 1 in the image indicated by the + mark in FIG. 4B.

Step 15 The CPU 20 sets a window for masking processing, that is, a processing area, using the coordinates of the center point of the workpiece obtained in Step 14 and the data regarding the area stored in the RAM 19.
Set as shown in Figure C.

Step 16 The CPU 20 inputs the data in the obtained processing area from the image memory 5, performs gradation image processing on only this area, and as shown in FIG. Create a histogram pattern of pixel number-density value.

Step 17 The CPU 20 compares the histogram pattern of pixel number-density value at the time of normal assembly of the snap ring 1 stored in the RAM 19 with the histogram pattern of pixel number-density value at the time of assembly obtained in step 16. . As a result of this comparison, if the histogram patterns match, the process proceeds to step 18; if not, the process proceeds to step 23.

Step 18 The CPU 20 calculates a threshold value for optimal binarization processing calculated by floating binarization etc. from the histogram obtained in Step 16, and uses this threshold value to determine the processing area input from the image memory 5. The image is binarized to obtain an image as shown in FIG. 4E.

Step 19 The CPU 20 calculates the area and position of the image existing in the obtained binarized image.

Step 20 The CPU 20 inputs the binarized pattern stored in the RAM 19 when the snap ring 1 is normally assembled.

Step 21 The CPU 20 generates a pattern based on the pattern of the normal assembly of the snap ring 1 input in Step 20 and the pattern calculated in Step 19 of the image shown in FIG. If the patterns match, the process goes to step 22, and if they do not match, the process goes to step 23.

Step 22 When the CPU 20 confirms that the pattern for normal assembly of the snap ring 1 matches the pattern for the snap spring assembly determined in step 18, it sends an OK signal to an external output device (not shown) via the output interface 15. Output.

Step 23 If the pattern when the snap ring 1 is normally assembled does not match the pattern when the snap spring is assembled that was determined in step 18, the CPU 20 sends an NG message to an external output device (not shown) via the output interface 15. Output a signal.

As described above, in this embodiment, first, the image of the captured workpiece is subjected to grayscale image processing to extract the outline of the workpiece and calculate the center point of the workpiece. Then, an image in a preset range of this center point is extracted, and the extracted image is subjected to gradation image processing to create a preset histogram pattern of pixel number-density value during normal assembly. The histogram pattern of pixel number-density value at the time of actual assembly is compared. next,
The extracted image is binarized and a preset pattern for normal assembly is compared with a pattern for actual assembly. Then, the preset histogram pattern during normal assembly and the histogram pattern during actual assembly match, and the preset pattern after the binarization process during normal assembly and When the patterns after the binarization processing during actual assembly match, the inspection result becomes OK.

As mentioned above, in the embodiment of the present invention, the case of inspecting the fitting state of the snap spring 1 was described as a specific example, but the present invention is not limited to this and can be applied to inspection devices for various parts. Of course it is.

(Effects of the invention) As is clear from the above explanation, according to the invention, only the area necessary for inspection of the workpiece is reliably extracted, and the pixel-density value histogram of the extracted area and the Since the inspection results are calculated by comparing the image patterns, there is no misjudgment of the inspection results due to the quality of the lighting conditions of the workpiece, non-uniformity of the material, positional fluctuations, etc., making it more reliable. It is possible to obtain extremely accurate test results.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a parts assembly state inspection apparatus according to the present invention, FIG. 2 is a schematic configuration diagram of a parts assembly state inspection apparatus according to the present invention, and FIG.
FIG. 4 is an operation flowchart of the parts assembly state inspection device according to the present invention, and FIGS.
1 is an operation flowchart of a conventional parts assembly state inspection device and a diagram for explaining the same. 1...Snat spring (work), 2...Serration fitting portion, 3...ITV camera (image input means), 5...Image memory (image storage means), 15
...Output interface (output means), 19...
...RAM (area storage means, storage means), 20...
CPU.

Claims (1)

[Scope of Claim for Utility Model Registration] Image input means for inputting an image of a workpiece having predetermined parts assembled and at least a portion of which has a circular outline; grayscale image processing means for quantizing the image; an image storage means for storing the image; a center point calculation means for calculating the center point of the circular contour based on the quantized image; and a processing area in the image input by the image input means. an area storage means for storing an image, a processing target area setting means for setting a subsequent image processing area based on the center point of the circular contour and the processing area; , a binarization processing means that quantizes only the image within the area set by the processing target area setting means and calculates a histogram pattern of density value per pixel; A storage means for storing a histogram pattern of density value-number of pixels in a normal case, and a histogram pattern of density value-number of pixels calculated by the binarization processing means and density values stored in the storage means. - a first determination means that compares a histogram pattern of the number of pixels and determines whether the assembled state of the component is good or bad; and a density value calculated by the binarization processing means based on the histogram of the number of pixels. , a binarized image creating means for binarizing an image within the image processing area set by the processing target area setting means and creating a binarized image; a pattern storage means for storing a reference pattern after binarization processing in a case, and a pattern of an image created by the binarized image creation means and a reference pattern stored in the pattern storage means are compared; , a second determining means for determining whether the assembled state of the parts is good or not; and an output means for outputting the result determined by the second determining means. Device.