JPS6176941A - Method and device for inspecting appearance failure of screw - Google Patents

Abstract

PURPOSE:To discriminate a defective part with high accuracy, and also to improve the work efficiency by detecting the side face development graphic form of a screw. CONSTITUTION:Reflected light which has been irradiated from an illuminating device 1 and has been reflected by a screw 2 is converted to an electric signal through a line camera 6. An image signal inputting circuit 7 is controlled by a CPU10, and a video signal V1 is fetched by operating the camera 6. The signal V1 is synchronized with a scanning signal and sampled, and thereafter, binary-coded in accordance with bright and dark levels. The circuit 7 has a floating binary-coding circuit for varying a threshold value for binary-coding the signal V1, in accordance with its input signal. Also, the circuit 7 detects the one-dimensional bright and dark graphic form of the signal V1, and thereafter, rotates the screw 2 by a prescribed angle by sending out a control signal V2 to a motor driving circuit 9, and detects the one-dimensional bright and dark graphic form in other side face. Such an operation is stored continuously in an image memory 8, the side face development graphic of the screw 2 is obtained, various deformations on this side face development graphic form are discriminated 40, and a defect of the screw can be detected.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method and apparatus for inspecting appearance defects such as missing threads.

[Prior Art] Conventionally, defects in the appearance of screws have been inspected by the operator's eyes. For this reason, all the screws had to be kept around the entire circumference without being overlooked, which required a lot of inspection time.

C points to be solved by the invention] In order to improve the inaccuracy of conventional human visual inspection and shorten the inspection time, the present invention detects the side development view of a screw, and detects the side development view of the screw. The purpose of this invention is to identify 10 missing portions with high accuracy based on the characteristics of the screw that appear in the figure.

The first invention is an invention of an inspection method, in which a screw is irradiated with light from a specific direction, and an image sensor placed on the side of the screw is used to inspect the screw. By detecting the reflected light from the screw, a one-dimensional light-dark figure along the axial direction of the screw is detected, and by the relative rotation of the screw and the image sensor, the one-dimensional shape of the side surface of the # marked screw is detected. This is an inspection method for detecting bright and dark figures over the entire circumference, thereby determining the side developed figure of the screw, and inspecting the screw for external appearance defects from the side developed figure.

The light is irradiated onto the screw so that the bright and dark shapes of the peaks and valleys of the screw can be clearly distinguished. is desirable. By irradiating in this way, only the tip of the thread of the screw can be brightened, and the missing part of the thread can be darkened by the shadow of the adjacent thread, so that the missing part can be accurately determined. However, the irradiation direction is not particularly limited, and irradiation may be performed from a direction perpendicular to the axis of the screw. In this case, the tips of the threads and the bottoms of the valleys generally become brighter, and the sloped parts of the threads become darker, but if there is a defect in the thread extension, the reflection angle changes and the brighter parts become darker, or Since dark areas become brighter, defective areas can be detected.

A television camera, line sensor, area sensor, etc. can be used to detect the one-dimensional bright and dark pattern along the axial direction of the screw. In the case of a sensor capable of detecting a two-dimensional figure, a signal detected by one scanning signal corresponding to an image in the axial direction may be extracted. To obtain a lateral developed figure, while detecting a one-dimensional bright and dark figure along the axial direction,
t change the relative positional relationship with the sensor. For example, rotating a screw. For convenience of signal processing, it is desirable that the side development figure of this screw be binarized in correspondence with the brightness level. Using the binarized side development figure, the missing part of the screw is extracted based on the features related to the shape of the screw. A method for extracting a defective part is, for example, a method in which a binarized side developed figure is attached to its screw line, moved by a predetermined amount, and exclusive OR is performed between the patterns before and after the movement. It is possible to extract the parts of the screws that are struck and have no periodicity as defective parts. Also, if the screw is installed at an angle to the rotation axis, the screw line that appears in the side development figure will be a curve, so
It is difficult to move the figure along the screw line. In this case, it is necessary to move the figure while correcting it following the deformation of the screw line. That is, Exp1: The amount of movement in the axial direction in the mouth figure and the movement perpendicular to the sleeve direction! Ilf? t and odor, the movement perpendicular to the axis is constant! ,: In the case, the amount of movement in the axial direction is a3 for each part of the figure and varies depending on the deformation of the screw line (straight).If such movement processing is performed, the regular screw line will necessarily be formed.Therefore, If an exclusive OR is performed between the development diagrams before and after the movement, the ?JS portion, that is, the missing 10 portions, will be extracted as defective portions.

Further, the present second invention 1 is an apparatus for realizing these methods, and the inventive concept is represented as a block diagram in FIG. That is, the device of the present invention includes an illumination device 1 that irradiates light onto a screw 2 from a specific direction, and detects a one-dimensional light-dark pattern along the axial direction of the screw 2 by detecting reflected light from the screw 2. A rotation control device that holds the screw 2 and controls the relative rotation between the screw 2 and the image sensor 60! 20, and a side surface that drives the rotation control device 20, detects the one-dimensional light-dark figure on the side surface of the screw 2 over the entire circumference of the screw, and generates a side developed figure 1q that is binarized according to the light and dark level. A screw driver comprising a developed figure detecting device 30, and a pattern determining device 40 that processes the divided side developed figure detected by the side developed figure detecting device 3o and determines whether the appearance of the screw 2 is good or bad. This is an appearance defect inspection device.

The configuration of the device of the present invention includes eight lights, a video sensor that detects a one-dimensional contrast pattern on the side of the screw, a rotation control device that holds the screw and controls its relative rotation with the image sensor, and This system consists of a side development view detection device that detects a side development figure, and a pattern judgment device that determines whether the external appearance of the screw is good or bad from the side development figure. It is desirable to irradiate from a direction slightly inclined to the direction of the axis of the screw.

Further, it is desirable that the pattern determination device uses the method described above.

[Examples] Hereinafter, the present invention will be described in detail based on specific examples.

FIG. 2 is a block diagram showing the configuration of an appearance defect inspection device according to a specific embodiment of the device of the present invention.

FIG. 3 is an explanatory diagram showing the positional relationship between the illumination device 1 and the screw 2° line camera 6. The illumination device 1 emits light from a direction at a certain angle (θ) with respect to the direction of the axis 13 of the screw 2. The line camera 6 is located on the same plane as the illumination device 1 and the screw 2, and is installed in a direction perpendicular to the screw 2. This illumination device used a filament lamp with little variation in the amount of light emitted. But 10K
It is also possible to use a light source in which fluctuations in the amount of emitted light do not affect the output signal of the image sensor 60, such as a high-frequency fluorescent lamp of I-Iz or higher. The screw 2 uses a bolt. The rotating head 20 includes a rotor 3 having a chuck for fixing the screw 2, a shaft 4 connecting the rotor 3 and the motor 5, a motor 5 for rotating the screw, and a motor drive circuit for driving the motor 5. It consists of 9. A line camera 6 is used as the video sensor 60. The line camera 6 uses light receiving elements (COD) arranged in a straight line to receive the reflected light from the screw 2 and convert it into an electrical signal. (The image signal input circuit 7 that inputs the image signal input circuit 7, the uniformity memory 8 that stores the output signal of the image signal input circuit 7, and the CP that controls both of them.
It consists of U 10 and RAM 131 that stores processing programs. ii! ii The image signal input circuit 7 is controlled by the CPU 10 and operates the line camera 6 to input the video signal v.
Take out 1. The video signal V1 is sampled in synchronization with the scanning signal, and then binarized in accordance with the brightness and darkness levels. The image signal input circuit 7 receives the video signal v1 at 2flri
It has an input circuit so as to vary the threshold field to be changed in accordance with the input amplitude. Further, after detecting the one-dimensional contrast figure 1, the image signal input circuit 7 sends the i, II, and III signals v2 to the motor drive circuit 9 to rotate the I set screw by a predetermined angle. After rotating by a predetermined angle, one-dimensional bright and dark figures on the other side are detected in the same manner as shown in FIG. C, the side development figure of the screw 2 is obtained. Further, the pattern rotation device 40 is a CPU
10 and the RAM that stores the determination processing program.
II 32.

FIG. 4 shows the image l5 (L) detected by the line camera 6.
i and the image signal input circuit 7 perform binarization processing L'[!
The figure shows the 2 (1n signals) to be detected and the 0I11 plane developed figure obtained by detecting these signals over the entire circumference of the screw.As shown in Fig. 4(a), the h1 radiation is The light flux 21 i; The curve 24 also represents the envelope of the video signal @V1 produced by the line camera 6.The video signal v1 is set appropriately according to the amplitude of the input signal. The resulting 21-value level 25 is used as the threshold to binarize the signal.As a result, a 21-value signal 26 is obtained as shown in FIG. 4(C), and a one-dimensional contrast figure is obtained from this signal. Next, the motor 5 is driven to rotate the motor 5 by minute angles (1.8 degrees in this example), and the signal 26 divided into two sides according to each rotation angle is detected.As a result, the signal 26 shown in FIG. ) As shown in Figure 2, the screw is expanded.
7 is IJ. In this figure, the thread part is the bright part 28, the valley is the two dark parts, and the bright part is '1' of the binary code.
”, dark areas are represented by ″○′′.

If there is a defect in the thread, a signal as shown in FIGS. 5(b) and (C) will be emitted.
, the distribution of the amount of reflected light 33 of the desired light beam 21 becomes as shown by the curve 34 in FIG. 5(b). Defect part 35
is hidden in the shadow of the adjacent screw thread and no light is irradiated, so the reflected light cannot be reflected.

Therefore, the video signal QV1 of that portion is at a dark level. When the video signal v1 is converted into 21 values at the binarization level 36, a signal as shown in FIG. 5(C) is obtained, and the bright level signal corresponding to the screw defect 35 is not used.

Further, as shown in FIG. 6(a), when there are dents 45 on the screw thread, the incident light beam 21 is reflected on the surface of the dents 45, but due to the deformation, the reflected light 43 is As shown in FIG. 6(b), the position of the bright portion is shifted from the normal position.

When this is binarized using the binarization level 46, it becomes as shown in FIG. 6(C), and the thread pitch 48 of the normal thread part and the thread pitch 49 of the defective part have different values.

If there is a defect in the screw as described above, its side development figure will be as shown in FIG. Of these, 53 [, 亥] represents a defective part where light is not reflected due to a missing screw thread, 54 represents a defect where a part of the screw thread is deformed due to a dent, and 55 represents a defect such as a foreign object on the screw ( = J W. By processing such a developed figure, it is possible to detect defects in the appearance of the screw. First, the pitch of the screw line 51 at each part is measured. Second, defects can be detected by moving and aligning the figures and detecting the portions that cannot be overlapped.

FIG. 8 is a flowchart showing the processing of the CPU 10 used in the apparatus of this embodiment. The CPU 10 starts execution from step 100 and initializes the eight scales. Next, in step 102, a signal v3 for driving the image input signal circuit 7 is output. The image input signal circuit A drives the line camera 6, inputs the video signal v1 as described above, binarizes it, and stores it in the image memory 8 as a 2111I one-dimensional bright on figure. Motor drive circuit 9
A control signal v2 is outputted to rotate the screw 2, and many one-dimensional bright and dark figures are inputted to the outline of the screw to obtain a side-developed figure. The CPU inputs the image input signal circuit 7 in step 104.
Detects the () symbol and determines whether the manual input of the developed diagram has been completed. If the input is complete, move on to the next step and judge the pattern (noun. First, Step 105 t- detects the thread line drawn by the thread represented as a bright part, and determines the pitch in each part.

As a result, a defective location in the pitch of the screw wire is detected. For example, in FIG. 7, where there is a notch 53, the pitch is twice as large as the normal pitch. Further, in the deformed defect 54, the pitch on one side is large and the pitch on the other side is small.Furthermore, in the defect 55 where dust or the like has adhered, the pitch is small.

Next, the defective portion is extracted using another method.

The following method can also detect defects where the thickness of the thread line changes. That is, in step 106, the inclination angle of the thread line is measured. The obtained developed figure is moved in the horizontal and vertical directions for a predetermined distance along the direction of the measured inclination angle. Next, an exclusive OR is calculated between the screen after movement and the screen before movement. The result 1q is the 9th figure.
As shown in the figure. In other words, since a normal part has the same pattern before and after the movement for the above-mentioned constant movement, that part becomes 'O''' due to the excretory logic m.However, the defective part Since there is no periodicity, the figures do not overlap due to movement.Therefore, the defective part is 62.63.6
4, it is clearly visible. With this method, only defective parts can be extracted. In addition, the thickness of the screw wire fluctuates by about 1 bit of the line camera due to quantization errors, etc., resulting in noise such as 65, but such noise and minute defects can be removed by filtering. can. By comparing the graphic area of the defective portion thus calculated by 1q with a predetermined reference value, it is possible to accurately determine whether the appearance is good or bad. The device of this embodiment has the groove structure and operation described above.

In the above embodiment, the screen is moved along the direction of the screw line to extract the defective part, and an exclusive OR is calculated. However, depending on the arrangement of screw 2 to rotor 3, screw 2 may be attached to rotor 3.
may be tilted from the axis of rotation. In such a case, the developed figure will be a curved shape with the screw line deformed into a curved shape as shown in FIG. 10(a). For this reason, the screen cannot be moved along the screw line direction, and the curvatures of the screens are different, so they do not overlap.

Therefore, these appear as detection errors of the defect 81.

To avoid this problem, the following method is required. As shown in FIG. 10(b), sample points 82 are extracted from the thread line at constant m+ intervals, and the thread line is approximated by a broken line 83 connecting the sample points 82.

Then, the approximate polygonal line overlaps the polygonal line of the corresponding area after movement.) [The amount of displacement of these three images is determined. That is, the first
The amount of movement in the X-axis direction in Figure 0 (b) is constant (13 for all addresses), but the amount of movement in the Y-axis direction is changed according to the x coordinate of the figure to be moved. If you do this, if you have J3 on an image that is inflected in a curved shape, you can superimpose regular parts on V.Also, if the curvature differs depending on the Y coordinate (L, 4, G and Y
If the movement in the Y direction is changed according to the coordinates, perfect matching can be obtained.

The light irradiation method is as shown in Figure 11, using light source 1 and screw 2.
A scattering plate 71 may be provided between the two.

If there is uneven illumination, there are surfaces other than the screw that strongly reflect toward the line camera, which is particularly effective if the image of the screw is affected.

[Effects of the Invention] The present invention provides a method and apparatus for detecting a side-view developed figure of a screw, and detecting a defective appearance of the screw by taking characteristics of the screw into consideration from the side-view developed figure. Since side development figures are processed, defective parts can be detected with the same accuracy in all surrounding areas.

This device automates the detection of appearance defects, which improves work efficiency compared to human visual inspection.
High detection accuracy.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the concept of the device of the present invention. FIG. 2 is a block diagram showing the configuration of an inspection device according to a specific embodiment of the present invention. Figure 3 is an explanatory diagram showing the irradiation angle of the light emitted from the lighting device and the positional relationship between the screw and the line sensor that detects the reflected light from it. Figure 4 is the outline of the screw and the image created by the reflected light. FIG. 2 is an explanatory diagram showing a side development figure consisting of a signal, a binary signal obtained by processing the signal, and a binary signal thereof. Figures 5 and 6 are explanatory diagrams showing the waveforms of the external image signal and binary signal of the screw when there is a defect, and Figure 7 is a side development view of the screw when there is a detected defect. FIG. Figure 8 shows lζ used in the same embodiment device.
FIG. 9, a flowchart showing the processing of the metering machine, is a defective product extraction figure in which defective parts are extracted by performing exclusive OR on the same packaging. FIG. 10 is an explanatory diagram illustrating a method of moving and overlapping another developed figure, and FIG. 11 is an explanatory diagram illustrating another irradiation method. Patent applicant Nippondenso Co., Ltd. Agent Patent attorney Hirodo Okawa Patent attorney Shudo Fujitani Patent attorney Akio Maruyama2] Figure 4 (a) Figure 4 (b) Figure 4 (C) Figure 5 2] (a ) (b) (C) Figure 6 shows the amount (a) ('b) (C) Figure 8 Figure 9 $10 Figure y (a) (b) -a Figure 11 shows the amount (a)

Claims (10)

[Claims] (1) A one-dimensional light-dark figure along the axial direction of the screw is created by irradiating light onto the screw from a specific direction and detecting the reflected light from the screw using a video sensor placed on the side of the screw. , and detecting the one-dimensional light-dark figure on the side surface of the screw over the entire circumference by the relative rotation between the screw and the video sensor, thereby obtaining a side development figure of the screw; An inspection method for inspecting external appearance defects of the screw from a side development figure. (2) The side development figure is binarized according to the brightness level, and the determination of the appearance defect of the screw is based on the periodicity of the pattern appearing in the binarized side development figure. Claim 1 characterized in that it is determined based on quality.
Method for inspecting external appearance defects of screws as described in section. (3) The side development figure is binarized according to the brightness level, and the judgment of the appearance defect of the screw is made by converting the binarized side development figure to the screw line appearing on it. Claim 1, characterized in that the process is carried out based on a figure in which a defective part is extracted by moving a predetermined amount along the same direction and performing an exclusive OR with the side development figure before the movement, and extracting the defective part. Method for inspecting the appearance defects of the described screws. (4) The side developed figure is binarized according to the brightness level, and the binarized side developed figure is moved by moving a unit amount in the axial direction of the screw and in the axial direction perpendicular thereto. To obtain the figure, the amount of movement in the axial direction is changed according to the inclination of each part of the screw line, and by taking the exclusive OR of the side development figures before and after the movement, the defective part is extracted. 2. The inspection method according to claim 1, wherein the screw is inspected for defects in appearance based on the figure. (5) The specific direction in which the light beam is irradiated is a direction slightly inclined with respect to the axis of the screw so that the reflection of light from the tip of the screw thread is increased. A method for inspecting external appearance defects of screws as described in Scope 1. (6) a lighting device that irradiates light onto the screw from a specific direction; an image sensor that detects a one-dimensional bright and dark pattern along the axial direction of the screw by detecting reflected light from the screw; and the screw. a rotation control device that holds the screw and controls relative rotation between the screw and the image sensor; and a rotation control device that drives the rotation control device to detect the one-dimensional brightness pattern on the side surface of the screw over the entire circumference of the screw. a side development figure detection device that obtains a binarized side development figure according to the brightness level; and a side development figure detection device that processes the binarized side development figure detected by the side development figure detection device to determine the appearance of the screw. A screw appearance defect inspection device comprising a pattern determination device for determining quality. (7) The pattern determining device determines whether the appearance of the screw is good or bad based on whether the periodicity of the pattern appearing in the binarized side development figure is good or bad. Inspection device for defective appearance. (8) The pattern determination device moves the binarized side developed figure by a predetermined amount along the screw line appearing thereon, and calculates an exclusive OR with the side developed figure before movement. 7. The apparatus for inspecting external appearance defects of screws according to claim 6, wherein a figure is obtained by extracting a defective part, and the quality of the appearance is determined based on the figure. (9) In moving the binarized side developed figure by a unit amount in the axial direction of the screw and in an axial direction perpendicular thereto to obtain a moving figure, the pattern determination device may move the binarized side developed figure by the amount of movement in the axial direction. is changed according to the inclination of each part of the screw line, and by taking the exclusive OR of the side development figures before and after movement, a figure that extracts the defective part is obtained, and based on this figure, the screw 7. The screw appearance defect inspection device according to claim 6, wherein the screw appearance defect inspection device determines appearance defects. (10) The specific direction in which the light beam is irradiated by the illumination device is a direction slightly inclined with respect to the axis of the screw so that the reflection of light from the tip of the thread of the screw is large. An apparatus for inspecting external appearance defects of screws according to claim 6.