JPH08254503A - Apparatus and method for automatic fabric inspection - Google Patents

Abstract

PURPOSE: To prevent an incorrect judgment without generating an abnormal waveform as long as a woven cloth is normal even when an illuminating environment is changed by a method wherein an image signal in which a waveform component inherent in a normal woven cloth has been removed from an image signal obtained as an imaged result is compared with a tolerance. CONSTITUTION: A CCD camera 4A detects a flaw and a small dotlike flaw in the weft direction of a woven cloth 1, and the whole width of the woven cloth 1 is covered by three cameras. A CCD camera 4B is used to detect a large flaw such as a stain, a spot or the like, and the whole width of the woven cloth 1 is covered by two cameras. A CCD camera 4c is used to detect a flaw in the warp direction, and the whole width of the woven cloth 1 is covered by three cameras. Then, in a signal processing system 13, an image signal in which a waveform component inherent in a normal wove cloth has been removed from an image signal obtained as an imaged result is compared with a threshold value by a digital comparator so as to judge an abnormality, and the abnormality is taken into an image memory so as to be output to the outside.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic inspecting apparatus and an automatic inspecting method for detecting an image of a woven cloth by an image pickup apparatus and automatically detecting a defect in the woven cloth based on the image pickup result.

[0002]

2. Description of the Related Art There has been proposed an inspection device (Japanese Patent Laid-Open No. 62-93637) for taking an image of a woven fabric using an image pickup device such as a video camera and inspecting the result of the image pickup. Conventionally, if there is no defect in the woven fabric with this type of automatic inspection device,
Paying attention to the fact that the magnitude (level) of the image signal obtained as a result of imaging falls within a predetermined level, the magnitude of the image signal is compared with a threshold value (allowable range), and the magnitude of the image signal is greater than the threshold value. When it is smaller than the threshold value, it is determined to be normal.

[0003]

Since the magnitude of the image signal is proportional to the intensity of the illumination, the magnitude of the image signal also changes depending on the installation environment of the video camera. Therefore, the threshold value must be adjusted and changed for each installation location of the image pickup apparatus.

Since woven fabrics of thousands of different colors are to be inspected, if the woven fabric to be inspected has a different color, the threshold value must be changed. If you don't change the threshold,
A normal woven fabric may be erroneously determined to be abnormal, or an abnormal woven fabric may be erroneously determined to be normal. There are many types of defects of woven cloth, such as different sizes, different strengths, unevenness, and uneven colors, and the directions in which they occur also differ. It is desired to improve the detection accuracy.

[0005]

In order to achieve such an object, the invention of claim 1 images the woven cloth to be inspected by an image pickup device, and the level of an image signal obtained as an image pickup result is within an allowable range. And a signal processing circuit that removes the waveform component of the normal woven fabric from the image signal obtained as a result of the imaging in the automatic inspection device that determines that the woven fabric is abnormal when And a comparison circuit for comparing the generated image signal with an allowable range.

The invention of claim 2 is characterized in that, in addition to the invention of claim 1, a plurality of the image pickup devices are installed corresponding to the types of defects.

According to a third aspect of the present invention, in addition to the second aspect of the invention, the plurality of image pickup devices have different installation positions, visual field sizes, and illumination positions according to the types of defects to be detected. Characterize.

According to a fourth aspect of the invention, in addition to the first aspect of the invention, a signal system for outputting the image signal obtained as the image pickup result to the signal processing circuit and the image signal directly for outputting to the comparison circuit. It is characterized by further comprising a signal system and a switching circuit for selectively switching the two signal systems.

According to a fifth aspect of the present invention, in addition to the first aspect, the woven fabric is a glass cloth, and the first image pickup apparatus for obtaining a transmission image of the glass cloth as the image pickup apparatus and the glass cloth. It has a 2nd imaging device which acquires the reflection image of.

According to a sixth aspect of the present invention, the woven cloth to be inspected is picked up by an image pickup device, and when the level of the image signal obtained as an image pickup result exceeds an allowable range, the woven cloth is determined to be abnormal and automatic detection is performed. In the method, the waveform component of the normal woven fabric is removed from the image signal obtained as the image pickup result, and the image signal from which the waveform component is removed is compared with an allowable range.

According to a seventh aspect of the invention, in addition to the sixth aspect of the invention, a filter circuit is used to remove the waveform component, and the content of the filtering process is variably set according to the type of defect to be detected. Is characterized by.

According to the invention of claim 8, in addition to the invention of claim 6, a filter circuit is used to remove the waveform component, a normal woven fabric is imaged prior to automatic inspection, and It is characterized in that the filtering process content of the filter circuit is determined using the resulting image signal.

According to a ninth aspect of the invention, in addition to the sixth aspect of the invention, a filter circuit is used to remove the waveform component so that the level of the image signal input to the filter circuit falls within a predetermined range. In addition, the filtering process contents of the filter circuit are determined.

[0014]

According to the present invention, the characteristic component of the normal waveform is removed from the image signal, and only the abnormal waveform portion caused by the defect is inspected. Therefore, the normal portion of the woven cloth is changed due to the change of the lighting environment. Even if the level of the signal waveform itself changes, an abnormal waveform does not occur as long as the woven cloth is normal, and the erroneous determination as in the conventional case is eliminated.

According to the second aspect of the present invention, a plurality of defects are detected to improve the detection accuracy.

According to the third aspect of the invention, the inspection accuracy is improved by defining the imaging environment according to the type of defect. In the invention of claim 4, the signal system for inspecting the signal waveform of only the defective portion and the conventional signal waveform in which the signal waveforms of the normal portion and the defective portion are mixed can be selectively used according to the type of the defect to be inspected. it can. Therefore, the inspection device can be used universally without being limited to the type of woven fabric.

According to the invention of claim 5, it is possible to improve the detection accuracy of the glass cloth.

According to the invention of claim 7, in addition to the effect of claim 6, it is possible to detect a plurality of defects with one inspection device, and the inspection device can be used for general purposes.

In the eighth aspect of the invention, since the filter circuit is provided with the function of automatically adjusting the processing content, manual adjustment can be omitted.

According to the ninth aspect of the present invention, it is possible to deal with a change in illumination by providing an automatic level adjusting function of the image signal, and the inspection accuracy is improved.

[0021]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(First Embodiment) FIG. 1 schematically shows a side surface of an internal structure of an automatic inspector to which the present invention is applied. FIG. 2 is a top view thereof. In FIG. 1 and FIG. 2, the woven cloth 1 to be inspected is supplied by the guide roller 2 to the inspection surface where automatic inspection is carried out, receives the inspection, and goes through the guide roller 3 to the next step. Three line type CCD cameras 4A and two line type CCD cameras 4B are arranged on the upper part of the inspection surface so as to be inclined at a constant angle with respect to the conveying direction of the woven fabric 1. The line CCD cameras 4A and 4B receive the reflected light from the illumination 5A. Further, three line type CCD cameras 4C are arranged at right angles to the conveyance direction of the woven fabric 1,
The reflected light from the two lights 5B and 5C is received.

As the illuminations 5A, 5B and 5C, a long fluorescent lamp, an optical fiber type halogen light source or the like is used. As each line CCD camera, one having a function of changing the field of view of the camera such as zoom is used.

The line-type CCD camera 4A is used for detecting defects in the weft direction of the woven cloth 1 and small dot-like defects. As the line CCD camera 4A, one having 2048 pixels is used and three units cover the entire width of the woven fabric 1. Also, 1
The exposure time per line is set short in order to detect a weft direction defect. In addition, line type CCD camera 4B
Is used to detect large defects such as stains and stains, and a line type CCD camera with 2048 pixels is used to cover the entire width of the woven fabric 1 with two units. Further, the exposure time per line is set long in order to detect a large defect. The line-type CCD camera 4C is used to detect defects in the warp direction, and three line-type CCD cameras having 4096 pixels are used to cover the entire width of the woven fabric 1. For this reason, the resolution of the line CCD camera 4C is increased by assigning 2 pixels to one warp. Further, the exposure time per line is set long in order to detect a defect that is long in the longitudinal direction.

In addition to the above, a seam detector 6 for detecting a seam between cloths and an encoder 7 for measuring the length of the cloth are arranged on the inspection surface. A video signal 8 output from a camera arranged in the center of the three line CCD cameras 4A and 4C is branched and connected to an illuminance controller 9. The illuminance controller 9 receives the video signals 8 output from the line CCD cameras 4A and 4C (middle), and lights 5A, 5B, 5 so that the average level of the video signals 8 becomes a preset level.
Feedback control for raising or lowering the brightness of C is performed. The video signal 8 output from the line type CCD cameras 4A, 4B and 4C is the control panel 1 which is the central part of the automatic inspection machine.
It is connected to a signal processing system 13 for detecting defects in 0.

The output 11 of the seam detector 6 is the control panel 1
0 is connected to a digital interface (I / O) 14 and the output 12 of the encoder 7 is connected to a counter 15 in the control panel 10. The counter 15 starts operating in response to the output of the seam detector 6, and the length of the woven cloth 1 is measured by counting the output of the encoder 7. Further, the output of the counter 15 when a defect is found in the woven cloth 1 is converted into a defect position.

FIG. 3 shows the configuration of a signal processing system relating to the detection of defects. This processing system comprises eight digital filter boards 16, two image processing boards 17, and one central control board 18. The filter substrate 16 is a line type CC
One D camera is provided for each of the D cameras 4A, 4B, and 4C. The digital filter substrate 16 is a line type C
A still image is extracted from the moving image captured by the CD camera,
Burst transfer to the image processing board 17. In the digital filter substrate 16 equipped with a digital filter circuit, the main component (feature) of a predetermined signal waveform of a normal woven cloth having no defect is removed from the captured result (still image) by the filter circuit. Further, in order to keep the remaining waveform components within the preset measurement range, those with a level higher than the measurement range are attenuated, and conversely, those with a level lower than the measurement range are amplified. As described above, in this embodiment, even if the cloth color changes and the output level of the CCD camera changes, the abnormal level can be detected more accurately by amplifying or attenuating the output level so that the level falls within the measurement range. .

In the image processing board 17, the received still image is image-analyzed, and the presence or absence of a defect in the woven fabric, and if there is a defect, the content of the defect, for example, brightness gradation, area, perimeter length, line length, width, The long side angle and the like are discriminated, and the characteristic parameter indicating the type and the degree of the defect is calculated. When the central control board 18 receives the characteristic parameters from the image processing board 17,
This characteristic parameter is converted into a code signal for display or the like. This code signal is output as a detection result to an external device.

The detailed signal processing contents of the digital filter substrate 16 will be described below with reference to FIG. FIG. 4 shows a circuit configuration of the digital filter substrate 16.

The output V for one line input from the line type CCD cameras 4A to 4C is the shading correction unit 19.
Then, shading (unevenness) is corrected by and an output E is obtained. The shading correction pattern to be sent to the shading is created by the central control board 18. The output E is converted into a digital signal by the A / D converter 20 for each pixel. A /
The output D from the D conversion unit 20 is the digital filter unit 2
1 is smoothed by digital FIR filtering. In the present embodiment, the processing content of the digital filter is given by the filter coefficient C, and for example, the coefficient C is determined by a smoothing method called the moving average difference method. When the number of filter taps is 8, the coefficients c1 to c4 are set to a and the coefficients c5 to a8 are set to -a.

In this embodiment, an image is input for a woven cloth having a normal appearance, and the coefficient a is adjusted in advance. By this adjustment, the main component of the waveform component of the image signal is erased in the digital filter unit 21, and the image signal is amplified / attenuated so that the remaining waveform component falls within a predetermined measurement range. A method of adjusting the coefficient a in consideration of the case where the color of the woven cloth 1 is different will be described below. That is, the woven fabric 1 is actually measured, and when the seam detector 6 operates, the output F of the digital filter unit 21 is forcibly changed by switching the switch 23.
In the image memory 24. The central control board 18 calculates a representative value of the brightness value distribution, for example, a standard deviation from the brightness value indicated by the output F. The coefficient a is set so that this standard deviation value is equal to the standard deviation value. When a0 is used as a coefficient when used for sampling, σ0 is a standard deviation value of a target (reference), and σ is a standard deviation value of a sampled image, a is calculated by (σ0 × a0) / σ.

Since a plurality of image pickup systems are provided in this embodiment, the above-mentioned filter coefficient suitable for defect detection, that is, the processing content of the present invention is set for each type of defect to be detected. For example, when erasing the warp pitch in the image, the number of filter taps is set small. The upper part of FIG. 5 shows the image signal waveform before the filtering process, and the lower part shows the waveform after the filtering process. The example of FIG. 5 is an example in which the woven cloth 1 has a defect, and it is possible to easily visually determine that the defective portion cannot be identified by the waveform before the filter processing, but the characteristic of the defective portion appears in the waveform after the filter processing. .

In this way, the setting of the filter is completed in advance, and the signal system is switched by the changeover switch 23 so that the output D of the A / D converter can be taken into the image memory 24. The output F output at the time of the inspection is compared with a threshold value in the digital comparator 22, and when the abnormality determination is obtained, the trigger signal T is generated and the image of the woven fabric 1 (normal component Is output to the outside. This image is saved as defect data or displayed on a monitor or the like.

By sending the output F after the filter processing to the image memory 24 by the changeover switch 23, it is possible to judge the presence or absence of a defect and select an image in which the feature amount of the defect is easily calculated.

As described above, in the first embodiment, since the main component of the frequency component of the normal portion is deleted from the waveform of the acquired image signal, the image signal of the image signal changes due to the change of illumination and the change of cloth color. Even if the waveform level itself changes, only the abnormal part remains. As a result, the inspection can be performed without being affected by the lighting environment.

A flow chart for executing the control of the present invention is shown in FIG. S1 to S16 in the figure indicate the steps of the flowchart.

The operation of the present invention is performed by, for example, the seam detector 6
It is executed when a signal is received from. First, step S2
Then, the signal processing conditions such as the shading correction pattern P, the filter coefficient C, the target standard deviation σ0, and the detection threshold value are read and set. In step S3, the image processing conditions are read and set. In step S4, a signal from the seam detector 6 is awaited, and if the signal is ON, the process proceeds to the next step S5. In step S5, unnecessary parts near the seam of the cloth are passed. The image is captured in step S6, the standard deviation of the image brightness values is calculated in step S7,
Based on the result, in step S8 a new filter coefficient C
Calculate the size of ´. In step S9, the filter coefficient C'is set.

In step S10, it is confirmed whether or not the next joint signal is ON, and if it is OFF, the procedure goes to the inspection execution section. O
If it is N, the next inspection cloth is obtained, so that the old filter coefficient C is set again in step S11, and the process returns to step S5. In step S12, the external inspection end signal is turned on,
Alternatively, when the inspection of the predetermined quantity set in advance is completed, the inspection is completed. If the end signal is OFF, it is confirmed in step S13 whether the trigger signal T is ON. If it is OFF, the process returns to step S10, and if it is ON, image processing is performed in S14. In step S15, it is determined whether or not it is a defect based on the result of the image processing. If it is a defect, in step S16, defect information such as the position where the defect has occurred, the type of defect, and the degree of defect is recorded. If it is not defective in step S15, the process returns to step S10.

Steps S4 to S9 are the preparation section for the inspection, and steps S10 to S15 are the inspection execution section. Steps S6 to S9 are steps for increasing the accuracy of the inspection of the present invention.

When the cloth color changes, that is, when the seam is turned on, a new shading correction pattern may be created based on the image captured in step S6.
Furthermore, in order to improve accuracy, it is also possible to take in a plurality of images, obtain an average value based on the result of step S7 for each image, and use it.

In addition to the first embodiment described above, the following example can be implemented.

1) In this embodiment, the image obtained by the line CCD camera is filtered, but it may be better not to perform the filtering depending on the kind of the defect. In this case, the output D of the A / D converter 20 is directly changed to the digital comparator 22 via the changeover switch 23 by the changeover switch 23 of FIG.
Bypass transfer is recommended. This makes it possible to provide a general-purpose inspection device that detects a plurality of types of defects.

(Second Embodiment) A test device suitable for the case where the woven cloth is glass cloth will be described below as a second embodiment. The glass cloth usually has a white color, and as a result of an experiment conducted by the inventor of the present application, the sensitivity of warp defects of the woven fabric (defects continuously continuing in the cloth advancing direction) is high in transmission imaging, and weft defects (cross width direction). It was found that the sensitivity of (defect of) is weak. It is considered that this is because the warp defect is often the case where the single yarn constituting the yarn is cut and becomes thin, which makes it unlikely to be a raised defect. Therefore, in the present embodiment, the image pickup system is configured to acquire a transmission image of the woven cloth for warp defects and acquire a reflection image for other defects. Further, since the captured image is easily affected by the illumination, feedback control is performed so that the illuminance becomes constant. Arrangement examples of the line CCD camera for this purpose are shown in FIGS. Figure 7
Indicates an arrangement for acquiring a reflection image. The light emitted from the light source 103 is reflected by the glass cloth 104,
It is taken into the line CCD camera 100. As shown in the arrangement (A) of FIG. 7, the light source 103 and the line type CCD camera 100 are arranged so as to have a certain angle with respect to the woven cloth 104. Further, in order to enhance the detection sensitivity of the weft direction defect, the glass cloth 104 as shown in the arrangement (B) of FIG.
Also, the scanning direction is arranged so as to have a constant angle with respect to the traveling direction. In this embodiment, in order to detect a plurality of types of defects, as shown in the arrangement (C) of FIG. 7, a plurality of line type CCD cameras 101 are prepared as in the first embodiment.

FIG. 8 shows an arrangement example for acquiring a transparent image. Components similar to those in FIG. 7 are designated by the same reference numerals and need not be described in detail. In the case of acquiring a transmission image, a better inspection result was obtained when the scanning direction was 90 degrees with respect to the traveling direction of the glass cloth.
FIG. 9 shows an example in which both the light source 111 for capturing the transmission image, the line CCD camera 110, the light source 113 for capturing the reflection image, and the line CCD camera 112 are arranged. A circuit for performing detection using an image signal obtained from such an image pickup system can be the same as that of the first embodiment shown in FIG. However, in the second embodiment, FIG.
A controller 122 for controlling the light source as shown in FIG.
This controller 122 is provided with a light source 12
The light sensor 121 provided on the zero side measures the light amount of the light source 120, and controls the power supply to the light source 120 so that the light amount of the light source 120 becomes constant according to the measurement result. As a result, particularly high quality images can be obtained in a stable illumination environment in an image pickup system for acquiring a transmission image.

[0045]

As described above, according to the first and sixth aspects of the present invention, the characteristic component of the normal waveform is removed from the image signal, and only the abnormal waveform portion caused by the defect is inspected. Even if the level of the signal waveform of the normal portion of the woven fabric changes due to the change of, the abnormal waveform does not occur as long as the woven fabric is normal, and the erroneous determination as in the prior art is eliminated.

According to the second aspect of the present invention, a plurality of defects are detected to improve the detection accuracy.

According to the third aspect of the invention, the inspection accuracy is improved by defining the imaging environment according to the type of defect.

According to the fourth aspect of the present invention, the signal system for inspecting the signal waveform of only the defective portion and the signal waveform in which the signal waveforms of the conventional normal portion and the defective portion are mixed according to the type of the defect to be inspected. Can be used properly. Therefore, the inspection device can be used universally without being limited to the type of woven fabric.

According to the invention of claim 5, it is possible to improve the detection accuracy of the glass cloth.

According to the invention of claim 7, in addition to the effect of claim 6, it is possible to detect a plurality of defects with one inspection device, and the inspection device can be used for general purposes.

In the eighth aspect of the invention, since the filter circuit is provided with the function of automatically adjusting the processing contents, manual adjustment can be omitted.

According to the ninth aspect of the invention, it is possible to deal with a change in illumination by providing an automatic level adjusting function of the image signal, and the inspection accuracy is improved.

According to the above invention, it is possible to provide an automatic inspection device that is resistant to changes in the illumination environment and contribute to the improvement of inspection accuracy.

[Brief description of drawings]

FIG. 1 shows a system configuration of a first embodiment of the present invention.

FIG. 2 is a plan view showing the arrangement of a line CCD camera according to the first embodiment of the present invention.

FIG. 3 is a block diagram showing a circuit configuration of a defect detection unit of the first embodiment of the present invention.

FIG. 4 is a block diagram showing a circuit configuration of a digital filter substrate.

FIG. 5 is a waveform diagram showing a measurement signal waveform according to the first embodiment of the present invention.

FIG. 6 is a flowchart showing an operation procedure of the embodiment of the present invention.

FIG. 7 is a configuration diagram showing an arrangement of a line CCD camera according to a second embodiment of the present invention.

FIG. 8 is a configuration diagram showing an arrangement of a line CCD camera according to a second embodiment of the present invention.

FIG. 9 is a configuration diagram showing an arrangement of a line CCD camera according to a second embodiment of the present invention.

FIG. 10 is a block diagram showing a configuration of an illumination control system according to a second embodiment of the present invention.

[Explanation of symbols]

 1 Woven Fabric 2, 3 Guide Roller 4A-4C Line Type CCD Camera 5A-5C Illumination 6 Seam Detector 9 Illuminance Controller 10 Control Panel 15 Counter 16 Digital Filter Board 17 Image Processing Board 18 Central Control Board

Claims (9)

[Claims] 1. An automatic inspection device, wherein an image of a woven cloth to be inspected is picked up by an image pickup device, and the woven cloth is determined to be abnormal when the level of an image signal obtained as an image pickup result exceeds an allowable range. It is characterized by comprising a signal processing circuit for removing the waveform component of the normal woven cloth from the resulting image signal and a comparison circuit for comparing the image signal from which the waveform component is removed with an allowable range. Automatic inspection device. 2. The automatic inspection device according to claim 1, wherein a plurality of the image pickup devices are installed corresponding to types of defects. 3. The automatic inspection device according to claim 2, wherein the plurality of image pickup devices have different installation positions, field of view sizes, and illumination positions according to the types of defects to be detected. . 4. A switching system for selectively switching between a signal system for outputting an image signal obtained as the image pickup result to the signal processing circuit, a signal system for directly outputting the image signal to the comparison circuit, and the two signal systems. The automatic inspection device according to claim 1, further comprising a circuit. 5. The woven cloth is glass cloth, and has a first imaging device that acquires a transmission image of the glass cloth and a second imaging device that acquires a reflection image of the glass cloth as the imaging device. The automatic inspection device according to claim 1, wherein: 6. An automatic detection method in which a woven fabric to be inspected is imaged by an imaging device, and when the level of an image signal obtained as an imaging result exceeds an allowable range, the woven fabric is determined to be abnormal. An automatic detection method characterized in that a waveform component of a normal woven cloth is removed from the resulting image signal, and the image signal from which the waveform component is removed is compared with an allowable range. 7. The automatic inspection apparatus according to claim 6, wherein a filter circuit is used to remove the waveform component, and the content of the filtering process is variably set according to the type of defect to be detected. Method. 8. A filter circuit is used to remove the waveform component, a normal woven fabric is imaged prior to automatic detection, and an image signal obtained as a result of the imaging is used to filter the filter circuit. The automatic inspection method according to claim 6, wherein the processing content is determined. 9. A filter circuit is used to remove the waveform component, and a filtering process content of the filter circuit is determined so that a level of an image signal input to the filter circuit falls within a predetermined range. The automatic inspection method according to claim 6, which is characterized in that.