KR100441313B1 - Apparatus for inspecting a defect of textile - Google Patents

Abstract

An object of the present invention is to provide a woven fabric examination apparatus capable of simultaneously extracting warp abnormality and warp abnormality with high precision under the same optical condition, and automatically inspecting them at low cost, without depending on the woven fabric.

A woven apparatus for woven fabric to image woven fabrics and to inspect the woven fabric based on image data of the woven woven fabric, the method comprising the steps of: calculating a tissue cycle of the woven fabric from image data of the woven fabric; And a CPU 17 for extracting defects based on the statistical amount of the image data in the comparison area extracted by the feature amount extracting circuit 12.

Description

[0001] APPARATUS FOR INSPECTING A DEFECT OF TEXTILE [0002]

The present invention relates to a specimen testing apparatus for woven fabrics, and more particularly to a specimen testing apparatus for automatically checking the presence or absence of defects in a woven fabric or woven fabric that has been woven.

Conventionally, as an apparatus or method for inspecting the appearance of a woven fabric, there has been known an apparatus or a method for inspecting the appearance of a woven fabric by imaging an image of the woven fabric surface with a camera and comparing the image density data obtained from the image pickup result with a threshold value, A method for automatically detecting an abnormality by irradiating a woven fabric, receiving the reflected light or transmitted light by a light receiving element, comparing the level of the received light amount with a threshold value, and detecting an abnormality. However, defects that can be detected by the examination apparatus or the examination method are limited to relatively large defects such as thread loosening which can be easily observed by a human eye. In addition, there has been a problem that the detection accuracy is significantly lowered due to vibration or disturbance.

As a method for improving the detection accuracy, for example, there is an invention disclosed in Japanese Patent Application Laid-Open No. 4-148852. There is disclosed a method of receiving light from a light source through a light receiving element through an optical slit arranged in a yarn direction of a subject to be inspected and transmitting the light transmitted through the woven fabric and detecting an abnormality from a comparison between the light receiving waveform and the reference waveform . This method is intended to determine the presence or absence of defects on the basis of the characteristic amount of the light transmitting portion, that is, the extracted woven fabric opening portion.

In addition, Japanese Patent Application Laid-Open No. 3-249243 discloses a method of detecting an abnormality based on a comparison between a difference value between the output of the light-receiving sensor and a preset threshold value, with two pairs of known light-receiving sensors. According to this method, since the density information obtained by dividing the narrow region of the woven fabric into two portions is reflected in the two pairs of comb-type light receiving sensors, even if there is vibration or disturbance light, the effect is obtained that the difference value output is canceled.

In addition to the characteristic inspection of the conventional woven fabric opening, it is necessary to extract the characteristic of the actual component itself in order to secure the accuracy equivalent to the inspection performed by observation. For example, as the observation inspection for defects such as warp inflow defects, the presence or absence of defects is determined from the confusion of the periodicity of the vertical relationship of the inspection target direction yarn on the yarn incorporation point. Therefore, if this defect is to be detected mechanically, it is better to extract only the yarn component above or below the warp yarn on the warp point coordinate. In order to detect this defect by image processing, binarization processing of the gradation image is indispensable, but in the conventional fixed binary coding method, extraction of only the warp component above or below the weft is not possible because the density of the object portion is not uniform .

The applicant of the present application has disclosed a device for examining a woven fabric which has been improved on the above-described problems in Patent Application No. 7-198171. In the present invention, the light irradiated to the woven fabric by the light projecting means is picked up by a CCD (Charge Coupled Device) element, and the woven fabric information (density of yarn, yarn inclination, woven structure, etc.) is calculated on the basis of the image data obtained thereby , The correlation value is compared with the set threshold value over the entire image data of the area having the actual pitch width obtained from the woven fabric information and the other area at the position apart by an integer multiple of the actual pitch width, And the other area is detected with high accuracy with respect to the entire width of the woven fabric under the same optical conditions without discrimination between warp and weft yarns.

In the invention disclosed in the above-mentioned Japanese Patent Application Laid-Open No. 4-148852, for example, defects can not be detected in the case of defects in which the opening of the woven fabric does not match the acceptable product like a warp inflow defect, and the detection accuracy is significantly lowered There was a problem. Further, in the present method, it is presupposed that the linear density is constant and the optical slit and the thread in the direction of the inspection object are parallel. However, there are various types of yarn density in actual woven fabrics, and there is a problem that the optical slit needs to be replaced every time. In addition, the actual weave finished yarn, particularly the warp yarn, is curved at both sides of the woven fabric, so that the above conditions can not be maintained and the detection accuracy is lowered.

Further, in the invention disclosed in Japanese Patent Application Laid-Open No. 3-249243, as in the invention disclosed in Japanese Patent Application Laid-Open No. 4-148852, there is a limitation in the defects that can be extracted, a change in the density, If the parallelism between the light receiving sensor and the room to be inspected can not be maintained, there is a problem that the detection accuracy is lowered.

In addition, all of the above-described inventions have a problem in that it is impossible to simultaneously detect the warp and weft defects with the same sensor.

In addition, in the invention disclosed in Japanese Patent Application No. Hei 7-198171, it is recognized that there is not much effect on defects of woven fabrics other than plain weave, such as a satin weave or a twill weave.

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above problems, and it is an object of the present invention to provide an image forming apparatus capable of simultaneously detecting warp abnormality and warp abnormality with high accuracy under the same optical condition, And to provide a device for examining a woven fabric.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic block diagram of a woven fabric testing apparatus according to an embodiment of the present invention; Fig.

2 is a diagram showing an example of a raw waveform and a filter waveform for explaining binarization processing;

Figs. 3 (a) to 3 (c) are diagrams showing an example of a binary image for explaining binarization processing and logical operation processing; Fig.

4 (a) to 4 (c) are diagrams for explaining a rectangular area set for extracting a statistic amount;

5 (a) and 5 (b) are diagrams showing examples of normal weave and defective weave.

6 is a flow chart showing a processing procedure of the apparatus for examining a woven fabric according to an embodiment of the present invention.

7 is a view showing a case in which a specimen processing apparatus according to the present embodiment is applied to a weaving machine.

Description of the Related Art

1: Light source

2: woven fabric during weaving

3: Optical lens

4: CCD camera

5: A / D converter

6: Frame memory

7: preprocessing circuit

8: FFT circuit

9: Concentration projection circuit

10: 2 matching circuit

11: Combined information integration circuit

12: Feature amount extraction circuit

13: Filter circuit

14: Image logic operation circuit

15:

16: CPU bus

17: CPU

18: ROM

19: RAM

20: I / O device

21: Shield plate

30: 1D / 2D converter

The invention described in claim 1 is an apparatus for examining a woven fabric for imaging a woven fabric and inspecting the woven fabric based on the image data of the worn woven fabric, the method comprising the steps of calculating the tissue cycle of the woven fabric from the image data of the woven fabric A comparison area setting unit for setting a comparison area of image data based on the organization period; a defect extracting unit for extracting a statistic amount from the image data in the comparison area and extracting a defect based on the statistical amount; .

By calculating the tissue cycle of the woven fabric and setting the comparison region of the image data based on the tissue cycle, the abnormality of the plurality of kinds of woven fabrics having different linear density, woven structure, It becomes possible to detect abnormality with high accuracy over the width.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic block diagram of a study apparatus according to an embodiment of the present invention. Fig. The examination apparatus includes a camera lens 3, a CCD camera 4, an A / D converter 5, a frame memory 6, a preprocessing circuit 7 for performing preprocessing such as differential emphasis, FFT (Fast Fourier Transform ) Circuit 8, a density projection circuit 9 for projecting the density of the inspection target yarn to the coordinate in the direction of the room to be inspected, a binarization circuit 10 for binarizing the density image, labeling the binary image in the rectangular area A characteristic amount extraction circuit 12 for extracting various characteristic quantities of the two-value image to which the label is attached, a filter circuit 13 for cutting off the intermediate frequency region, An image bus 15, a CPU bus 16, a CPU (Central Processing Unit) 17, a ROM (Read Only Memory) 18, a RAM An access memory) 19 and an input / output device 20 such as a keyboard or a display.

The light emitted from the light source 1 is transmitted through the gaps of the woven fabric 2 and is focused on the CCD element in the camera 4 by being converged by the camera lens 3. The light source 1 may be either a transmission type or a reflection type. However, the light source 1 can take a clear image of the upper and lower faces of the yarn at the yarn wedge point, and the wedge openings, The transmission method is preferable. When an area-type CCD camera is used as the light source 1, scattered light with uniform surface roughness is preferable. When a line-type CCD camera is used, the light source 1 may be a light source that irradiates a semiconductor laser, a HeNe laser, or the like in a slit shape using a lens, or a special light source A light source irradiated in a slit shape by a slit-type scattering paint or a slit-type optical fiber illumination may be used. Among them, the light source using the rod is most preferable in that the orientation characteristics in the width direction are uniform and high luminance can be obtained. In this embodiment, an example of using an area type CCD camera is shown. If a line-shaped CCD camera is used on the assumption that the woven fabric runs at a low speed, one-dimensional image data is divided into two (1) dimensional image data between the A / D converter 5 and the frame memory 6 Dimension / two-dimensional transformer 30 for transforming the image data into the two-dimensional image data.

There is no particular limitation on the amount of light passing through the woven fabric, but there is no particular problem insofar as it is a light quantity at a level capable of accumulating sufficient charge within the update period of the CCD. The enlargement magnification of the camera lens 3 is determined on the basis of the finest direct density among the direct density of the product. In general, an image having a higher expansion ratio of a warp image tends to be easier to extract a defect. However, as described later, since the size of the rectangular area from which the statistical quantity is extracted needs to be matched to the weaving cycle, it is premised that the number of yarns in the inspection object direction in the captured image is at least the number of real multiples of the tissue cycle .

The gradation image data photographed by the CCD camera 4 is converted into 8-bit digital image data by the A / D converter 5 and then stored in the frame memory 6. When the CCD camera 4 is of the line type as described above, the one-dimensional / two-dimensional converter 30 converts the one-dimensional image data into two-dimensional image data and stores the two-dimensional image data in the frame memory 6. The preprocessing circuit 7 performs preprocessing such as differential emphasis for emphasizing the outline in order to effectively extract defects. A texture period is required to set the size of the rectangular area from which the statistical amount to be described later is extracted from the image subjected to the preprocessing. This can be realized by the following method.

The short axis of the rectangular area is set in the vertical direction with respect to the thread to be inspected. The number of pixels in the minor axis direction of the comparison area is increased in order from the smallest value of the width of the minor axis (the number of pixels corresponding to the pitch size of the thread having the highest density of the thread to be inspected among the woven images picked up under the same optical condition) , The correlation value of image data in a pair of comparison areas (two comparison areas of the same size set in different areas) is obtained. It is found that the width of the short axis which is the maximum value among the required correlation values coincides with the organization period. The width of this short axis is defined as the short axis size of the comparison area. In addition, a method of obtaining the tissue period may be performed after conversion into a binary image, not a gradation image. Also, for example, the periodicity can be calculated by extracting the characteristic of the concentration waveform in the vertical direction and the periodicity in the direction perpendicular to the inspection target yarn, or in the method of obtaining the periodicity in the FFT, but the method is not particularly limited.

Next, in order to automatically calculate the average actual pitch and the average thread inclination amount in the inspection object thread direction, the concentration projection circuit 9 performs concentration projection (concentration addition processing) on the coordinate in the room to be inspected, . Fourier transform is performed on the density data by the FFT circuit 8, and real data and imaginary data of the spectrum maximum frequency value are required. From the real data and the imaginary data, the actual average pitch (straight density) and the tilt amount in the direction of the room to be inspected are required. In order to obtain the tilt amount, the above-described processing is performed by setting at least two or more image regions in the image region in the direction of the object to be inspected. Based on the obtained imaginary data, that is, the phase component data, The tilt is obtained by

Since the fast Fourier transform does not need to be performed all the time, it is also possible by soft processing by the CPU 17. In addition, this method has an advantage of being required to have high accuracy even when defect information such as annealing is included in image data.

As another method for the calculation of the room-direction, there are, for example, a method of differentiating the vertical direction of the room to be inspected and emphasizing the outline thereof and then tracking the peak value of the axial waveform, And a method of finding the yarn direction by a Hough transform that determines a straight line passing through the yarn. However, the method is not particularly limited.

Next, binarization processing is performed by the binarization circuit 10 on the image data. The binarization is performed to extract the characteristic quantities of the real components in the inspection direction. In particular, in the case of warp, it is necessary to extract only the component above or below the entanglement point with the weft yarn. This process will be described with reference to Fig.

First, density data in the X-axis direction of the image data at the cut portion of the woven fabric are sequentially extracted to generate a raw waveform. Then, a filter waveform is generated by passing the raw waveform through the filter circuit 13. The filter circuit 13 is a band elimination filter for blocking an intermediate frequency region containing a period component in the room direction of the inspection object to be inspected in advance. A slight offset value is added to the filter waveform obtained through the filter circuit 13 to obtain a threshold value. Then, whether or not the raw waveform is larger than the threshold value, the image data in the cut portion is binarized. This process is repeated while scanning the cut portion in the Y-axis direction to generate a first binary image. Likewise, a second binary image is generated by taking the cut section in the Y-axis direction, binarizing the image data at the cut section, and repeating it while scanning in the X-axis direction.

Since the first binary image is filtered and processed in the direction perpendicular to the warp direction, the binary image of only the warp component above the warp yarn on the blur point is generated in the first binary image. Further, in the second binary image, a binary image of only the weft component including the upper and lower portions of the warp yarn on the blended point is generated. In addition, the above-mentioned dichroism method has an advantage that only the actual information can be reliably extracted even if there are variations in illumination, unevenness in illumination, uneven local density of the woven fabric, or difference in glossiness of the yarn.

An example of an image after binarization is shown in Figs. 3 (a) and 3 (b). When a rectangular region process, which will be described later, is performed on the basis of the first binary image and the second binary image described above, defect information of the woven fabric is obtained. However, in the binary image in the warp direction, But also an intermediate image that penetrates downward. Since this information is unnecessary for defect detection of the woven fabric, in the case of the warp inspection, by performing the AND logic operation on the first binary image and the second binary image by the image logic operation circuit 14, It is possible to convert only the warp image on the warp weft yarn. An example of the binary image after conversion is shown in Fig. 3 (c).

A rectangular area is automatically generated for extracting statistical quantities from the organization cycle and real direction of the required yarn direction. An example of a rectangular area in the case where the weave is a plain weave is shown in Figs. 4 (a) and 4 (b). The length of the rectangular region in the major axis direction is not particularly limited, but the accuracy of defect detection tends to be improved as the defect is continuously generated such as thread pulling. In addition, with respect to defects caused by locally occurring fluffs or the like, the accuracy is improved in a short period of time. The length in the major axis direction may be determined in accordance with the characteristics of the woven fabric to be inspected, and the length may be varied during the inspection to repeat the same process a plurality of times.

The method of setting the rectangular area is not limited to a pair of adjacent rectangular areas (area 1 = A 占 H, area 2 = B 占 H) shown in Figs. 4 (a) and 4 (b). That is, as shown in FIG. 4 (c), the rectangular area to be compared may be alternately set (area 1 = A × H + C × H, area 2 = B × H + D × H). However, it is presupposed that the size (number of pixels) in the short axis direction of the rectangular area coincides with the rectangular organization period of the yarn in the direction of the yarn to be inspected. By setting the rectangular area in this manner, it is possible to perform the same inspection regardless of the shape of the woven fabric. However, if the phase of the direction of the major axis of the rectangular area deviates from the direction of the room to be inspected, the detection accuracy of defects lowers. In order to avoid this problem, as described above, the direction of the room to be inspected is automatically calculated to follow the long axis direction of the rectangular area in the target room direction, or the image is rotated by the tilt amount of the target room, Align the direction of the room to be inspected. As a result, for example, it is possible to avoid deterioration of the defect detection accuracy due to the inclination of the warp yarns occurring on both sides of the woven fabric during weaving and the missed misses when the sensor is fixed.

Further, by automatically calculating the average density of the yarns to be inspected and automatically optimizing the long axis of the rectangular region for calculating the statistical value, even if the woven fabrics having different straight densities are inspected, the same inspection is performed without any adjustment of the optical conditions There is an advantage to be able to. In addition, even when the light amount of the light source is relatively lowered or increased during the inspection, for example, during the statistical amount comparison process between adjacent rectangular areas, there is an advantage that these influences are offset. However, in the case where light causing halation is present, the shielding plate 21 shown in Fig. 1 may be provided.

The combined information integration processing circuit 11 performs labeling of a binary image in a rectangular area. For example, in the case of the binary image shown in Fig. 3 (c), a label is attached to each of binary images extracted as warp components above the weft yarn on the interlaced points. Then, the feature quantity extraction circuit 12 extracts the feature quantity (statistical quantity) of the binary image in the rectangular area.

The statistic amount in the rectangular area includes the total number of pixels of black or white, the total number of labels, the area of the two-dimensional image attached with the label, the area of the pitch, the aspect ratio, the main axis angle, Distance, etc., and the defect is extracted by comparing the difference value of these statistical amounts, the image pattern correlation or the statistical amount with the reference data. However, the image pattern correlation may be performed not only in the binary image but also in the original gradation image. Further, the piercing means the length of the projection part in the X-axis direction and the length of the projection part in the Y-axis direction of the binary image of one lump with the label attached thereto.

FIGS. 5A and 5B are diagrams showing examples of defect extraction, in which FIG. 5A shows a normal weave pattern, and FIG. 5B shows an example of a defective weave pattern. Such a defect is difficult to be extracted in the comparison of the conventional woven cloth openings. However, as shown in Fig. 5, when the binary image by the above-mentioned binarization processing is compared within a rectangular area, The position pattern matching in the rectangular area is completely different. That is, the positional pattern of the warp component (indicated by a black square in FIG. 5) on the weft yarn on the intertwin point is completely different. Thus, it can be seen that defects in the woven structure can be easily extracted by calculating the pattern consistency between the pair of rectangular regions. The defect is extracted by the CPU 17 based on the above-described statistical amount. The calculation result is output by the display of the input / output device 20 or the like.

When the woven fabric being weighed is inspected inline, foreign matter such as dust may adhere to the surface of the woven fabric. As described in the background art, if defects are to be extracted only by comparing the presence or absence of defects based on the characteristic amounts of the woven fabric openings, or by comparing the density data in the two pairs of neighboring regions, these defects may be erroneously determined as defects Throw away. In order to avoid this problem, a comprehensive comparison of the above-described statistical amounts and a comparison between a statistical amount and a reference value are performed. For example, in the inspection of the transmission method, when there is foreign matter, when the warp defect of the difference in lead penetration and the defect due to the foreign object are compared between the above-mentioned rectangular areas, the average density value which is an average value of the density data in the rectangular area clearly differs. Therefore, in addition to the conventional arithmetic method, by adding the comparison of these statistic amounts to the judgment, it becomes possible to separate the disturbance elements such as foreign bodies from defects. The statistical quantity serving as a parameter at the time of the judgment shown here is not particularly limited, but a statistical quantity showing a characteristic of a target defect may be obtained by experiment or the like, or a bar graph may be created during inline, , Or a combination thereof may be used. In addition, since this method can extract rich statistics at the same time, it is also possible to identify defects in fuzzy inference or multiple regression analysis, for example.

Fig. 6 is a flowchart showing a processing procedure of the apparatus for examining a woven fabric according to the present embodiment. First, the grayscale image data picked up by the CCD camera 4 is converted into 8-bit digital image data by the A / D converter 5, and then received in the frame memory 6. At this time, if the CCD camera 4 is of the line type, the one-dimensional image data is converted into two-dimensional image data by the one-dimensional / two-dimensional converter 30 and then received in the frame memory 6 (S1).

Next, the preprocessing circuit 7 reads the grayscale image data received in the frame memory 6 and performs preprocessing such as differential emphasis for emphasizing the outline of the image (S2). The CPU 17 extracts the tissue period of the inspection target thread by the above-described method based on the density image data subjected to the preprocessing process. Further, after the density projection data 9 is projected by the density projection circuit 9, the FFT circuit 8 performs Fourier transform on the density-projected image data to calculate the real direction (S3).

The binarization circuit 10 generates a binary image (first binary image) of only a warp component above the interlaced point weft by performing filter processing on the gradation image data (S4) (Second binary image) of only the weft component including the upper and lower portions of the warp yarns of the warp yarns (S5).

The image logic operation circuit 14 performs arithmetic operation of the first binary image and the second binary image to generate a binary image of only a warp component above the warp point weft yarn, And a third binary image which is a binary image of only the weft component is generated (S6).

Next, the CPU 17 sets a rectangular area to the third binary image from the texture period and the actual direction of the room to be inspected (S7). The feature quantity extraction circuit 12 extracts various statistical quantities from the binary image data within the set rectangular area (S8). When a pair of rectangular areas is set, the presence or absence of a defect is determined by comparing various statistical values of the respective rectangular areas. When the reference value is set in advance, the presence or absence of the defect is determined by comparing the various statistical quantities of the rectangular area with the reference value.

In step S1, it is determined whether or not inspection has been performed on all of the received image data. If the inspection is not completed (S10, No), the rectangular area set as the third binary image is set to be shifted in the direction perpendicular to the direction of the room to be inspected (S12). Then, the processing in steps S8 and S9 is repeated. If the inspection is completed (S10, Yes), it is determined whether or not to inspect the inspection target yarn in the other direction. If the inspection is to be carried out (S11, Yes), the processes in and after step S7 are repeated. If the inspection is not to be performed (S11, No), the inspection result is output to a display or the like on the input / output device 20 (S13).

The program for performing the entire control is stored in the ROM 18. When inspecting the entire width of the woven fabric, the sensors may be traversed in the woven fabric width direction or a plurality of sensors may be arranged at uniform intervals in the woven fabric width direction.

Although the above description relates to an inline inspection process of woven fabric during weaving, it can also be applied to an automatic inspection of a woven fabric that has been completed. It is also applicable to a sheet having features other than woven fabric.

Fig. 7 is a view showing a case in which a specimen analyzing apparatus according to the present embodiment is applied to a weaving machine. The light source 1 is irradiated to the woven fabric 2 being weaved from below, and the image is picked up by the CCD camera 4. The CCD camera 4 is mounted so as to be movable along the moving shaft 40.

INDUSTRIAL APPLICABILITY As described above, according to the present invention, it is possible to reliably detect other defects in the weaving structure in which it is difficult to detect. In addition, even if the straight density changes or the yarn direction of the inspection object changes, the defects can be extracted at the same time and with high accuracy without changing the conditions of the optical system at all. In addition, by extracting the abundant statistical amount, it becomes possible to separate the disturbance elements such as surface deposits from the defects, and more accurate identification becomes possible.

Claims (1)

A woven apparatus for a woven fabric for imaging a woven fabric and inspecting the woven fabric based on the image data of the worn woven fabric, Tissue cycle calculating means for calculating a tissue cycle of the woven fabric from the image data of the woven fabric; Comparison region setting means for setting a comparison region of the image data based on the organization cycle; And defect extraction means for extracting a statistic amount from the image data in the comparison region and extracting a defect based on the statistical amount.

Images