JPH011939A - Yarn package inspection device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Field of Application] The present invention is for automatically inspecting the surface of synthetic fibers, especially polyester fibers, in the form of products such as burnt, cheese, or corn (hereinafter referred to as packages). related to a device for

[Prior art] Technology has been developed to automatically inspect the appearance of polyester m-fibers and polyamide fibers rolled up into packages using the v1 machine. The method used is to irradiate the package with light and measure the difference in brightness of the soiled area using a television camera. Since black stains such as oil and hand grime have a large difference in brightness, this method provides sufficient practical detection power.

However, with the conventional method described above, there are the following abnormalities that are difficult to detect. In other words, this is a case where an abnormal product is mixed into the raw material polymer. In the case of darning, if the abnormality is mild, it will not affect the stringiness and can be rolled up into a package, but there will be a difference in dyeing performance and the product value will be greatly impaired. If you carefully observe a package that contains such a part, you will notice that the abnormal part is wrapped in layers, so the edge appears in a ring shape on the surface of the package, and the color tone has changed very slightly compared to the base part. 6 However, as mentioned above, it is difficult to automatically detect slight color tone abnormalities with the existing methods. Due to the magnitude of the impact, inspectors are currently conducting inspections with great care, but the detection rate is not necessarily satisfactory.

Purpose of the Invention The present invention has been made in view of the current situation, and it is an object of the present invention to provide a yarn palage inspection device that can reliably detect the image color tone abnormality.

Structure and operation of the invention] The present invention provides a yarn package inspection device that photographs a yarn package with an image sensor camera and determines abnormalities based on the detection signal, which irradiates the package with near-ultraviolet light. The yarn package inspection apparatus is characterized in that the irradiated surface of the package is photographed through a blue transmission filter.

In the present invention, it is advantageous for the structure to be simplified if the yarn package is rotated about its axis and an image is taken by a line sensor that scans the package surface in the axial direction.

In addition, a gain compensation circuit is installed to adjust the level of the detection signal from the image sensor camera, and the gain is determined at the first scan of the package to be inspected, and the measurement is kept fixed during the inspection of the bag. It has the advantage of being able to perform stable inspections even in the face of fluctuations.

The present invention described above was accomplished as follows.

The present inventors have discovered that in polyester fibers, thermal decomposition products generated on the surface during the production process become phosphors, which are excited by near ultraviolet rays and have wavelengths of 400 to 420 nlU.
Focusing on the fact that it generates blue fluorescence with a peak at
We succeeded in developing a fully practical automatic inspection device as shown below.

In other words, in a package with no abnormalities, if the irradiation intensity is uniform when illuminated with near ultraviolet light, fluorescence will be generated uniformly, but if there is an abnormality in the polymer, the fluorescence in that area will weaken and become darker. appear. This difference in brightness is larger than that caused by visible light illumination when observed with the naked eye, and when measured using an image sensor equipped with a short-wavelength transmission filter (hereinafter referred to as a blue transmission filter) that includes the fluorescent region mentioned above, It turns out that it is even easier to observe. In addition, this blue transmission filter is specifically 500n
It is sufficient that it transmits wavelengths of 1 or less, more preferably 450n1 or less. When using the same method to observe dirt attached to the surface of a package, it was also confirmed that the sensitivity was higher than with the previous method. In addition, when a fluorescent substance stronger than the base fibers, such as oligomers, is included, the abnormal area becomes brighter, so it has been confirmed that the same configuration can detect abnormalities, and it is easy to identify foreign substances that cause color abnormalities in general. This invention has been made to enable automatic detection.

Needless to say, in the case of detecting contamination of a fluorescent substance, the present invention is also applicable to objects such as polyamide fibers in which the normal pores do not contain fluorescent substance.

The details of the present invention will be explained below using an example in which the present invention is applied to the burning of polyester fibers.

FIG. 1 is an explanatory diagram of the basic configuration of an embodiment for measuring a pan, and numeral 1 indicates a pan, which is rotated by a rotating device (not shown). Reference numeral 10 denotes an image sensor camera using a line sensor as a light receiving element, and three cameras are arranged to take images of the upper tapered part, body part, and lower tapered part of the piern 1, respectively. The focus is set on the pan surface of each of the above-mentioned parts so as to indicate the photographing position S, and scanning is performed in the direction of the pan axis. The lens system of each camera 10 is equipped with a blue transmission filter that includes a fluorescent region.

2 is a black light that generates near ultraviolet rays. In the 6 examples, two straight tube lights 2 are placed parallel to each side of the photographing position S, and illuminate the part of the pan surface that corresponds to the camera field of view. A black light with a wavelength of 360 nn was used. Reference numeral 100 denotes a determining means that processes the output of the camera 10 to determine whether there is dirt or abnormality in color tone.

In the above configuration, when the surface of the pern is irradiated with near-ultraviolet rays, fluorescence is emitted from the fluorescent substance contained in the fibers.

In the case of polyester fiber, some of it thermally decomposes during the manufacturing process and becomes a fluorescent substance, which is evenly distributed within the fiber.
When exposed to near-ultraviolet light of around 1M, it is observed with the naked eye that the entire paan turns blue.

If dirt adheres to the pan surface, fluorescent coloring will be inhibited and it will become dark. Furthermore, in the case of poly-abnormality, although the difference in brightness is small, it takes the form of stripes that go around the pan once, so it can be easily determined.

The run sensor of the camera 10 needs to have sufficient sensitivity on the short wavelength side in order to detect the above-mentioned fluorescence, and also to limit the sensitivity to the fluorescence region and block the influence of stray light from the surroundings reflected on the pan surface. A blue transmission filter is installed. The cutoff wavelength of the blue transmission filter was set to 500n11.

Specifically, the field of view of the line sensor of the camera 10 is along the rotation axis of the surface of the pan as shown in the figure, and as it rotates around the axis, adjacent parts of the surface of the pan are scanned. This means that the entire circumference of the pan was photographed.

The size of the dirt you want to detect is also closely related to the optical system and pan rotation system. If you want to detect small dirt, use a sensor with a large number of elements, reduce the magnification, increase the scan frequency, and slow the rotation speed. However, in this example, each camera 10 using a 512-element line sensor has a pan length of about 150 cities! ! It is set so that it can cast shadows, the scanning frequency is 100 times/second, the pan is rotated at 1 rotation/second, and the pan surface is 3.6
Photographs are taken every once in a while.

Since the camera field of view is insufficient to measure the total length of the pan with one camera, three cameras 2 are set as described above to measure the total length of the pan.

The determining means 100 processes the brightness signals obtained from each camera 10 in the following manner to determine dirt and color tone abnormalities. In other words, the brightness signal from the camera 10 reflects dirt or color tone abnormalities on the pan surface during scanning.
For example, when measuring dirt A and color tone abnormality B on the pan surface in Figure 2 (2), the output signals for each scan a, b, c, and d are respectively a, b, and c in Figure 2 (1). , d, stain A generally falls deeply, and color tone abnormality B falls shallowly.

Therefore, it is sufficient for the stain A determination means to determine the area where the amount of drop from the normal area is greater than the appropriately set darkness value Thd. On the other hand, the means for determining color tone abnormality B has a small amount of drop in the abnormal area, so if you try to determine a threshold that can detect this, it will end up detecting excessive brightness difference noise in the normal area, making it difficult to make a correct judgment. It's difficult. However, as shown in e in Figure 2 (1), when an appropriate number of scans are integrated, the noise in the normal area is averaged out and disappears, whereas the abnormal color tone B area exists over the entire circumference. This makes the amount of decline clear.

In this example, an abnormality is determined by comparing 100 pan-circle scans with an appropriate threshold value Ths after integrating 100 scans, thereby enabling accurate and sensitive detection.

A specific circuit example of the above-described determining means 100 is shown in FIG. 3, and its operation will be described below. Assume that the line sensor 11 installed in the camera 10 is being scanned and the output of the i-th element is currently being read.

The output of the element is converted into an 8-bit digital value by an A10 converter 110, and then transferred to a 4-stage 8b digital value via a changeover switch 113.
It is input to an 8-bit subtracter 115 and a shift register 114 . 'The other input of the LAX unit 115 is the (i-4)th element output output from the shift register 114, and the subtracter 115 calculates the difference between the i-th and (i-4)th outputs, A comparator 118 compares the threshold value Thd set in the latch 116 in advance.

In this way, if the brightness difference at a point separated by four line sensor elements exceeds the threshold value, it is determined that the stain is A.

When the brightness of the pan surface fluctuates or when the brightness differs depending on the location, investigating the difference between several distant points in this way is effective for stable detection. On the other hand, data is stored in the memory 112 at address i corresponding to the i-th element. The numerical value stored at address i is read out and added by adder 111 to the output of the i-th element converted to a digital value by A/D converter 110, and the result is stored at address i again. . The contents of memory 12 are cleared before each pan measurement, so when 100 scans are completed, memory 11 is cleared.
2, the product X value of brightness for one round of pan for each element of the line sensor 11 is recorded for each element. Pan-
When the measurement of the circumference is completed, the selector switch 113 is switched to the memory 112 side, and the output integration result of 100 times or the total length of the burn is measured for one circumference of each element stored in the memory 112. is read from 0 to 512# and shifts to sister 114:gx
, and is sent to the device 115, and the difference for every four elements is calculated in the same way as in the case of dirt A.

This output difference is compared by the comparator 118 with a threshold value Ths preset in the latch 117 selected by the changeover switch 119 that is being switched at the same time as the changeover switch 113, and if it exceeds the threshold value, the color tone abnormality B is detected. It will be judged.

Note that the intensity of fluorescence varies depending on the proportion of additives mixed into the polyester fiber raw material, so it varies depending on the brand, and it also changes if the intensity of the near-ultraviolet light to be irradiated changes. By creating an algorithm that detects dirt A based on the difference in intensity from the normal part, the influence of such intensity differences can be considerably suppressed, but if the signal output is compensated appropriately, the reliability of the determination can be further improved. Needless to say, it is possible to increase the brightness of the black light that is emitted.The luminous intensity of the black light gradually decreases over time, but in order to compensate for this change, the power supply voltage of the black light that is emitted so that the brightness at a fixed location remains constant. A method is used to adjust the

In order to compensate for differences in brands, the strength of each brand is investigated in advance, and the level is selected and set in accordance with the brand to be tested. However, this method requires adjustment in two places: the light source and the measurement system or signal processing system of the determination means, so the inventors developed the following method to compensate by directly using the measurement output. .

That is, this is a method in which an appropriate degree of increase in rIJ is directly calculated from the fluorescence intensity obtained by scanning once before starting PANN measurement, and the value is fixed at that value during the measurement.

The configuration of the gain compensation circuit will be explained below with reference to FIG. FIG. 4 (1) is a block diagram of the basic configuration for explaining its operation, and FIG. 4 (2) is a block diagram of an example of the circuit. A gain compensation circuit, specifically a gain setting circuit 121 for automatically setting the gain, and an increaser [tJ unit 122] are provided between the /D converter 110 and the gain compensation circuit 110. The output of this amplifier riJ is supplied to the aforementioned A/D converter 110 and comparison amplifier 123, where the difference between it and an appropriate target value Vr is calculated and amplified rl.

Before starting measurement for inspection, in one scan, in the central element of the line sensor 11, specifically in the section from the 128th pixel to the 383rd pixel, so that this difference becomes zero.
The input division ratio of the lJ amplifier 122 is automatically adjusted by a gain setting circuit 121 consisting of an electronically controlled potentiometer.

The movable terminal of the gain setting circuit 121 is fixed by the adjustment port R124 during the inspection of one pan, that is, during 100 scans. The circuit adopted to carry out these functions will be explained in more detail with reference to FIG. 4(2).

The output of the line sensor 11 is input-divided by a gain setting circuit 121 and then supplied to the A/D converter 110 through an rtI amplifier 122.

The purpose of the compensation circuit is to stabilize the judgment by setting the voltage level of the input to the A/D converter 110 to an appropriately set value Vr even when there are changes in lighting or differences in brands when measuring a normal part. It measures the Therefore, the difference between the output of the amplifier 122 and the reference voltage Vr is amplified by the Amplifier 123, and the output is sent to the sample and hold circuit 124a. The sample bold circuit 124a samples and holds the sample at a timing synchronized with the scan timing t of the line sensor 11, and connects only the parts where the output is stable for each element of the line sensor to make it continuous.

The signal thus made continuous is supplied to the code detector 124b and the absolute value circuit 124C. The output of the absolute value circuit 124C is converted by a V/F converter 124d into a rectangular pulse train signal having a frequency proportional to the voltage. V/F converter 124d
There is a gate that prohibits operation, and oscillation is stopped by a prohibition signal Si. This rectangular pulse train signal and the output of the sign detector 124b are used to control the position of the variable resistance terminal of the gain setting circuit 121, specifically, an electronically controlled potentiometer incorporated as part of the input resistance of the measurement amplifier. An electronically controlled potentiometer has a structure in which the brush position of the variable resistance terminal changes by 1/100 for each pulse applied to the input terminal, and the direction of change is determined by the on/off of the signal applied to another terminal. It becomes.

In this example, commercially available R2PO'r' manufactured by XI COR was used. According to this, when the pan 10 is in place,
The gate is opened by the Si signal and the V/F converter 124d is operated, and the output of the amplifier 122 and the set voltage Vr at that time are
The gain setting circuit 12 generates a pulse train with a frequency proportional to the difference between
At this time, when the measured output is larger than Vr, the resistance value increases and the amplification rate decreases, and when it is smaller than Vr, the amplification rate is increased. Ru.

By constructing a control loop in this way and appropriately selecting loop gains such as the amplification rate of the differential amplifier 123 and the conversion rate of the V/F converter 124d, the line sensor 11 can be By sufficiently setting the gain of the gain setting circuit 121 to an appropriate point, the output of the amplifier 122 can be made to match Vr. When selecting the loop gain, it is necessary to set the loop gain appropriately, since if the response speed is increased too much, it will become too sensitive to noise contained in the detection signal.

When the gate is closed by the Si signal at the end of the scanning of the above-mentioned element, specifically the 383rd pixel, after allowing a suitable settling time, the oscillation of the V/F converter 124d stops and the resistance value of the gain setting circuit 121 changes. Therefore, the division ratio of the sensor output is fixed.

After that, the line sensor is scanned 100 times, and from the results, dirt A or color tone abnormality B is detected by the signal processing described above. By the method described above, it is possible to simultaneously compensate for light source deterioration and brand differences.

When we inspected polyester fibers with the above configuration, we confirmed that we could detect dirt up to 0.5 mm in diameter, and that color tone abnormalities could only be detected with the naked eye of the middle two cabinets if we were very careful. It was confirmed that light color tone abnormalities could be detected.

Note that in the above embodiment, each circuit of the determination means 100 is provided for each of the cameras 10, but
The processing may be performed by a common circuit for the camera 10, and the determination means 100 may be configured by a digital computer.

The number of cameras used for inspection is not limited at all, and should be appropriately designed based on the shape and size of the yarn package to be inspected.

As described above, the present invention uses a light source of near-ultraviolet light to photograph the yarn package with an image sensor through a blue filter. This has created an inspection device that can also inspect for color tone abnormalities, making a major contribution to streamlining inspection work.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of the basic configuration of an embodiment of the present invention, FIG. 2 (11i2) is an explanatory diagram of the processing contents of the determination means in the embodiment, and FIG. 3 is a block diagram of the processing circuit of the determination means. , FIG. 4 (11, (2+ is the basic r of the gain compensation circuit of the determination means)
FIG. 4 is a block diagram of four components and a specific configuration. 1: Pan 2 Black light 10: Camera 100: Judgment means 110: A/D converter 1
21 gain setting circuit 122: amplifier (1)
(-111-Izu (72nd)
(Korao 30・

Claims (5)

[Claims] (1) In a yarn package inspection device that photographs a yarn package with an image sensor camera and determines abnormalities based on the detection signal, near-ultraviolet light is irradiated onto the package, and the package is irradiated with near-ultraviolet light. A yarn package inspection device characterized by photographing the surface through a blue transmission filter. (2) Claim 1, in which the package is rotated around its axis and an image sensor camera arranged to scan the package surface in the axial direction captures an image.
The yarn package inspection device described in Section 1. (3) The determining means for determining an abnormality based on the detection signal includes a color tone abnormality determining means for determining a color tone abnormality based on the integrated value of the detection signal for one rotation at a fixed position in the axial direction of the package, 3. The yarn package inspection device according to claim 2, further comprising a dirt determining means for determining dirt based on the difference between the detected value and the detected value a predetermined time ago. (4) The yarn package inspection device according to any one of claims 1 to 3, wherein the determining means includes a gain compensation circuit that adjusts the level of a detection signal from an image sensor. (5) The gain compensation circuit is a circuit that determines a gain for each package to be inspected based on the detection signal of the first scan of the package, and maintains the gain at the determined gain during the package inspection. The yarn package inspection device described in Section 1.