JP2906662B2 - Tape wrinkle inspection method and apparatus - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a method and an apparatus for inspecting a tape wrinkle performed in a step of inspecting and adjusting a wrinkle state of a tape in the production of a video deck, and an adjusting device using the same. .

2. Description of the Related Art A conventional tape wrinkle inspection method and an inspection and adjustment device therefor will be described with reference to the drawings.

FIG. 12 is a diagram showing a configuration of a wrinkle inspection and adjustment apparatus using a conventional tape wrinkle detection unit by moire fringe processing. FIG. 13 is a diagram showing a state of light projected on the tape by the tape wrinkle inspection adjusting device shown in FIG.

In FIG. 12, 1 is a light projector, 2 is an imaging device, 3 is a diffraction grating for generating moiré fringes, 4 is a tape, 5 is an image processing device, 6 is an actuator control device, and 7 is a tape wrinkle adjustment. An adjustment actuator.

In FIG. 13, reference numeral 8 denotes moire fringes projected on the tape 4 by the diffraction grating 3. Reference numeral 9 denotes reflected light from a deck (not shown). Reference numeral 10 denotes the center line of the moiré fringes 8. Reference numeral 11 denotes an approximate straight line of the moiré fringes 8. Reference numeral 12 denotes a displacement of the moiré fringes 8 with respect to the approximate straight line 11.

In FIG. 12, the light projected from the light projector 1 passes through the diffraction grating 3 and is projected on the tape 4 as moire fringes. The projected light is taken in by the imaging device 2, and wrinkle detection of the tape 4 is performed by the image processing device 5. The criteria for determining wrinkles on the tape 4 are the size of the moire fringes and the position of the bent fringes. In FIG. 13, the degree of bending of the moire fringes is given by the maximum value of the displacement 12 of the center line 10 of the moire fringes 8 with respect to the approximate straight line 11 of the moire fringes 8. In addition, the position of the bent stripe is the displacement 12 that takes the maximum value.
Is given by In this manner, the image processing apparatus 5 identifies the state of the wrinkle and performs the wrinkle inspection. When the wrinkles of the tape 4 are not in an appropriate state as a result of the inspection, the actuator control device 6 moves the adjusting actuator 7 to adjust the wrinkles. Thereafter, the inspection adjustment is repeated until the wrinkles are in an appropriate state.

However, in the method of projecting the light on the tape 4 through the diffraction grating 3 as shown in FIG. 12, the reflected light 9 from the deck is also projected as shown in FIG. Also,
The moire fringes 8 projected on the tape 4 by the light projector 1 do not have a clear boundary with the tape 4 as a background, and have poor contrast.

Due to the matters described above, the image captured by the imaging device 2 is not in good condition, and the image processing device 5 requires a complicated image to calculate the maximum displacement of the moiré fringe 8 with respect to the approximate straight line 11. Processing algorithms are required. Therefore, in the apparatus shown in FIG. 12, it takes about several tens of seconds to detect wrinkles. This speed is not practical. Further, in order to detect the wrinkle state of the running tape, it is necessary to detect a dynamic change in the wrinkle of the tape. For this reason, the wrinkle detection time must be short.

Further, as shown in FIG.
And it is difficult to position it.

Further, in the identification of wrinkles, the change in the state of minute wrinkles is complicated, and the state is not determined only by the maximum displacement amount of the moiré fringe 8 with respect to the approximate straight line 11 and its position. In other words, it is necessary to extract more feature amounts of wrinkles, and to appropriately match a plurality of feature amounts to changes in the state of complex wrinkles.

In addition, the change in the state of a minute wrinkle is complicated, and the pass / fail criterion of the wrinkle state currently depends only on human sensual judgment. That is, it is necessary to correlate the wrinkle identification result with a human sensual judgment.

Further, due to the matters described above, the apparatus shown in FIG. 12 cannot properly detect a subtle wrinkle change state, so that tape wrinkles can be performed at high speed based on the wrinkle state detected using this apparatus. It is difficult to adjust properly.

In order to solve this problem, the present invention provides a tape wrinkle inspection method for facilitating the positioning of a stripe pattern, inspecting a wrinkle state with a projected image in which a pattern boundary having good contrast is clear, and providing a plurality of feature amounts. Is a tape wrinkle inspection device that can identify complex and minute wrinkle state changes at high speed by appropriately matching the wrinkle state and perform high-speed inspection by associating it with human judgment. To provide.

Further, another object of the present invention is to provide a tape wrinkle adjusting device capable of adjusting the wrinkles of a tape quickly and appropriately based on the state of the identified wrinkles as described above.

DISCLOSURE OF THE INVENTION The tape wrinkle inspection method of the present invention projects reflected light of a striped pattern onto a running tape, and sets a plurality of pitches of the striped pattern projected on the tape orthogonal to a running direction of the tape. Each is calculated on the inspection line, and a histogram relating to the number of abnormal pattern pitches having a value smaller or larger than the reference value among the calculated pattern pitches is created on a coordinate axis parallel to the inspection line, and the stripe pattern is formed. The tape is inspected for wrinkles by extracting a plurality of feature values of the pattern.

Further, the tape wrinkle inspection method of the present invention is the above-described inspection method, wherein the inspection is performed by identifying a wrinkle state of the tape using a FUZZY inference mechanism based on information on a plurality of feature amounts of the striped pattern. It is.

Further, the tape wrinkle inspection apparatus of the present invention is a device for projecting reflected light of a striped pattern onto a running tape, an imaging device for capturing and capturing an image of the striped pattern projected on the tape, and a captured image. An image binarization unit for binarizing
An image data storage unit for storing binarized image data, and a plurality of image data stored in the image data storage unit, wherein a plurality of pitches of a stripe pattern projected on the tape are orthogonal to a running direction of the tape. A pattern pitch calculation unit that calculates each on the inspection line, a pattern pitch storage unit that stores data of the calculated pattern pitch, and a value smaller than a reference value from the data stored in the pattern pitch storage unit. Or, an abnormal pattern extraction unit for extracting a large abnormal pattern pitch, an abnormal pattern pitch storage unit for storing the extracted abnormal pattern pitch, and a histogram relating to the number of abnormal pattern pitch data stored in the abnormal pattern pitch storage unit An image feature amount calculation unit that creates a coordinate on a coordinate axis parallel to the inspection line; The image feature amount calculated by the collection amount calculation unit is input, and a plurality of inference rule calculation units that calculate the fitness level of the inference rule corresponding to each wrinkle state taught in advance, and the fitness level of each inference rule are input. And a wrinkle state determination unit that determines a wrinkle state by selecting an optimal rule.

Further, in addition to the above configuration, the inference rule calculation unit and the wrinkle state determination unit are configured by a FUZZY inference mechanism.

Further, the tape wrinkle adjusting device of the present invention uses the tape wrinkle inspection device having the above-described configuration, and based on the information about the tape wrinkle state output from the wrinkle state determination unit of the tape wrinkle inspection device, the tape Equipped with an adjustment actuator for adjusting the mounting position of the member supporting the traveling of the
It is configured to adjust the state of wrinkles on the tape.

Therefore, according to the present invention, the reflected light of the striped pattern is projected onto the running tape, and the pitch of the striped pattern projected on the tape is analyzed on a plurality of inspection lines, and a plurality of striped patterns are analyzed. By extracting the characteristic amounts of (1) and (2), an image with good contrast and clear pattern boundaries can be obtained, and the positioning of the striped pattern can be simplified. Further, it is possible to detect a change state of a minute wrinkle as a dynamic property of the running tape to a position and a degree in a short time.

In addition, since the state of wrinkles in the striped pattern is identified using the FUZZY inference mechanism based on information on multiple feature amounts, complex and minute wrinkle state changes can be quickly identified by appropriately matching multiple feature amounts. As well as
Inspection can be performed at a high speed in association with wrinkle conditions with human judgment.

Means for projecting the reflected light of the striped pattern onto the running tape, and calculating the pitch of the striped pattern projected on the tape on a plurality of inspection lines to extract a plurality of feature amounts of the striped pattern. Means and means for identifying the wrinkle state of the striped pattern using the FUZZY inference mechanism based on information of a plurality of feature amounts, so that the positioning of the striped pattern is simple, the contrast is good, and the pattern Images with clear boundaries can be obtained, and the change of minute wrinkles as a dynamic property of the running tape can be detected in a short time to its position and degree. By appropriately matching the amounts, it is possible to identify a plurality of small wrinkle state changes at high speed, and to inspect the wrinkle state at high speed in association with human judgment.

Further, using the above-described configuration, it is provided with an adjustment actuator for adjusting the mounting position of the member supporting the running of the tape based on the information of the detected wrinkle state, and is configured to adjust the wrinkle state of the tape, Adjustment can be performed until the wrinkles reach a desired state to complete the assembly, high-speed and appropriate adjustment can be performed, and a series of steps can be automatically performed.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a schematic configuration of a tape wrinkle inspection device according to a first embodiment of the present invention, and FIG. 2 is a schematic configuration of a tape wrinkle adjusting device according to a second embodiment of the present invention. FIG. 3 is a diagram showing the configuration of the image feature amount extraction unit 17 in FIG.
FIG. 4 is a diagram showing a data structure compressed by the image data compression unit 21 in FIG. 3, and FIG. 4 (a) is a diagram showing binary image data stored in the image memory, and FIG. FIG. 5B is a diagram showing compressed data stored in the image memory, FIG. 5 is a diagram for explaining how to extract a wrinkle characteristic amount by the tape wrinkle inspection apparatus shown in FIG.
The figure shows the structure of the wrinkle state determination unit 18 in FIG. 1,
FIG. 7 is a diagram for explaining how the degree of rule conformance is calculated in the inference rule calculation unit in FIG. 6, and FIG. 8 is a diagram illustrating how the membership function teaching unit in the membership function teaching unit in FIG. FIG. 9 is a diagram showing how the image feature amount histogram and the membership function are created in the image feature amount histogram creation unit and the membership function creation unit in FIG. FIG. 9 (a) is a diagram showing an image feature amount histogram,
9 (b) is a diagram showing a membership function, FIG. 10 is a diagram showing a configuration example using a microcomputer of the image processing device 19 in FIG. 1, and FIG. 11 is a main part near a loading post of a video tape. FIG. 12 is a perspective view, FIG. 12 is a diagram showing the configuration of a tape wrinkle inspection and adjustment device using a conventional tape wrinkle detection unit using moire fringe processing, and FIG. 13 is projected onto a tape by the tape wrinkle adjustment and inspection device shown in FIG. FIG.
Fig. 14 and Fig. 15 show the ambiguity of meaning quantitatively, and FUZZ
It is a figure explaining the concept of the membership function in Y theory.

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing an outline of a tape wrinkle inspection apparatus according to a first embodiment of the present invention. In the figure, 13 is a light projector, 14 is a reflection plate on which a striped pattern is printed, 15 is a tape, 16 is an imaging device, 17 is an image feature amount extraction unit, and 18 is FUZZY
A wrinkle state determination unit 19 using an inference mechanism is an image processing device including an image feature amount extraction unit 17 and a wrinkle state determination unit 18.

FIG. 3 is a diagram showing the configuration of the image feature quantity extraction unit 17 in FIG. In the figure, reference numeral 20 denotes an image binarizing unit, 21 denotes an image data compressing unit, 22 denotes an image data storing unit, 23 denotes an inspection line extracting unit, 24 denotes a pattern pitch calculating unit, 25 denotes a pattern pitch storing unit, and 26 denotes abnormalities. A pattern pitch extraction unit, 27 is an abnormal pattern pitch storage unit, and 28 is an image feature amount calculation unit.

FIG. 4 is a diagram showing a data structure compressed by the image data compression unit 21. FIG. 4 (a) shows binary image data stored in the image memory, and FIG. 4 (b) shows compressed data stored in the image memory.

FIG. 5 is a diagram for explaining how the wrinkle characteristic amount is extracted by the tape wrinkle inspection apparatus shown in FIG. In the drawing, reference numeral 29 denotes an inspection line, and reference numeral 30 denotes a stripe pattern pitch, which is the distance between the centers of the stripe patterns on the inspection line 29. 31 is an abnormal pattern pitch, and 32 and 33 are end points of the abnormal pattern pitch.

FIG. 6 is a diagram showing the configuration of the wrinkle state determination unit 18 in FIG. 3, reference numeral 34 denotes an image feature information transmission path calculated by the image feature extractor 17 in FIG. Reference numeral 35 denotes an inference rule calculation unit corresponding to each wrinkle state taught in advance. Reference numeral 36 denotes a storage unit for the membership function corresponding to the inference rule operation unit 35.

Here, the membership function is an important concept in FUZZY theory, and expresses the ambiguity of meaning such as "high" and "elderly" in terms of quantity. Height 140c
In the range from m to 200 cm, the degree that a person of x (cm) can be said to be high is μ, and the height x is made to correspond to the degree μ (where 0 ≦ μ ≦ 1). If x is plotted on the horizontal axis and μ is plotted on the vertical axis, a graph as shown in FIG. 14 can be drawn. This graph is a quantitative representation of the ambiguity of "high". Similarly, the ambiguity of the concept of "old man" can be represented by the graph in FIG. The abscissa indicates the age and the ordinate indicates the degree.

As shown in the graphs of FIGS. 14 and 15, the function that gives the degree of the meaning of a word is called a membership function in FUZZY theory. The degree μ may be called a membership value.

37 is an inference rule operation unit 35, a membership function storage unit 36
Is a rule matching degree calculation unit composed of 38 is a membership function teaching unit. Reference numeral 39 denotes a transmission line for the degree of adaptation of the inference rule calculated by the inference rule operation unit 35.
Reference numeral 40 denotes an optimal rule determining unit. 41 is the optimal rule decision unit 40
Is the wrinkle state information transmission path determined by

FIG. 7 is a diagram showing how the degree of conformity of the inference rule is calculated in the inference rule calculation unit 35 of FIG. In the figure, reference numeral 42 denotes a membership function corresponding to an image feature amount. Reference numeral 43 denotes a transmission path for information on the membership function conformance indicating the degree to which the image feature quantity conforms to the membership function. Reference numeral 44 denotes a rule conformance calculator for determining the conformity of the inference rule from the membership function conformance.

FIG. 8 is a view for explaining how the membership function teaching section 38 of FIG. 6 teaches the membership function. In the figure, the same components as those in FIG. 1 are denoted by the same reference numerals, and description thereof is omitted. In the figure, reference numeral 45 denotes an image feature amount histogram creation unit. 46 is a membership function creation unit. 47 is a TV monitor. Reference numeral 48 denotes a keyboard attached to the image processing device 19. 49 is a human teaching.

FIG. 9 is a diagram showing how the image feature amount histogram and the membership function are created in the image feature amount histogram creation unit 45 and the membership function creation unit 46 of FIG. FIG. 9A shows an image feature amount histogram, and FIG. 9B shows a membership function.

FIG. 10 is a diagram showing a configuration example using a microcomputer of the image processing device 19 in FIG. In FIG. 10, 50 is a microcomputer, 51 is a microcomputer memory, 52 is an image memory, and 53 is an image calculation device.

The operation of the tape wrinkle inspection apparatus configured as described above will be described below with reference to the drawings.

First, an outline of the operation will be described. The projector 13 irradiates the reflector 14 with the stripe pattern printed thereon with strong light, reflects the light, and projects the stripe pattern of the reflector 14 onto the tape 15. The projected light is reflected by the tape 15 and taken into the imaging device 16. Then, the wrinkle detection of the tape is performed by the image processing device 19 including the image feature amount extraction unit 17 and the wrinkle state determination unit 18, and the wrinkle state is identified and the wrinkle inspection is performed.

Next, the operation of the image feature amount extraction unit 17 will be described.
The image captured by the imaging device 16 is binarized and compressed by the image calculation device 53 shown in FIG. This means that the captured image is stored in the image data storage unit 22 which has been binarized and compressed by the image binarization unit 20 and the image data compression unit 21 shown in FIG. Next, a pattern pitch 30 on the inspection line 29 shown in FIG. 5 is calculated from the image data stored in the image
Store in 51. This is based on the data stored in the image data storage unit 22 and the pattern pitch on the inspection line 39 using the inspection line extraction unit 23 and the pattern pitch calculation unit 24.
This corresponds to calculating 30 and storing it in the pattern pitch storage unit 25. Here, since the compressed image data stored in the image memory 52 is compressed in the form as shown in FIG. 4 (b), the calculation cost of the pattern pitch calculation unit 24 is small.
High-speed processing becomes possible. Then, the abnormal pattern pitch is calculated from the pattern pitch 30 stored in the microcomputer memory 51.
31 is extracted using the microcomputer 50 and stored in the microcomputer memory 51. This is based on the data stored in the pattern pitch storage unit 25 and the abnormal pattern pitch extraction unit.
This is equivalent to extracting the abnormal pattern pitch 31 using 26 and storing it in the abnormal pattern pitch storage unit 27. here,
The abnormal pattern pitch 31 is a pattern pitch 30 whose value is smaller or larger than a certain reference value. Then, the microcomputer 50 calculates the image feature amount from the abnormal pattern pitch 31. This is the abnormal pattern pitch stored in the abnormal pattern pitch storage unit 27.
This corresponds to calculating an image feature amount from 31 using the image feature amount calculation unit 28. In this embodiment, the abnormal pattern pitch
Number of 31, abnormal pattern pitch end points 32, 33, inspection line 2
The coordinates of the center of gravity of the histogram with respect to the center coordinates of the abnormal pattern pitch 31 on the coordinate axis parallel to 9 and the number of portions where the histogram value is 0 are defined as image feature amounts. of course,
Depending on the state of the wrinkles that occur, it may be better to use something other than the feature amounts listed here. Further, in the present embodiment, the pattern pitch 30 is calculated by the microcomputer 50 from the compressed image data and stored in the microcomputer memory 51, but the pattern pitch 30 is calculated by the image calculation device 53 and stored in the image memory 52. Is also good. In this case, the processing speed is further improved. However, in the method of calculating the pattern pitch 30 by the microcomputer 50 from the compressed image data and storing the pattern pitch in the microcomputer memory 51, the center of the pattern is obtained with an accuracy of 1/2 pixel, In the method of calculating the pitch 30 and storing it in the image memory 52, the accuracy of the pattern pitch 30 is reduced because the accuracy of the center of the pattern is one pixel.

Next, the operation of the wrinkle state determination unit (FUZZY inference unit) 18 will be described. These operations are performed by the microcomputer 50
By using the microcomputer memory 51. The image feature amounts calculated by the image feature amount extraction unit 17 are respectively input to a plurality of inference rule calculation units 35 corresponding to the wrinkled state taught in advance. There is an inference rule in a one-to-one relationship with respect to a wrinkle state pattern to be detected, and there are as many inference rule calculation units 35 as inference rules. The inference rule is specified by a membership function 42 shown in FIG. 7 corresponding to the image feature amount. That is, teaching the wrinkle state in advance means teaching the shape of the membership function 42 corresponding to the image feature amount. The membership function 42 is stored in the membership function storage unit 36 in advance by the membership function teaching unit 38. The teaching of the membership function 42 is as follows. First, as shown in FIG.
The image captured by step 16 is displayed on the TV monitor 47.
The human 49 looks at the wrinkle state displayed on the TV monitor 47,
It is determined which of the wrinkle states classified in advance corresponds to the wrinkle state, and is input to the image processing apparatus 19 using the keyboard 48. On the other hand, inside the image processing device 19, the TV
The image feature amount is calculated for the wrinkle displayed on the monitor 47 using the image feature amount extraction unit 17 in advance. In the image feature amount histogram creation unit 45, the image feature amount
An image feature amount histogram corresponding to the wrinkle state inference rule input in step 48 is created. By executing the teaching a plurality of times, an appropriate image feature amount histogram is created. Then, the membership function creating unit 46 smoothes and normalizes the image feature amount histogram as shown in FIG. 9A, and converts the trapezoidal membership function as shown in FIG. 9A. Create automatically. The created membership function 42 is stored in the membership function storage unit 36. By using the membership function 42 created in this manner, wrinkles can be almost identified.However, in the case where more delicate identification is required or unexpected wrinkles appear, In the embodiment, a mechanism is also provided that enables a human 49 to modify the membership function 42 into an arbitrary shape using the keyboard 48. As described above, in the present embodiment, it is possible to teach the membership function 42 to an appropriate shape for a wrinkled state arbitrarily classified based on human senses. Now, as shown in FIG. 7, the inference rule calculation unit 35 refers to the image feature amount to the membership function 42 corresponding to the image feature amount, and determines how much the image feature amount matches the membership function 42. For each image feature amount, a membership function suitability value indicating whether or not the image feature amount is obtained is input to the rule suitability calculation unit 44 to output the suitability of the inference rule for the image feature amount. In FIG. 7, the membership function conformances correspond to g ₀ , g ₁ , and g ₂ . Here, the rule conformance calculating unit 44 performs an operation to set the minimum value among the membership function conformances to the inference rule conformity 39 in the present embodiment. In the FIG. 7, the fitness of the inference rule corresponding to g _1. Further, the degree of conformity of the inference rule indicates to what extent the image feature amount conforms to the wrinkled state taught.
In this way, the fitness of each inference rule obtained by the rule fitness calculating unit 37 is input to the optimal rule (wrinkle state) determining unit 40, and the wrinkle state 41 is determined by selecting the optimal rule. You. Here, the optimal rule determination unit 40 considers that the taught wrinkle state has the same weight in the present embodiment, and determines the inference rule having the maximum value of the fitness level among the fitness levels of the inference rules. Is performed. By the way, in the present embodiment, there is an inference rule in a one-to-one relationship for a wrinkle state pattern to be detected, and the shape of the membership function 42 corresponding to the image feature amount is taught. May be fixed, and a plurality of inference rules may be set for a wrinkle state pattern to be detected by teaching. As described above, these operations are performed by the microcomputer 50 using the microcomputer memory 51. In order to reduce the calculation cost of the microcomputer 50 and the storage capacity of the microcomputer memory 51, in the present embodiment, the membership function 42 The image feature amount as input and the fitness as output are represented by 10 discrete integers, and the membership function 42 itself is also 10
Expressed as a set of × 10 integer arrays. For this reason, the storage capacity of the microcomputer memory 51 can be saved, and high-speed calculations of several ms or less can be performed because data comparison is mainly performed in calculations.

As described above, by projecting the reflected light of the striped pattern onto the tape, it is possible to obtain an image in which the boundary between the patterns is clear and the pattern is clear, and therefore, when detecting wrinkles using an image processing apparatus. The wrinkle state can be detected with a simple algorithm. Further, it is not necessary to bring the striped pattern into close contact with the tape, and the positioning is simplified. In addition, by analyzing the pitch of the striped pattern projected on the tape on multiple inspection lines and extracting multiple feature amounts of the striped pattern, based on the necessary and sufficiently compressed wrinkle information Wrinkles can be detected, and a change state of minute wrinkles can be detected in a short time. In addition, by identifying the wrinkle state of the striped pattern using the FUZZY inference mechanism based on information of a plurality of feature amounts, a plurality of feature amounts can be appropriately matched in a form that is easy to realize by a computer, thereby making it complicated and minute. The wrinkle state change can be identified at high speed, and the wrinkle state can be inspected at high speed in association with human judgment.

As described above, the effect of the tape wrinkle inspection method according to the present embodiment can be exerted even when the operator directly visually inspects and identifies the state of the striped pattern projected on the tape.

FIG. 2 is a diagram showing an outline of a second embodiment of the present invention. In the figure, the same items as those in FIG.
Description is omitted. In the figure, reference numeral 64 denotes an actuator control device, and reference numeral 65 denotes an adjusting actuator for adjusting wrinkles of the tape.

FIG. 11 is a diagram showing the vicinity of the loading post of the video tape. In the figure, 54 is a loading post main body, 55 is a loading post part, 56 is a base part, 57 is a wrinkle adjusting post part, 58 is a pace fixing screw, and 59 is a stopper.

The operation of the tape wrinkle adjusting device configured as described above and using the tape wrinkle inspection device shown in the first embodiment will be described below with reference to the drawings.

First, an outline of the operation will be described. As in the first embodiment, strong light from the light projector 13 is reflected by the reflector 14 on which the stripe pattern is printed, and the stripe pattern of the reflector 14 is projected onto the tape 15. The projected light is reflected by the tape 15 and taken into the imaging device 16. And an image feature amount extraction unit
The wrinkle detection of the tape is performed by the image processing device 19 including the wrinkle state determination unit 18 and the wrinkle state determination unit 18, and the wrinkle inspection is performed by identifying the wrinkle state. If the wrinkles on the tape are not in an appropriate state as a result of the inspection, the actuator control device 64 moves the adjusting actuator 65 to adjust the wrinkles. Thereafter, the inspection adjustment is repeated until the wrinkles are in an appropriate state.

Next, operations of the actuator control device 64 and the adjustment actuator 65 will be described. In FIG. 11, the loading post body 54 is pressed against a stopper 59, and a base portion 56 is fixed by a solid base screw 58. Therefore, by loosening the base fixing screw 58, the base 56
It can rotate left and right around the point of contact with 54. Further, since the wrinkle adjusting post 57 is fixed obliquely to the base 56, when the base 56 is rotated, the angle of the tape 15 in contact with the wrinkle adjusting post 57 changes, and the loading post 55 The state of the wrinkles of the nearby tape 15 changes. In other words, in order to adjust the state of wrinkles, the base fixing screw 58 may be loosened by rotating the base fixing screw 58, and the screw may be tightened when the adjustment is completed. Actuator control device 64 and adjustment actuator in this embodiment
65 has such a function. As for the wrinkles near the loading post of the video tape to be controlled in the present embodiment, there is a correlation between the rotation direction of the base portion 56 and the state of occurrence of wrinkles. Provide a table of the adjustment rotation amount for the state and set the base part
An adjustment mechanism for rotating 56 is provided. Further, depending on the video deck, a difference appears in the adjustment rotation amount even if the wrinkle state is changed in the same manner. For this reason, depending on the video deck, even if a standard adjustment amount is applied, the number of adjustments increases, and the time required for the adjustment may take too long. In the tape wrinkle adjustment inspection apparatus according to the present embodiment, it is possible to know how much the rotation amount of the base portion 56 is too large or too small by comparing the wrinkle state before and after the adjustment. The actuator control device 65 of the present embodiment has a mechanism for applying an appropriate adjustment amount to the video deck based on such information.

As described above, by adjusting the wrinkles of the tape based on the subtle wrinkle change information detected appropriately, the wrinkles of the tape can be adjusted quickly and appropriately in accordance with the adjustment characteristics of the video deck.

INDUSTRIAL APPLICABILITY As described above, according to the present invention, reflected light of a striped pattern is projected onto a running tape, and the pitch of the striped pattern projected on the tape is set on a plurality of scanning lines. By analyzing and extracting a plurality of feature amounts of the striped pattern, it is possible to obtain an image with good contrast and clear pattern boundaries, and therefore, when detecting wrinkles using an image processing device. The wrinkle state can be detected with a simple algorithm. Further, it is not necessary to bring the striped pattern into close contact with the tape, and the positioning is simplified. Furthermore, it is possible to detect wrinkles based on necessary and sufficiently compressed wrinkle information, and to change the minute wrinkle change state as the dynamic property of the running tape to its position and degree in a short time. Can be detected.

In addition, since the state of wrinkles in the striped pattern is identified using the FUZZY inference mechanism based on information on a plurality of feature values, multiple feature values are appropriately matched in a form that is easy to represent by a computer, and there is no complicated and minute amount. The wrinkle state change can be identified at high speed, and the wrinkle state can be inspected at high speed in association with human judgment.

Also, means for projecting the striped pattern on the running tape by reflecting the striped pattern by the projector, and calculating a plurality of feature amounts of the striped pattern by calculating the pitch of the striped pattern projected on the tape on a plurality of inspection lines. Since it has a means for extracting and a means for identifying the state of wrinkles of the striped pattern using the FUZZY inference mechanism based on information of a plurality of feature amounts, the positioning of the striped pattern is simple, and the contrast is good. It is possible to obtain an image with clear pattern boundaries, and to detect the change in the minute wrinkles as the dynamic properties of the running tape to its position and degree in a short time. By appropriately matching the characteristic amounts of the wrinkles, a plurality of changes in the state of fine wrinkles can be identified at high speed, and the state of wrinkles can be inspected at high speed in association with human judgment.

Further, using the above-described configuration, it is provided with an adjustment actuator for adjusting the mounting position of the member supporting the running of the tape based on the information of the detected wrinkle state, and is configured to adjust the wrinkle state of the tape, In accordance with the adjustment characteristics of the video deck, the adjustment can be performed until the wrinkles are in a desired state, the assembly can be completed, the adjustment can be performed quickly and appropriately, and a series of steps can be automatically performed.

Continuation of the front page (56) References JP-A-60-61936 (JP, A) JP-A-62-28042 (JP, A) JP-A-59-202045 (JP, A) JP-A-61-264412 (JP) JP-A-60-218291 (JP, A) JP-A-2-658 (JP, U) (58) Fields investigated (Int. Cl. ⁶ , DB name) G01N 21/84-21/90 G11B 15/18 G11B 15/43

Claims (5)

(57) [Claims] 1. A method of projecting reflected light of a striped pattern on a running tape, and calculating a pitch of the striped pattern projected on the tape on a plurality of inspection lines orthogonal to a running direction of the tape. From the calculated pattern pitches, a histogram relating to the number of abnormal pattern pitches having a value smaller or larger than the reference value is created on a coordinate axis parallel to the inspection line to extract a plurality of feature amounts of the striped pattern. A tape wrinkle inspection method for inspecting the wrinkle state of the tape. 2. The tape according to claim 1, wherein the tape is inspected by identifying a wrinkle state of the tape by using a FUZZY inference mechanism based on information of a plurality of feature amounts of the striped pattern. Wrinkle inspection method. 3. A means for projecting reflected light of a striped pattern onto a running tape, an imaging device for capturing an image of the striped pattern projected on the tape, and an image 2 for binarizing the captured image. A binarizing unit, an image data storage unit for storing binarized image data, and a running speed of the tape, based on data stored in the image data storage unit. A pattern pitch calculation unit that calculates each on a plurality of inspection lines orthogonal to the direction, a pattern pitch storage unit that stores data of the calculated pattern pitch, and a reference value from data stored in the pattern pitch storage unit. An abnormal pattern extractor that extracts a small or large abnormal pattern pitch, and an abnormal pattern pitch that stores the extracted abnormal pattern pitch A storage unit, an image feature amount calculation unit that creates a histogram regarding the number of abnormal pattern pitch data stored in the abnormal pattern pitch storage unit on a coordinate axis parallel to the inspection line, and the image feature amount calculation unit. A plurality of inference rule calculation units for inputting the calculated image feature amount and calculating the fitness of the inference rule corresponding to each wrinkle state taught in advance, and the fitness of each inference rule are input, and the optimal rule is input. And a wrinkle state determination unit that determines a wrinkle state by selecting a wrinkle state. 4. An inference rule calculation unit and a wrinkle state determination unit are FUZZ
4. The tape wrinkle inspection apparatus according to claim 3, wherein the apparatus is configured by a Y inference mechanism. 5. A tape wrinkle inspection device according to claim 3 or 4, wherein the tape wrinkle inspection device outputs the tape wrinkle state information output from a wrinkle state determination unit of the tape wrinkle inspection device. A tape wrinkle adjustment device using a tape wrinkle inspection device, comprising an adjustment actuator for adjusting a mounting position of a member supporting a tape running, and configured to adjust a tape wrinkle state.