JPH07248214A - Method for correcting measuring timing, apparatus for correcting measuring timing and automatic appearance-inspection apparatus - Google Patents

Abstract

PURPOSE:To obtain a measuring timing enabling the judgment of an appearance of a to-be- measured object on a continuously moving conveyor, by performing an analysis of variance by a first-dimensional arrangement of a plurality of characteristic amounts of the object at every point where the measurement is supposed to be started. CONSTITUTION:A CCD camera 3 is arranged immediately above a to-be-measured object on a conveyor. Two inverter fluorescent lamps 31, 32 for image processing which are sufficiently longer than the to-be-measured object are set at both sides of the CCD camera 3 in parallel to the conveyor. An image-taking device 41 in an image-processing device 4 takes an image photographed by the camera 3, and a converting device 42 converts the image to a binary image. A characteristic amount-operating device 43 operates a characteristic amount of the binary image. An analyzing device 44 carries out an analysis of variance of a first-dimensional arrangement of the operated characteristic amounts. A determining device 45 determines a measurement-starting point based on the result of the analysis of variance. As a result of the analysis of variance of the first-dimensional arrangement, a point where the measurement is supposed to be started and showing not smaller than a predetermined reliability estimation value is determined as the measurement-starting point to set a starting switch.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a measurement timing correction method, a measurement timing correction device and a measurement timing correction device for determining a measurement start position which enables an automatic visual inspection of a similar measurement object flowing by a conveyor. The present invention relates to an automatic visual inspection apparatus in which measurement is started by an imaging camera when a measurement start switch set to is turned on.

[0002]

2. Description of the Related Art In an automatic appearance measuring apparatus using a conventional binarized image processing technique, in order to detect a plurality of meta-reflective parts, a camera and an object to be measured as shown in FIGS. Since the image is different even for the same measurement target due to the difference in the distance from the measurement target, the measurement target was stopped for measurement.

[0003]

The automatic appearance measuring apparatus utilizing the above-mentioned conventional binarized image processing technology can measure the presence or absence of the measurement object on the continuously moving conveyor, but the continuously moving conveyor. It was difficult to judge the appearance of the above measurement target and the appearance of similar parts with similar appearance.

Therefore, the present inventor stops the measurement object in order to determine the appearance of the measurement object on the continuously moving conveyor by the automatic appearance measurement using the binarized image processing technique. The measurement timing (measurement start point) that enables measurement under the same conditions as those of No. 1 is determined by the analysis of variance by the one-dimensional arrangement of the feature quantities of the plurality of measurement targets at each measurement start scheduled point. Focusing on the technical idea, when the measurement target reaches the determined measurement start point, measurement is started, and a plurality of similarities obtained experimentally with the feature amount of the obtained binarized image in a still state in advance. Further, further research and development was carried out, focusing on the second technical idea of the present invention of determining the degree of conformity with the reference value corresponding to the feature amount obtained for the measurement target and determining the similar measurement target. As a result, binary image processing Operative with obtaining the measurement timing which allows the appearance determination of the measurement object on the conveyor which moves continuously by automatic visual measurement utilizing, have reached the present invention to achieve the object of realizing an appearance judgment.

[0005]

The measurement timing correction method of the present invention (the first invention according to claim 1) comprises a setting step of setting a plurality of measurement start points, and each set measurement start point. A measurement step of measuring a plurality of measurement objects flowing on the conveyor based on the camera, a conversion step of converting captured images of the plurality of measurement objects measured by the camera into a binary image, Analysis of variance by performing a feature amount calculation step of calculating the feature amounts of the plurality of measurement targets from a binarized image, and performing an analysis of variance by unifying arrangement of the feature amounts of the plurality of measurement targets for each of the obtained measurement start scheduled points. And a measurement start point determining step of determining a measurement start scheduled point equal to or higher than a certain reliability estimation value as a result of the one-way analysis of variance as a measurement start point for setting a measurement start switch. Are those that permit the determination by the feature of the measurement target similar flowing continuously by A.

In the measurement timing correction method of the present invention (the second invention according to claim 2), in the first invention, the converting step removes white or black isolated points in the converted binary image. It includes a step of performing.

The measurement timing correction method of the present invention (the third invention according to claim 3) is the method according to the first invention, wherein the variance analysis step calculates the characteristic quantities of the plurality of measurement targets at each of the measurement start scheduled points. The process includes a step of storing the measurement position and the number of times of measurement in a unified arrangement table.

The measurement timing correction apparatus of the present invention (the fourth invention according to claim 4) is configured to set a plurality of measurement start scheduled points and flow on a conveyor based on each set measurement start scheduled point. A camera that measures a plurality of measurement targets, a conversion unit that converts captured images of the plurality of measurement targets measured by the camera into a binarized image, and the plurality of measurement targets from the converted binarized image Amount calculating means for calculating the characteristic amount, an analysis of variance by means of a one-dimensional arrangement of the obtained characteristic amounts of the plurality of measurement targets for each of the obtained measurement start scheduled points, and an analysis of variance by the one-way arrangement As a result, a measurement start point determining means for determining the measurement start scheduled point that is equal to or higher than a certain reliability estimated value as a measurement start point for setting the measurement start switch, and a type that flows continuously by a conveyor. It is those that permit the determination by the feature of the measurement target to be.

An automatic visual inspection apparatus according to the present invention (a fifth invention according to claim 5) is a measurement start switch set at the measurement start point determined by the first invention and the fourth invention,
An imaging camera that measures the measurement target when the set measurement start switch is turned on by the measurement target flowing through the conveyor, and a captured image of the measurement target captured by the imaging camera is converted into a binary image. Conversion means, feature quantity calculation means for calculating feature quantities of the plurality of measurement targets from the binarized image converted by the conversion means, and feature quantity calculated by the feature quantity calculation means and types of measurement targets It comprises a determination means for determining the degree of conformity with the corresponding reference value, and a display means for displaying the obtained feature amount and the determination result.

The automatic visual inspection apparatus according to the present invention (the sixth invention according to claim 6) is the automatic appearance inspection apparatus according to the fifth invention, wherein the imaging camera is arranged in parallel with the flow direction of the conveyor while covering the measurement visual field. It is arranged outside the direct ray region between the linear light sources.

The automatic appearance inspection apparatus according to the present invention (the seventh invention according to claim 7) is the automatic appearance inspection apparatus according to the fifth invention, wherein the feature amount computing means has an area to be measured, a perimeter, a number of holes, a hole area, and At least one of the angles is calculated.

[0012]

According to the measurement timing correction method of the first invention having the above structure, a plurality of measurement start points are set, and a plurality of measurement targets flowing through the conveyor are measured by the camera based on the set measurement start points. Then, the captured images of the plurality of measurement targets measured by the camera are converted into a binarized image, the characteristic amounts of the plurality of measurement targets are calculated from the converted binarized image, and the obtained respective measurements are performed. An analysis of variance is performed by a one-way arrangement of the feature amounts of the plurality of measurement targets for each scheduled start point, and the result of the one-way arrangement analysis of variance is set to a measurement start scheduled point that is equal to or higher than a certain reliability estimation value. It determines the starting point.

The measurement timing correction method according to the second aspect of the present invention, which has the above-described structure, adds to the first aspect the action of removing white or black isolated points in the converted binarized image. .

According to the measurement timing correction method of the third invention having the above-mentioned configuration, the feature quantities of the plurality of measurement objects at the respective measurement start points are stored in a unified arrangement table of the measurement position and the number of times of measurement, in the first invention. The action of storing is added.

In the measurement timing correction apparatus of the fourth invention having the above-mentioned structure, the setting means sets a plurality of measurement start points, and the plurality of cameras flow on the conveyor based on the set measurement start points. Of the measurement object, the conversion means converts the captured images of the plurality of measurement objects measured by the camera into a binarized image, and the feature amount calculation means converts the binarized image into the binary image. The feature amount of a plurality of measurement targets is calculated, and the variance analysis unit performs a variance analysis based on the unitary arrangement of the feature amounts of the plurality of measurement targets for each of the determined measurement start scheduled points, and the measurement start point determination unit. As a result of the analysis of variance by the one-way arrangement, the planned measurement start point that is equal to or higher than a certain reliability estimation value is determined as the measurement start point at which the measurement start switch is set.

In the automatic visual inspection apparatus according to the fifth aspect of the present invention having the above-described structure, the measurement start switch set at the measurement start point determined by the first and fourth aspects is turned on by the measurement target flowing through the conveyor. Then, the imaging camera measures the measurement target, the conversion unit converts the captured image of the measurement target captured by the imaging camera into a binarized image, and the feature amount calculation unit causes the conversion unit to convert the captured image. The feature amounts of the plurality of measurement targets are calculated from the converted binarized image, and the determination unit determines the compatibility between the feature amount calculated by the feature amount calculation unit and the reference value according to the type of the measurement target. It is determined, and the display means displays the obtained feature amount and the determination result.

In addition to the function of the fifth aspect of the invention, the automatic visual inspection apparatus of the sixth aspect of the invention having the above-mentioned configuration uses the linear light source arranged in parallel with the direction of flow of the conveyor to cover the measurement visual field. The light is evenly emitted to the object, and the measurement target on the conveyor is measured by the imaging camera arranged outside the direct light ray region between the linear light sources.

In addition to the operation of the fifth aspect of the invention, the automatic appearance inspection apparatus of the seventh aspect of the invention having the above-mentioned configuration has at least the area of measurement target, the peripheral length, the number of holes, the hole area, and the angle. This has the effect of computing one or more.

[0019]

According to the measurement timing correction method of the first aspect of the present invention, which has the above-described operation, the measurement start scheduled point that is equal to or higher than a certain reliability estimated value as a result of the one-way analysis of variance is set as the measurement start point at which the measurement start switch is set. Since it is determined, it is possible to obtain the measurement timing that enables the appearance determination of the measurement target on the conveyor that moves continuously by the automatic appearance measurement using the binarized image processing technology, and to realize the appearance determination. Play.

In addition to the effect of the first aspect of the invention, the measurement timing correction method of the second aspect of the present invention, which has the above-described operation, removes white or black isolated points in the converted binarized image, thus eliminating noise. There is an effect that can be.

In addition to the effect of the first aspect of the invention, the measurement timing correction method of the third aspect of the present invention having the above-described operation has a feature that the characteristic quantities of the plurality of measurement targets at each of the measurement start points are uniformly arranged at the measurement position and the number of measurements. Since the data is stored in the table, there is an effect of facilitating the above-mentioned one-way analysis of variance.

The measurement timing correction apparatus of the fourth invention having the above-mentioned operation determines the measurement start point which is equal to or higher than a certain reliability estimation value as a result of the one-way analysis of variance as the measurement start point for setting the measurement start switch. Therefore, it is possible to obtain the measurement timing that enables the appearance determination of the measurement target on the continuously moving conveyor by the automatic appearance measurement using the binarized image processing technology, and it is possible to realize the appearance determination. .

In the automatic visual inspection apparatus according to the fifth aspect of the present invention having the above-mentioned operation, the judging means judges the degree of conformity between the characteristic amount calculated by the characteristic amount calculating means and the reference value according to the kind of the measuring object. Thus, it is possible to distinguish measurement targets having similar appearances.

In addition to the effect of the fifth aspect of the invention, the automatic visual inspection apparatus of the sixth aspect of the invention having the above-described operation is characterized in that the object to be measured is covered by a linear light source which is arranged in parallel with the conveyor flow direction so as to cover the measurement visual field. With the effect of being able to uniformly irradiate a light ray with respect to, because the measurement target on the conveyor is measured by the imaging camera arranged outside the direct ray region between the linear light sources. This has the effect of realizing highly accurate measurement.

In addition to the effect of the fifth aspect of the invention, the automatic appearance inspection apparatus of the seventh aspect of the invention having the above-mentioned operation has at least the area of measurement target, the peripheral length, the number of holes, the hole area, and the angle by the feature quantity computing means. Since one or more pieces are calculated and the appearance determination of the measurement targets having similar appearances is performed based on this feature amount, there is an effect of realizing highly reliable appearance determination.

[0026]

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

(First Embodiment) A measurement timing correction method and a measurement timing correction device according to the first embodiment move on a limit switch 1 arranged at a measurement start scheduled point and a conveyor as shown in FIG. CC that measures the measurement target
D camera 3, sequence controller 2 connected to the limit switch 1, CCD camera 3 and a feature amount of a binary image obtained by converting a captured image measured by the CCD camera 3 connected to the sequence controller 2 The image processing apparatus 4 determines the measurement start point by performing the analysis of variance of the one-way arrangement, and the display unit 5 including the TV monitor 51 and the CRT terminal 52 for displaying the image processing result.

As shown in FIG. 1, the limit switch 1 is composed of limit switches 11 to 1m arranged at respective positions of x1 to xm which can be the starting point of measurement of a stationary portion facing the conveyor.

The CCD camera 3 is arranged directly above the object to be measured on the conveyor, and both sides of the CCD camera 3 are image processing impertainer fluorescent lamps 31 and 3 which are sufficiently longer than the object to be measured.
Two 2 are arranged in parallel with the conveyor.

As shown in FIG. 1, the image processing device 4 includes an image capturing device 41 for capturing the captured image measured by the CCD camera 3, and a conversion device 42 for converting the captured image into a binary image. A feature amount computing device 43 that computes a feature amount of a binarized image, an analysis of variance device 44 that performs a one-way analysis of variance of the computed feature amount, and a measurement that determines a measurement start point based on the result of the variance analysis. Starting point determination device 45
It consists of and.

Prior to the measurement of the first embodiment, measurement is performed by a CCD camera placed right above an object to be measured, which is stationary experimentally, and image processing similar to that described above is performed to calculate a feature amount. That is, as shown in FIG. 2A, for example, in the case of a measurement target on a disk, measurement is performed for a total of four cases, that is, the case where the CCD camera matches the core of the measurement target and the case where the CCD camera is displaced from the center. Is performed, (b), (c),
As shown in (d) and (e), the shadow patterns are different,
The digitized image also differs accordingly. Therefore, the characteristic amount of the binarized image when the CCD camera and the core of the measurement target are coincident with each other is previously obtained experimentally in advance.

The flow chart shown in FIG. 3 which is stored in the ROM of the memory of the image processing apparatus 4 in the measurement timing correction method and the measurement timing correction apparatus of the first embodiment will be described.

That is, initialization is performed in step 101, and various parameters such as the actual measurement environment such as the moving speed of the conveyor and the reflection state of the illumination light are set by the manual measurement in step 103 prior to the automatic measurement in step 102, and the above-mentioned experiment is performed. The characteristic amount of the measurement target obtained in advance is set in the memory, and then the unitary initial setting is performed in step 104, the unitary arrangement parameter is set in step 105, and the CCD camera 3 measures in step 106. Then, the feature amount is extracted, the unification is performed in steps 107 and 108, the one-way arrangement is calculated and determined, and the one-way arrangement analysis of variance is performed.

This one-way ANOVA will be described in more detail with reference to FIG. 4. Initialization is performed in step 201, and m, which is the number of planned measurement start points x1 to xm, and the measurement object measured by the CCD camera 3. Enter n, which is the number of

In step 203, an instruction to install the limit switch at the scheduled measurement start point is displayed on the CRT terminal 52, and the limit switch is installed according to the instruction.

Next, in step 204, the above-mentioned CC
The captured image measured by the D camera 3 is captured by the image capturing device 41, the captured image is converted into a binarized image by the conversion device 42, and the feature amount calculation device 43 is used.
The feature amount of the binarized image converted by is calculated, and in steps 205 and 206, the n feature amount of the measurement target at each of the m planned measurement start points is collected.

Next, in step 207, the 95% confidence estimated value bar Mi ± Δ of the feature amount by the one-way ANOVA of the collected feature amount is calculated based on the equations 1 to 4.

[Equation 1]

[Equation 2]

[Equation 3]

[Equation 4]

Next, in steps 208 to 210, the average value bar Mi of the feature amount at each measurement position is stored in the memory with respect to the reference value Mm which is the above-obtained experimentally obtained feature amount of the measurement object. It is judged whether the value is within the allowable value Δ or not, and if it is within the allowable value Δ, the measurement start point where the limit switch 1 can be set is determined.

The operation procedure and action of the measurement timing correction method and the measurement timing correction apparatus of the first embodiment having the above-mentioned configuration will be described with reference to FIGS. 5 and 6.

In step 301, it is checked whether or not the one-way analysis of variance is being processed. If not, in step 302, a guidance message is sent to the CRT terminal 52 to set the number of times of measurement n and the number of measurement start positions m. Display it.

In steps 303 and 304, it is checked whether or not the input values of the number of times of measurement and the number of measurement start positions are appropriate, and if they are appropriate, it is checked in step 305 whether or not measurement is being performed at the measurement position xi.

When the measurement is not in progress, a guidance message "Install the limit switch 1 to the measurement position xi or put the limit switch at the measurement position xi into the operating state" is displayed in step 306.

In step 307, it is checked whether or not the measurement start limit switch is on, and if it is on,
In step 308, the memory area for capturing the image is erased.

In step 309, the image capturing device 41 captures a captured image of a reference portion from the CCD camera 3, and in step 310, the conversion device 42 converts the captured image into 2 based on a binarization threshold value. Convert to a digitized image.

The converted 2 in step 311
Black and white isolated points of the binarized image are removed, and step 312
At, the window area is set and measurement is performed.

In step 313, the feature quantity computing device 33 extracts the feature quantity such as the area value and the perimeter of the binarized image, and in step 314, the feature quantity is stored in the unitary arrangement table as shown in Table 1. To do. In Table 1, Mij indicates the feature amount obtained by the j-th measurement of the measurement position xi. i = 0 to m, j = 1 to n

[Table 1]

In step 315, it is checked whether or not the measurement at the measurement position xi is the nth time, and when it is the nth time, the counter of the number of times of measurement is initialized in step 316, and the count of the measurement position is incremented in step 317. .

If it is not the nth time in step 318, the count of the number of times of measurement is incremented, and step 3
It is checked in 19 whether or not the collection of feature quantities is completed, and if completed, in step 320, the one-way ANOVA is performed by the above-mentioned ANOVA 44 according to Eqs. 1 and 2.

In step 321, it is checked whether or not there is a difference in the measurement positions x0 to xm. If the relationship of expression 3 is satisfied and there is no difference, then in step 322 expression 4 is obtained.
Calculate the 95% confidence estimate at each measurement location according to
In the above equation 4, t (Φe, 0.05) is a value with a risk rate of 5% in the t distribution.

If the measurement positions x0 to xm are different from each other without satisfying the relation of Expression 3, it is displayed on the CRT terminal 52 that the measurement at the measurement positions x0 to xm has a variation (danger rate 5%) in step 324. To do.

In step 323, as shown in FIG. 7 which shows the relationship between the measurement position and the feature quantity, 9
It is checked whether or not the difference is within the allowable value Δ corresponding to the 5% reliability estimated value, that is, whether the difference between the average bar Mi of the feature amount at each measurement position and the reference value Mm is within the range of the variation Δ. It is checked whether or not there is a corresponding measurement position in step 325. If there is a corresponding measurement position, the corresponding measurement position and estimated value are displayed on the CRT terminal 52 in step 326.

[Equation 5]

If there is no applicable measurement position, step 3
27, the CRT said “There is no corresponding measurement position”
It is displayed on the terminal 52.

The results of the one-way ANOVA are shown in Table 2 and displayed on the CRT terminal.

[Table 2]

In the measurement timing correction method and the measurement timing correction apparatus of the first embodiment having the above-mentioned operation, the result of the analysis of variance by the one-way arrangement out of a plurality of m-divided planned measurement start points is equal to or higher than a certain reliability estimation value. Since the scheduled measurement start point is determined as the measurement start point at which the measurement start switch is set, it is possible to determine the appearance of the measurement target on the continuously moving conveyor by the automatic appearance measurement using the binarized image processing technology. The measurement timing, that is, the measurement position can be accurately obtained. That is, the measurement position for accurately measuring the image pickup position of the CCD camera 3, for example, the center of the visual field and the measurement position of the characteristic amount, for example, the center of the measurement target, is accurately obtained.

Further, the measurement timing correction method and apparatus of the first embodiment removes white or black isolated points in the binarized image converted by the converter 42, and therefore is included in the binarized image data. The effect that the noise generated can be removed is produced.

Furthermore, the measurement timing correction method and apparatus of the first embodiment stores the feature quantities of the plurality of measurement targets at the respective measurement start points in the unitary arrangement table of the measurement positions and the number of times of measurement. This has the effect of facilitating the analysis of variance of the arrangement.

Further, since the measurement timing correction device of the first embodiment is configured to calculate the measurement timing, that is, the measurement position in the image processing device 4, a special computer or the like is used to calculate the measurement timing, that is, the measurement position. Has an effect that is unnecessary.

(Second Embodiment) As shown in FIGS. 8 to 14, the automatic visual inspection apparatus of the second embodiment includes a limit switch 101 arranged at the measurement position determined by the first embodiment, and a conveyor. CC that measures the lever housing of the transmission that is the measurement target 7 moving up
The ID reading start L that determines the D camera 3, the ID tag 83 in which the discrimination data of the measurement target 7 is recorded, and the area where the discrimination data can be read from the ID tag 83
S80 and the ID head 82 for actually reading the discrimination data
And an ID read end LS 81 that determines the limit of the area where the determination data can be read, and the limit switch 1
01, 80, 81, the sequence controller 2 connected to the ID head 82, the CCD camera 3 and the binary image obtained by converting the captured image measured by the CCD camera 3 connected to the sequence controller 2 An image processing device 104 that determines the degree of conformity between the amount and a reference value according to the type of the measurement target, and a display unit 5 that includes a TV monitor 51 that displays the image processing result and a CRT terminal 52.
It consists of and.

In the conveyor, the same type of measurement object 7 does not flow continuously, so that the measurement object 7 shown in FIG.
As shown in FIG. 2, an ID tag 83 in which discrimination data representing the type is recorded in advance for each measurement target 7 is installed.
An ID read start LS 80 and an ID read end LS 81 are installed at positions that determine the area where the discrimination data can be read from the D tag 83.
The identification data is read from the ID tag 83 by the ID head 82 until the ID reading end LS is turned on after S80 is turned on. These limit switches 8
When the discrimination data cannot be read from the ID tag 83 in the area determined by 0 and 81, it is determined that an abnormality has occurred. As shown in FIG. 9, the limit switch 101 is set at a measurement position determined by an analysis of variance of the one-way arrangement of the first embodiment in a stationary portion facing the conveyor.

The CCD camera 3 is shown in FIGS.
As shown in FIG. 3, two fluorescent lamps 31 and 32, which are arranged directly above the measurement target on the conveyor and are sufficiently longer than the measurement target, are arranged in parallel on the conveyor. The fluorescent lamps 31 and 32 are provided with 18 W in order to obtain an illuminance that can be discriminated even if reflection occurs on the upper and side surfaces of the measurement target.
Inverter fluorescent lamp for image processing was adopted.

The CCD camera 3 and the two fluorescent lamps 31,
As for the positional relationship of 32, as shown in FIG. 13, fluorescent lamps 31 and 32 are arranged parallel to the flow direction of the conveyor at positions separated from the center of the CCD camera 3 by d, respectively. Is arranged so as to cover the entire area of the measurement target at a position not irradiated by the irradiation light.

As shown in FIG. 8, the image processing device 104 includes an image capturing device 41 that captures the captured image measured by the CCD camera 3, and a conversion device 42 that converts the captured image into a binarized image. A determination device 4 that determines the degree of conformity between the feature amount calculation device 43 that calculates the feature amount of the binarized image and the reference value that corresponds to the feature amount according to the type of the measurement target.
It consists of 6 and.

The reference value corresponding to the feature quantity corresponding to the kind of the measurement target is experimentally measured in advance by a CCD camera in a state where the measurement target to be measured is stationary, and a captured image is taken in and converted into a binary image. Then, the feature amount is obtained for each measurement target and stored in the memory of the image processing device 4 as a reference value.

A flow chart shown in FIG. 15 is stored in advance in the memory of the image processing apparatus 104.

The operation of the automatic visual inspection apparatus of the second embodiment having the above construction will be described with reference to FIGS. 15 and 16.

When the limit switch 101 for starting the measurement is turned on in step 401, the memory area for capturing the image is erased in step 402, and the image capturing device 41 operates the CC in step 403.
The captured image from the D camera 3 is A / D-converted and stored in the image memory, the grayscale image is preprocessed, and the region is divided.

Step 404 by the conversion device 42.
In, the captured image is converted into a binarized image based on the binarized threshold value.

The black and white isolated points of the binarized image converted in steps 405 and 406 are removed, the window area is set, and measurement is performed.

In step 407, the feature quantity computing unit 43 determines the suitability from the binarized image in the window area among the area value, the perimeter, etc., that is, the total area, the contour length, the number of holes, the hole area and the like. At least one suitable feature amount is extracted. In other words, if it is not possible to accurately determine the goodness of fit with one feature amount, two or more feature amounts, such as total area and perimeter, total area and number of holes or hole area, perimeter, may be used as necessary. Judgment is made by combining length, number of holes, total area, etc.

In steps 408 and 409, the degree of conformity with the reference value of each measurement object (lever housing) 7 determined experimentally in advance by the determination device 46 is determined, and in steps 410 and 411, the degree of conformity is determined based on the degree of conformity. The conformity is determined to be good (OK) or bad (NG).

In step 412, the extracted feature amount and the determination result are displayed on the CRT terminal of the display device 6.

The automatic appearance inspection apparatus of the second embodiment having the above-described operation is the experimental value stored in the memory and the characteristic amount of the measuring object (lever housing) 7 having a similar appearance continuously flowing by the conveyor. It is possible to determine which type of measurement target (lever housing) it is by determining the degree of conformity with the reference value of each measurement target obtained in (1). That is, in the past, it was realized that the measurement target had to be kept stationary and the measurement determination of the measurement target having a similar moving appearance was difficult.

Further, the automatic visual inspection apparatus of the second embodiment uses a fluorescent lamp 31 or 32, which is a linear light source and covers the measurement field of view and is arranged in parallel to the flow direction of the conveyor, to irradiate the measuring object with a light beam. Is uniformly illuminated, and the measurement target on the conveyor is measured by the CCD camera 3 arranged outside the direct ray region between the fluorescent lamps. It is possible to realize highly accurate measurement without being affected by the rays of

Furthermore, the automatic visual inspection apparatus of the second embodiment is
The feature quantity computing device 43 extracts at least one of the area to be measured, the perimeter, the number of holes, the hole area, and the angle from the binarized image in the window, and the appearance is similar based on the feature quantity. Since the appearance determination of the measurement target is performed, the highly reliable appearance determination is realized, and it is possible to determine the approximate measurement target.

(Third Embodiment) The measurement timing correction apparatus and the automatic appearance inspection apparatus of the third embodiment are the same as the measurement timing correction apparatus of the first embodiment and the automatic timing inspection apparatus of the second embodiment by one and the same image processing apparatus. The program of the first embodiment and the second embodiment corresponds to a subroutine when the main program as shown in FIG. 17 is newly added to the image processing apparatus 4 which constitutes the appearance inspection apparatus.

In step 501 of the main program, parameters and variables are initialized, and step 50
2 asks whether to select the automatic operation mode, so if the NO manual mode is selected, step 5
In step 03, processing is performed according to the command input from the CRT terminal.

The manual mode in step 503 is for setting a window area and a binary threshold value, specifying parameters and execution of measurement timing correction, setting and storing a reference value of a feature amount, and the like.

When the automatic operation mode is selected in step 502, whether or not to correct the measurement timing is confirmed in step 504. Therefore, when the one to be performed is selected, the first execution is performed in step 505. The correction processing of the measurement timing similar to that described in the example is executed.

If the correction of the measurement timing is not selected in step 504, the same automatic appearance inspection process as that described in the second embodiment is executed in step 506.

Immediately after the power is turned on, the automatic operation mode is set,
When initial setting has already been performed and measurement timing correction processing has already been performed, a program is built so that the appearance inspection can be immediately executed.

The device of the third embodiment having the above-described structure has the same effects as those of the first and second embodiments, and since it has one image processing device, it is easy to handle and maintain. The effect of reducing the cost is achieved.

Further, in the apparatus of the third embodiment, when the initial settings have already been made and the measurement timing correction processing has already been made, the automatic operation mode is set immediately after the power is turned on, and the appearance inspection can be executed immediately. This has the effect of enabling quick measurement.

The above-described embodiments are merely examples for the purpose of explanation, and the present invention is not limited to them. Those skilled in the art will recognize from the claims, the detailed description of the invention and the description of the drawings. Modifications and additions can be made without departing from the technical idea of the present invention.

In the above embodiment, the limit switch has been described as an example of the switch for detecting the arrival of the measurement object at the measurement start position, but the present invention is not limited to this, and as shown in FIG. A magnetic switch, a photoelectric switch, or another proximity switch 102 can be used.

In the second embodiment described above, the lever housing of the transmission has been described as an example of the object to be measured, but the present invention is not limited to this. For example, the appearance is similar depending on the vehicle type and other types. If there are many shapes that can be recognized by the camera, they can be measured.

In the above embodiment, the measurement timing, that is, the measurement position is calculated in the image processing apparatus, but it is also possible to adopt a mode in which the image processing apparatus also has the function of the sequence controller.

In the above embodiment, the reliability estimated value of 95% has been described as an example. However, the reliability estimated value of 99%, 98%, 90% or the like can be adopted, and the reliability estimated value can be measured. It is also possible to install multiple cameras for use as needed.

Even if there is a difference in the measurement positions x0 to xm described in the first embodiment, the 95% confidence estimation value is examined and compared with the reference value Mm to determine the position where the limit switch is installed. Since it can be used, it may be adopted if accuracy is not so required.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment device of the present invention.

FIG. 2 is an explanatory diagram for explaining a change in an image due to a positional relationship among a measurement target, a camera, and a measurement target.

FIG. 3 is a chart showing a basic flow of the first embodiment.

FIG. 4 is a chart showing a flow of a one-way ANOVA in the first embodiment.

FIG. 5 is a detailed chart showing an operation sequence in the first half of the first embodiment.

FIG. 6 is a detailed chart diagram showing an operation sequence in the latter half of the first embodiment.

FIG. 7 is a diagram showing a relationship between a measurement position and a feature amount.

FIG. 8 is a block diagram showing a second embodiment device of the present invention.

FIG. 9 is a perspective view showing a limit switch of the second embodiment.

FIG. 10 is a perspective view showing an arrangement relationship among a camera, a fluorescent lamp, and a measurement target in the second embodiment.

FIG. 11 is a plan view showing an arrangement relationship among a camera, a fluorescent lamp, and a measurement target in the second embodiment.

FIG. 12 is a side view showing a positional relationship among a camera, a fluorescent lamp, and a measurement target in the second embodiment.

FIG. 13 is a front view showing a positional relationship among a camera, a fluorescent lamp, and a measurement target in the second embodiment.

FIG. 14 is a block diagram showing a connection relationship between each constituent element of the second embodiment device of the present invention and a limit switch.

FIG. 15 is a chart showing a flow of automatic visual inspection according to the second embodiment.

FIG. 16 is a chart showing the flow of processing in the image processing apparatus of the second embodiment.

FIG. 17 is a chart showing a main program according to the third embodiment of the present invention.

FIG. 18 is a cross-sectional view showing a modified example of the switch.

[Explanation of symbols]

 1, 101 Limit switch 2 Sequence controller 3 CCD camera 4, 104 Image processing device 5 Display device 51 TV monitor 52 CRT terminal

Claims (7)

[Claims] 1. A setting step of setting a plurality of planned measurement start points, a measuring step of measuring a plurality of measurement targets flowing on a conveyor by a camera based on the set respective measured start points, and a measurement by the camera. A converting step of converting the captured images of the plurality of measured objects into a binarized image; a feature amount calculating step of calculating a feature amount of the plurality of measuring objects from the converted binarized image; An ANOVA step of performing an analysis of variance by one-way arrangement of the feature quantities of the plurality of measurement targets for each of the measurement start scheduled points, and a result of the one-way arrangement analysis of variance A certain confidence estimation value or more of the measurement start scheduled points The measurement start switch is set, and the measurement start point determination step is performed to determine the measurement start point. Measurement timing correction method characterized by. 2. The measurement timing correction method according to claim 1, wherein the converting step includes a step of removing white or black isolated points in the converted binary image. 3. The method of claim 1, wherein the analysis of variance step includes a step of storing the characteristic amounts of the plurality of measurement targets at each of the measurement start scheduled points in a unified arrangement table of measurement positions and the number of times of measurement. How to correct measurement timing. 4. A setting means for setting a plurality of scheduled measurement start points, a camera for measuring a plurality of measurement targets flowing on a conveyor based on the set respective scheduled measurement start points, and the camera measured by the camera. A conversion unit that converts the captured images of the plurality of measurement targets into a binarized image, a feature amount calculation unit that calculates the feature amounts of the plurality of measurement targets from the converted binarized images, and the obtained each measurement An analysis of variance means for performing an analysis of variance by the one-way arrangement of the feature quantities of the plurality of measurement targets for each scheduled start point, and a measurement start switch for the scheduled measurement start point equal to or higher than a certain reliability estimated value as a result of the analysis of variance by the one-way arrangement And a measurement start point determining means for determining a measurement start point for setting the measurement start point. Measurement timing correction device. 5. A measurement start switch set at the measurement start point determined according to claim 1 or 4, and a measurement target when the set measurement start switch is turned on by a measurement target flowing on a conveyor. An image pickup camera that measures the image, a conversion unit that converts the imaged image of the measurement target imaged by the image pickup camera into a binarized image, and a plurality of measurement targets of the plurality of measurement targets based on the binarized image converted by the conversion unit. A characteristic amount calculation unit that calculates a characteristic amount; a determination unit that determines the compatibility between the characteristic amount calculated by the characteristic amount calculation unit and a reference value according to the type of the measurement target; and the obtained characteristic amount. An automatic visual inspection device comprising a display unit for displaying a judgment result. 6. The image pickup camera according to claim 5, wherein the image pickup camera is arranged outside a direct light ray region between linear light sources which are arranged in parallel with a conveyor flow direction so as to cover the measurement visual field. An automatic visual inspection device. 7. The automatic appearance inspection apparatus according to claim 5, wherein the feature quantity computing means computes at least one of an area of measurement target, a perimeter, a number of holes, a hole area, and an angle. .