JP3235387B2 - Lighting condition setting support apparatus and method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing system for capturing an image of an object using an image pickup device under illumination of an illumination device and processing image data obtained by the image pickup. To determine or set the installation position, installation angle, luminous intensity, etc. of the illuminating device, and to select or determine the camera, lens, and close-up ring constituting the imaging device in order to perform imaging suitable for the purpose. , Device or method for supporting a decision or setting.

[0002]

2. Description of the Related Art In order to automate visual inspection of industrial products and the like, automatic inspection devices have become widespread, and an image processing system has been introduced into this automatic inspection device. In order to realize high-precision inspection with an automatic inspection device, it is necessary to obtain an optimal image suitable for the purpose of inspection. In order to obtain an optimal image, it is important to set optimal lighting conditions, imaging conditions, and the like.

[0003] Specifically, an example will be described in detail. In a visual inspection device and a positioning device, an object to be recognized is imaged by a television camera under illumination by a lighting device. The obtained image data is taken into an image processing device and subjected to a binarization process. Based on the binary image data, feature quantities such as the area of the object and the position of the center of gravity are measured, and the quality of the object is determined and the position of the object is determined.

In order to measure feature amounts with high accuracy, it is necessary to obtain good binary image data, and therefore, it is necessary to set the illumination of the observation system to an optimal state. Here, a good binary image means an image in which the edges of the object in the binary image are smoothly continuous and have little isolated noise (small white point or black point).

Therefore, when such a binary image is obtained, it is determined that the illumination of the observation system is set to an optimum state. Even if some disturbance light is added, it can be obtained 2
The fact that the value image does not change is also a requirement for optimal illumination.

The type of lighting device and the lighting method are selected according to the purpose of inspection or the like. Conventionally, guidelines for selecting a lighting device and a lighting method have not been clearly arranged, so that a lighting device and a lighting method to be used have been selected by trial and error.

[0007] Even if a lighting device and a lighting method are appropriately selected, it is not easy to set the lighting device to an optimum installation position and an optimum installation angle. Conventionally, an operation of adjusting the position and angle of the lighting device while watching an image projected on a monitor has been performed.

However, according to such a conventional method, skill is required to determine the position and angle of the lighting device,
It was difficult for inexperienced workers to adjust to the optimum condition. In addition, there is a problem that the position and angle of the lighting device tend to vary depending on the operator, the lighting device is not always installed at the best position and angle, and the recognition accuracy changes every time adjustment is performed.

The lighting conditions include not only the position and angle of the lighting device but also its luminous intensity as important factors.

[0010] It is preferable to determine whether the illumination conditions are appropriate, more practically, by evaluating the obtained image. For example, the optimal lighting conditions for non-defective product inspection are that the non-defective product and the defective product can be reliably discriminated, that is, the lighting conditions under which the discrimination accuracy is high are optimal.

The optimum illumination conditions for the measurement can be said to be the illumination conditions in which the dispersion of the measured values is small, that is, the illumination conditions in which the measurement accuracy is high.

However, according to the above-described conventional adjusting method, there is no guarantee that the lighting conditions under which the non-defective product and the defective product can be reliably discriminated and the lighting conditions under which the variation in the measured value is small are always obtained.

The setting of appropriate photographing conditions or photographing environments is also an important matter. For example, it is necessary to install a camera at a convenient position in the configuration of the inspection system and to select an optimal lens in order to secure a field of view suitable for the purpose of the inspection. Also, it may be necessary to use a close-up ring to achieve the desired installation distance and field of view. Under the restrictions of using existing lenses, it may be necessary to solve the problem of how to properly set the installation distance and the close-up ring in order to achieve a desired field of view.

[0014]

DISCLOSURE OF THE INVENTION The present invention does not require any skill, and determines or sets the lighting conditions such as the installation position, installation angle, luminous intensity, etc. of the illuminating device, the lens, the close-up ring, the type of the camera, and the imaging distance including the camera setting distance It is an object of the present invention to enable automatic or semi-automatic determination or selection of conditions.

[0015] The present invention also provides an illumination condition based on image data obtained by actually imaging an object.
It is an object of the present invention to evaluate and optimize imaging conditions.

More specifically, for a non-defective inspection,
The illumination conditions can be set from the viewpoint of how reliably a non-defective product can be distinguished from a defective product, and in the case of measurement, how small measured values can be obtained.

The present invention provides an apparatus and method for evaluating and optimizing illumination conditions based on actual image data obtained by actually imaging an object, or supporting the same. In particular, in the present invention, the position of the lighting device is adjusted.

The lighting condition setting support device according to the present invention comprises:
An illuminating device for illuminating an object, a supporting device for supporting the illuminating device so as to be adjustable in position, an imaging device for imaging an object under illumination by the illuminating device, and outputting a video signal representing an image of the object, the imaging Image quality histogram generation means for generating an image quality histogram indicating the degree of quality of the image with respect to brightness based on a video signal output from the apparatus; An evaluation value calculating means for calculating; a comparing means for comparing the current evaluation value calculated by the evaluation value calculating means with an evaluation value before the position change of the lighting device; and an output means for outputting a comparison result by the comparing means. Have. In other words, the lighting condition setting support device according to the present invention includes a lighting device for illuminating an object, a supporting device for supporting the lighting device in a position-adjustable manner, and an object under illumination by the lighting device. An imaging device that images an object and outputs a video signal representing an image of the object; an evaluation value calculation unit that calculates an evaluation value of illumination based on the video signal output from the imaging device; It can be said that the apparatus includes a comparing unit that compares the calculated current evaluation value with the evaluation value before the position change of the lighting device, and an output unit that outputs a comparison result by the comparing unit.

In a first embodiment, the output means is a display device for instructing a position adjusting direction of the lighting device in accordance with a result of comparison by the comparing means.

In a second embodiment, the support device includes a driving device for changing the position of the lighting device. The output means is a display device for instructing a position adjustment direction of the lighting device according to a comparison result by the comparison means. Further, there are further provided an input device for inputting a command for driving the driving device, and a control device for controlling the driving device in response to the command input through the input device.

In a third embodiment, the supporting device includes a driving device for changing the position of the lighting device. Further, there is further provided control means for controlling the driving device in accordance with a result of the comparison by the comparing means and adjusting the position of the lighting device to an optimum position.

The lighting condition setting support method according to the present invention comprises:
An illumination device for illuminating the object, a support device for supporting the illumination device so that the position can be adjusted, and an imaging device for imaging the object under illumination by the illumination device and outputting a video signal representing an image of the object. Image processing system,
Based on the video signal output from the imaging device, an image quality histogram indicating the degree of quality of the image with respect to brightness is generated, and an evaluation value of illumination is calculated based on the generated image quality histogram. The evaluation value is compared with the evaluation value before the position change of the lighting device, and this comparison result is output. According to another aspect of the lighting condition setting support method according to the present invention, an illuminating device for illuminating an object, a supporting device for supporting the illuminating device in a position-adjustable manner, and imaging of the object under illumination by the illuminating device Then, in an image processing system having an imaging device that outputs a video signal representing an image of an object, an illumination evaluation value is calculated based on the video signal output from the imaging device, and the calculated current evaluation value is calculated. It can be said that the value is compared with the evaluation value before the position change of the lighting device, and the comparison result is output.

According to the present invention, the illumination conditions are optimized so that a good image can be obtained based on the image data obtained by actually imaging the object, so that no skill is required.
Even if the workers are different, the same good image is always obtained.

In the first embodiment, the operator manually operates while watching the displayed instructions, and in the second embodiment, the operator operates semi-automatically.
The height of the lighting device is adjusted.

In the third embodiment, the height setting of the lighting device is completely automated, so that the position setting of the lighting device is further facilitated, and the work efficiency and rationalization can be realized.

The main part of the image quality histogram generating means described above can be used as an edge image generating device.

The present invention provides an edge image generating apparatus using image quality histogram generating means.

An edge image generating apparatus according to the present invention is an image pickup apparatus for picking up an object and outputting a video signal representing the picked-up image, and an A / D for converting the video signal output from the image pickup apparatus into digital image data. Conversion means, A / D
An image memory for storing the converted one-screen digital image data, a window means for setting a window in the one-screen image data and extracting pixel data for a plurality of pixels in the window, Means for generating data relating to the inclination of the image data; means for calculating the degree of coincidence between the data relating to the inclination and a plurality of edge patterns; means for computing an evaluation value of edge likelihood from the degree of coincidence; Means is provided for controlling the calculation of the evaluation value so as to be performed for all pixels for one screen.

According to the present invention, a window is set in image data for one screen, pixel data for a plurality of pixels in the window is extracted, and data relating to the inclination of the image data in the window is generated. Calculating the degree of coincidence between the data relating to the inclination and the plurality of edge patterns, calculating the evaluation value of the likelihood of the edge from the degree of coincidence, and calculating the evaluation value of the likelihood of the edge for all the pixels for one screen; Since edge image data composed of the above evaluation values is obtained, a good edge image in which good-quality edges are emphasized can be generated, and the influence of noise and the like can be removed.

According to the present invention, a sample of an inspection or recognition (measurement) object is actually imaged under illumination by an illumination device,
An apparatus and a method for optimizing illumination conditions (for example, luminous intensity of a lighting device) by actually inspecting or recognizing (measuring) a sample based on image data obtained by this imaging and evaluating the inspection or recognition result. providing.

First, an illumination condition setting support apparatus and method for a system for performing a non-defective product inspection for discriminating a non-defective product from a non-defective product will be described.

The lighting condition setting support device according to the present invention comprises:
A sample supply device for sequentially supplying a plurality of samples of non-defective products and defective products to a predetermined observation position, an illumination device with variable illumination conditions for illuminating the sample at the observation position, a non-defective product brought to the observation position by the supply device, An imaging device that images each sample of a defective product under illumination by the illumination device and outputs a video signal representing the sampled image; and a feature that measures a feature amount of each sample based on the video signal output from the imaging device. Quantity measuring means, means for calculating an evaluation value relating to the accuracy of discriminating a non-defective product from a defective product based on the feature quantity measured by the feature quantity measuring means, and each lighting condition obtained while changing the lighting condition of the lighting device. There is provided means for searching for an illumination condition having the best evaluation value based on the evaluation value.

The lighting condition setting support method according to the present invention comprises:
A lighting device with variable lighting conditions for illuminating a predetermined observation position is arranged, a plurality of non-defective and defective samples are sequentially supplied to the observation position, and the non-defective and defective samples supplied to the observation position are sampled as described above. An image is taken under illumination by a lighting device, a feature amount of each sample is measured based on a video signal obtained by the imaging, and an evaluation value relating to the accuracy of discriminating a non-defective product from a defective product is calculated based on the measured feature value. An evaluation value for each lighting condition is obtained while changing the lighting condition of the lighting device, and a lighting condition with the best evaluation value is searched.

According to the present invention, the lighting conditions are evaluated from the actually picked-up images of the non-defective and non-defective samples, and the lighting conditions which enable the discrimination between the non-defective and defective products to be performed as reliably as possible are searched. High quality inspection is possible.

Next, a description will be given of a lighting condition setting support apparatus and method for a system for measuring a feature amount of an object.

The lighting condition setting support device according to the present invention comprises:
A sample supply device for sequentially supplying a plurality of samples to a predetermined observation position, an illumination device with variable illumination conditions for illuminating the sample at the observation position, and each sample brought to the observation position by the supply device by the illumination device. An imaging device that captures an image under illumination and outputs a video signal representing the sampled image, a feature amount measurement unit that measures a feature amount of each sample based on the video signal output from the imaging device,
Based on the means for calculating the variance of the feature quantity measured by the feature quantity measuring means, and the variance for each lighting condition obtained while changing the lighting condition of the lighting device, a lighting condition with the smallest variance is searched. Means.

The lighting condition setting support method according to the present invention comprises:
An illumination device with variable illumination conditions for illuminating a predetermined observation position is arranged, a plurality of samples are sequentially supplied to the observation position, and each sample supplied to the observation position is imaged under illumination by the illumination device. The characteristic amount of each sample is measured based on the video signal obtained by the above, the variance of the characteristic amount of all the measured samples is calculated, and the variance for each lighting condition is obtained while changing the lighting condition of the lighting device. Is to be searched for a lighting condition that minimizes.

According to the present invention, a sample of an object to be measured is actually imaged, and the feature amount of the sample is actually measured using image data obtained by this imaging, and the illumination condition under which the variance of the measured feature amount is reduced. Since the search for is performed, the dispersion of the measured values is reduced, and highly accurate measurement can always be performed.

By configuring the sample supply device with a rotary table on which each sample is placed and a drive device for rotating the rotary table, the configuration can be simplified and space can be saved.

Finally, the present invention provides a photographing condition determination support apparatus for determining an appropriate photographing condition by a camera having a detachable lens and a necessary close-up ring.

This apparatus includes information on the visual field of the camera, information on the installation distance of the camera to the object to be photographed, information on the lens,
An input device for inputting at least two pieces of information of the close-up ring, an arithmetic device for calculating remaining information using at least two pieces of information input from the input device, and an arithmetic result by the arithmetic device. A display device is provided for displaying the remaining information represented.

According to the present invention, the visual field information of the camera,
The remaining information is calculated and displayed by inputting at least two of the camera installation distance information, the lens information, and the close-up ring information with respect to the object to be photographed. An appropriate lens, close-up ring, camera installation distance, and the like can be selected or determined.

In the evaluation and optimization of the imaging conditions (imaging system), similarly to the evaluation and optimization of the illumination conditions, the object is actually imaged under each of a plurality of types of imaging conditions. It is needless to say that the evaluation value of the imaging condition can be calculated based on the image data obtained by and the evaluation can be performed based on this evaluation value.

[0044]

【Example】

(1) Types of lighting devices and lighting methods Lighting devices (light sources) used in image processing systems
The main ones are fluorescent lamps and halogen lamps.

Fluorescent lamps are roughly classified into a ring type and a straight tube type according to their shapes, and those having various dimensions are commercially available. FIG.
Denotes a ring-type fluorescent lamp 11. FIG. 2 shows a straight tube type fluorescent lamp 12.

In a lighting device using a halogen lamp as a light source, light from the halogen lamp is guided to a light irradiation section using an optical fiber called a light guide. Ring-shaped, slit-type, straight-type, and bifurcated types with various shapes of light irradiators are commercially available.
There are variations in dimensions.

FIG. 3 shows an illuminating device 13 having a ring-shaped light irradiating portion and called a so-called ring light guide. A large number of optical fibers 17 are bundled, and their ends are arranged in a ring to form a light irradiation section.

FIG. 4 shows an illumination device 14 called a so-called slit ring guide. By arranging the tips of a large number of optical fibers 17 in a straight line, a light irradiation section is formed, which is suitable for generating slit light.

FIG. 5 shows an illuminating device 15 for projecting light having a circular cross section from the tip end where a number of optical fibers 17 are bundled.
It is called a straight light guide.

In the illumination device 16 shown in FIG. 6, a bundle of optical fibers 17 is divided into two, and light having a circular cross section is projected from each end. This is called a bifurcated light guide.

In particular, the slit light guide 14, the straight light guide 15, and the bifurcated light guide 16 are suitable when the illumination area is narrow.

The above is summarized as follows. Ring fluorescent light 11 (Fig. 1) Straight tube fluorescent light 12 (Fig. 2) Ring light guide 13 (Fig. 3) Slit light guide 14 (Fig. 4) Straight light guide 15 (Fig. 5) Bifurcated light・ Guide 16 (Fig. 6)

7 to 11 show various illumination methods described below. Transmitted light illumination (Fig. 7) Uniform light illumination (Fig. 8) Coaxial epi-illumination (Fig. 9) Parallel light illumination (Fig. 10) Dark field illumination (Fig. 11)

The transmitted light illumination (FIG. 7) irradiates the light of the light source 10 from below the diffusion plate 29 on which the object OB is placed. An imaging device (for example, a CCD television camera) 20 is disposed above the object OB, and the camera 20
An image is picked up by light transmitted through the diffusion plate 29 at a portion other than the portion occupied by B.

The uniform light illumination (FIG. 8) is for irradiating the object OB with light having a uniform intensity distribution from above the same object OB where the camera 20 is arranged. In FIG. 8, the illumination light is drawn as parallel light, but need not always be parallel light. Uniform light illumination is also called uniform illumination.

In the coaxial epi-illumination (FIG. 9), a half mirror 28 inclined at approximately 45 ° is arranged between the camera 20 and the object OB. Light is applied to the half mirror 28 from the lateral direction, and the reflected light from the half mirror 28 is applied to the object OB. Half mirror 28 reflected by the object OB
Is transmitted to the camera 20. Coaxial epi-illumination means illumination as if illumination light is emitted from a camera 20 as an imaging device.

The parallel light illumination (FIG. 10) uses the parallel light as the object OB.
It is a method of irradiating the light. In FIG. 10, parallel light is applied to the object OB from obliquely above. This is referred to as parallel projection illumination. Of course, the target object OB may be irradiated with parallel light from above vertically. Generally, the background BG on which the object OB is placed is made of a substance that causes diffuse reflection. As a result, both the object OB and the background BG appear bright to the camera 20.

For dark-field illumination (FIG. 11), a background BG having a mirror surface or a surface close thereto is used. The illumination light is emitted from obliquely above. Diffusely reflected light (diffuse reflected light) from the object OB enters the camera 20. Since the irregularly reflected light from the background BG is very small, the background BG looks dark to the camera 20.

(2) Illumination Condition Determination Support Device (Part 1: Support for Determining Optimal Height) An object is illuminated by an illumination device, and an image of the illuminated object is captured by an imaging device. In an image processing system that performs image processing such as various inspections using image data, a support device for determining an optimum height of an illumination device with respect to an object will be described.

FIG. 12 shows the overall configuration of the image processing system including the optimum height determination support device.

The image processing system includes an illumination device 11 for illuminating the object OB, a television camera 20 as an image pickup device for imaging the object OB, and an image processing device 40.

A ring fluorescent lamp shown in FIG. 1 is used as the illumination device 11, and illuminates the object OB from directly above. The ring center of the fluorescent lamp 11 is located right above the object OB. The distance between the lighting device 11 and the object OB, that is, the height of the lighting device 11 with respect to the object OB is defined as h. As the illumination device 11, a ring light guide 13 shown in FIG.

The lighting device 11 is supported by a column 30 and a supporting device 31 so as to be vertically movable. The support device 31 is a lighting device
There are provided 11 gripping portions 32 and sliding portions 33 fixedly connected to the gripping portions 32 via arms. The sliding part 33 is a cylindrical body, and the column 30 slidably penetrates through the hole of the cylindrical body. Therefore, the sliding portion 33 is supported by the column 30 so as to be vertically movable. The sliding portion 33 is provided with a stopper 34 such as a set screw or a fixing bolt, and the sliding portion 33 is fixed to the column 30 by the stopper 34. Therefore, the lighting device 11 is manually adjusted to an arbitrary height position and fixed.

The camera 20 serving as an image pickup device is
The image pickup center is arranged downward at a position directly above the object so that the center of the image coincides with the center of the ring of the illumination device 11, and the object OB is imaged from above under illumination by the illumination device 11. The illustration of the supporting device of the camera 20 is omitted to avoid complicating the drawing. The camera 20 is fixedly supported by a supporting device. Needless to say, the camera 20 may be supported so as to be vertically movable.

A video signal representing the object OB photographed by the camera 20 (a monochrome video signal is sufficient, but may be a color video signal) is converted to an image processing device 40
Given to. The image processing device 40 is preferably configured by a computer system and has a function as a lighting condition determination support device described later. The image processing device 40 is C
An RT display device 41 is provided. On this display device 41, an image of the object OB taken by the camera 20 is displayed, and the optimum height h _P of the lighting device 11 determined by the image processing device 40 is displayed as shown in FIG. The user is notified of the optimum height h _P.

FIG. 13 shows another example of the image processing system. The same components as those shown in FIG. 12 are denoted by the same reference numerals, and redundant description will be avoided. In this image processing system, the lighting device 11
Is automatically moved up and down by a command from the image processing device 40, and is positioned at the optimum height h _P.

The lifting mechanism 38 of the lighting device 11 includes a rack 35 fixed to the sliding portion 33 and a pinion meshed with the rack 35.
36 and. The pinion 36 is rotatably supported by a fixed bearing (not shown). The pinion 36 is also rotationally driven by a pulse motor 37.

When the image processing device 40 has the optimum height h of the lighting device 11,
Upon determining the _P, device 40 illumination device 11 gives a rotation command of the forward or reverse direction to the motor 37 so that the height h _P. Since the motor 37 rotates the rotation amount is commanded, the lighting device 11 is moved up and down through the pinion 36, the rack 35 is positioned at the position of the optimum height h _P. Elevating mechanism if necessary
A lock mechanism is provided on the motor 38 or the motor 37 to hold the lighting device 11 at the optimum height position.

FIG. 14 shows the structure of a lighting condition determination support device suitable for supporting the determination of the optimum height of the lighting device 11.

The lighting condition determination support device 50 includes a user machine interface 53 and an optimum height determination unit 54. These units 53 and 54 are included in the image processing device 40.

The input device 51 includes a keyboard, a light pen, a mouse, and the like. The output device 52 includes the CRT display device 41 and the pulse motor 37 described above. The user operates the display device by using the user machine interface 53.
The input device 51 is operated in accordance with the menus, instructions, marks, cursors and the like displayed on the display 41 to input various data and information in an interactive manner.

The optimum height h _P determined by the optimum height determination unit 54
Is displayed on the display device 41 (output device 52) or used as a command to the motor 37, as described above. The optimum height determination unit 54 preferably includes a CPU that operates according to a program, and a memory that stores a data base described later and various input data.

FIGS. 16 to 19 show the optimum installation height of an annular illumination device 13 (for example, a ring light guide illumination device) for an object OB having an annular groove (marked portion) g formed on the surface. shows a method for determining h _P is.

The optimum installation height h _P is the height of the illumination device 13 with respect to the object OB for obtaining illumination such that the groove g is dark and the surface other than the groove g of the object OB is uniformly bright. Say.

When the lighting device 13 is installed at a considerably high position as shown in FIG. 16, light from the lighting device 13 reaches the bottom surface of the groove g. Are illuminated, and the groove g does not become dark.

On the other hand, when the illuminating device 13 is lowered to a height equal to or less than a certain upper limit h1, as shown in FIG. 17, the illuminating light does not reach the bottom of the groove g, and the groove g becomes dark. . Therefore, the condition of the height h of the illuminating device 13 for obtaining the illumination state in which the groove g becomes a dark portion is given by h ≦ h1.

The upper limit h1 is represented by the following equation.

[0078]

H1 = d × {(R1-R2) / W} Equation 1

Where d is the depth of the groove g, W is the width of the groove g,
R1 is the distance from the center of the field of view to the edge furthest from the center of the groove g, and R2 is the radius of the illumination device 13.

On the other hand, in order to uniformly illuminate the surface other than the groove g of the object OB, conditions such that uneven illumination does not occur on the surface of the object OB in consideration of the equal illuminance curve of the illumination device 13. Needs to be satisfied.

FIG. 18 shows a ring light guide lighting device 13.
3 shows an iso-illuminance curve.

It can be seen from this figure that the radius R of the area illuminated with equal illuminance changes according to the vertical distance from the lighting device 13, that is, the installation height h of the lighting device 13.

FIG. 19 shows the radius R of an area illuminated with equal illuminance.
And the installation height h of the lighting device 13. The radius of the field of view (the range of illuminating the object) of the imaging device 20 is R0. In FIG. 19, the radius R of the area illuminated with equal illuminance
Is equal to the radius R0 of the visual field, the installation height h of the illumination device 13 is h0. Therefore, in order to illuminate the portion other than the groove g brightly and uniformly, the lower limit value of the installation height h of the illumination device 13 may be set to h0. This condition is h ≧ h
0.

From the above two conditions, the optimum installation height h _P of the lighting device 13 can be determined within the range of h0 ≦ h _P ≦ h1. In the memory in the optimum height determination unit 54, information on the radius R2 and the illuminance curve for each type of lighting device is stored as a data base.

FIG. 20 shows the optimum height determination unit 54 for taking an image of the surface g of the object OB having the above-mentioned groove g as a dark part and capturing the surface other than the groove g of the object OB as a uniform bright part. 7 shows a procedure of an optimum height determining process of a lighting device to be executed.

First, the depth d, the width W of the groove g, and the distance R1 from the center of the field of view to the farthest edge of the groove g are input by the user from the input device 51 as information relating to the shape characteristics of the object OB. Further, the user inputs the radius R0 of the field of view of the imaging device 13 through the input device 51 (step S1).
201). However, since the viewing radius R0 may be equal to half the radius of the object OB or the length of the diagonal line thereof,
It can also be input as one of the geometric features of the object OB.

Assuming that the type of the lighting device has already been input or determined, the radius R2 of the lighting device is read from the data base. Further, the lower limit value h0 of the height is derived from the data base relating to the illuminance curve of the illuminating device by using the already input field radius R0 (step 202).

Next, the calculation of the equation (1) is executed to calculate the upper limit value h1 of the installation height of the lighting device (step 203).
). It is determined whether or not the calculated upper limit h1 is equal to or greater than the lower limit h0 (step 204).

If h0 ≦ h1, the optimum height can be determined. Therefore, the average value of the lower limit value h0 and the upper limit value h1 is calculated as the optimum installation height h _P of the lighting device using Expression 2 (Step 205).

[0090]

H _P = (h0 + h1) / 2 Equation 2

The calculation of the optimum height h _P is not limited to the equation (2).
Needless to say, the height h _P satisfying ≦ h _P ≦ h1 can be selected as the optimum one.

The optimum height h _P thus determined
Is output through the output device 52 (step 206). For example, a message image as shown in FIG. 15 for notifying the user of the optimum installation height h _P of the lighting device is displayed on the display device.
Displayed at 41. Alternatively, the motor 37 is driven based on the optimum height h _P , and the lighting device is automatically positioned at the height h _P.

If h0 ≦ h1 is not satisfied, a message is displayed on the display device 41 indicating that the optimum height h _P cannot be obtained (step 207).

(3) Lighting Condition Determination Supporting Device (Part 2: Supporting Determination of Optimal Angle) Next, a description will be given of a supporting device for determining the optimum angle of the lighting device with respect to the object in the image processing system.

FIG. 21 shows the overall configuration of the image processing system including the optimum angle determination support device.

As the illumination device, a slit light guide 14 shown in FIG. 4 is used. The illuminating device 14 is held in a horizontal position in the longitudinal direction, and is arranged so as to irradiate the object OB with illumination light obliquely from above. The lighting device is a straight tube fluorescent lamp 12 shown in FIG.
Can also be used.

In FIG. 21, the same components as those shown in FIG. 12 are denoted by the same reference numerals, and redundant description will be avoided. Holder of lighting device 14
32A is tiltably held by the sliding part 33,
14 directions can be manually set to any angle and fixed.

As in the system configuration shown in FIG. 13, a mechanism for automatically moving the sliding portion 33 up and down and a mechanism for automatically positioning the grip portion 32A at an arbitrary angle are provided. Needless to say, the height and the angle can be automatically set according to the command.

FIG. 22 shows an example of the object OB. The target object OB is a plate-like body, and a linear groove G extending in the longitudinal direction is formed on the surface of the target object OB.

FIG. 23 shows the configuration of a lighting condition determination support device suitable for supporting the determination of the optimum installation angle of the lighting device 14. Compared to the configuration shown in Fig. 14, the optimum height determination unit
An optimum angle determination unit 55 is provided instead of 54. Other configurations are the same as those shown in FIG.

FIGS. 24 and 25 show an object O having a groove G.
A method for setting the installation angle θ of the illumination device 14 (slit light guide illumination device) with respect to B to an optimum value is shown.

The optimum installation angle θ of the illumination device 14 is for obtaining illumination such that the groove G of the object OB is a dark portion and the other surface portion is a bright portion. To realize this, assuming that the width of the groove G is S and the depth of the groove G is t, Equation 3
The relationship must be satisfied. Angle θ is lighting device 14
It is the angle between the light ray (optical axis) from the camera and the horizontal plane.

[0103]

Tan θ ≦ t / S Equation 3

On the surface of a normal object, as shown in FIG. 25, light incident on the object is incompletely diffusely reflected. That is, the intensity of the specularly reflected light equal to the incident angle is the strongest, and the intensity of the reflected light becomes weaker as the angle deviates from the specularly reflected angle. In order to image the surface other than the groove G of the object OB as a bright portion and obtain a clear contrast with the groove G of the dark portion, the imaging device 20 is required.
It is necessary to increase the intensity of light reflected from portions other than the groove portion G incident on the surface. Since the imaging device 20 is installed directly above the object OB or at a position close thereto, the imaging device 20
In order to make as much specularly reflected light as possible, the installation angle θ of the illumination device 14 should be as close as possible to 90 degrees. From this, the optimal installation angle θ of the lighting device 14
Is given by Equation 4.

[0105]

Equation 4 θ _P = tan ⁻¹ (t / S) Equation 4

FIG. 26 shows an object OB having a groove G.
The figure shows the procedure of the process executed by the appropriate angle determining unit 55 to find out the illumination conditions for imaging the groove G as a dark part and the other parts as a bright part.

First, the user inputs the object OB from the input device 51.
The data on the width S and the depth t of the groove G is input (step 211). Next, the calculation of Expression 4 is executed, and the optimum installation angle θ _P of the lighting device 14 is calculated (step 212).
). The calculated optimum installation angle θ _P is output from the output device 52 (step 213). For example, a message image (not shown) for notifying the user of the optimum installation angle θ _P of the lighting device 14 is displayed on the screen of the display device 41.

(4) Lighting Condition Determination Support Device (Part 3: Support for Determining Lighting Device and Lighting Method) In an image processing system, in order to obtain an image suitable for the purpose of inspection, which lighting device is used and which lighting device is used Whether to illuminate with such a lighting method is also a very important factor.

The present inventors examine in detail what kind of lighting device is used in what kind of lighting method in order to examine what and how, and based on the following three factors, an appropriate lighting We have learned that the device and lighting method can be determined.
That is, there are three factors: the purpose of the inspection, the difference in the optical properties between the target portion and the background, and the size of the visual field.

The lighting device and the lighting method can be determined temporarily based on the above factors. Strictly speaking, it is also important to determine the lighting device and lighting method based on the difference (optical transmittance, reflectance, reflection direction, color, etc.) among the differences in the optical properties of the focused part and the background of the factors. It is. This is because the target portion to be inspected and the background generally do not differ only in one of transmittance, reflectance, reflection direction, color, and the like.

Therefore, it is necessary to consider factors suitable for each inspection purpose based on the factors. For this purpose, expert knowledge and optical analysis results acquired in advance are used.

FIG. 27 shows an example of the configuration of a lighting condition determination support device suitable for supporting the determination of a lighting device and a lighting method. In comparison with FIG. 14 and FIG.
Instead of 55, a lighting device and lighting method determining unit 56 is provided.

FIG. 28 shows the appearance of two connectors as an example of an inspection object (work). This connector will be described later in an example in which a lighting device and a lighting method are specifically determined.

FIG. 29 shows the overall procedure of the processing for determining the lighting device and the lighting method by the lighting condition determination support device 50.

When the user inputs the inspection purpose from the input device 51 through a dialogue using the display screen of the display device 41, the process jumps to one of the processes in FIGS. 30 to 36 according to the input inspection purpose. (Step 221). The relationship between the inspection purpose and these processes is as follows.

Inspection of shape and dimensions Inspection of FIG. 30 Inspection of presence / absence or lack of articles Inspection of FIG. 31 Inspection of scratches Inspection of FIG. 32 Inspection of color Processing of FIG. 33 Positioning Processing of FIG. 34 (A), (B) Inspection of foreign matter Processing of Fig. 35 Inspection of characters and marks Processing of Fig. 36

Although the size of the field of view may be input, FIG.
As can be seen from the side of the display screen shown in FIG. 43, the lighting device suitable for the size of the visual field is displayed, so that the user can determine the lighting device without inputting the size of the visual field. . Of course, after inputting the size of the visual field, only the lighting device suitable for the input visual field size may be displayed.

As will be described later in detail, in the specific processing that matches the inspection purpose shown in FIGS. 30 to 36, the difference in the optical property between the focused part of the factor and the background is sequentially asked on the screen of the display device 41. When the user inputs a response to the response through the input device 51, the difference in the optical properties between the background of interest and the background to be focused on for the inspection is derived (step 222). The lighting device and the lighting method are determined (step 223). In this embodiment, the user ultimately determines the lighting device and the lighting method from some combinations of the lighting device and the lighting method displayed on the displayed screen.

In the determination processing for inspecting the shape and dimensions shown in FIG. 30, it is first asked whether silhouette inspection is possible (step 231). When the user inputs an answer to the effect that it is possible, the difference between the optical property of the target part and the background that should be focused on for the inspection is the transmittance, and the transmitted light illumination screen shown in FIG. 37 is displayed on the display device. Displayed at 41 (step 232).

This transmitted light illumination screen indicates that the transmitted light illumination shown in FIG. 7 is suitable if a silhouette inspection is possible. The transmitted light illumination includes a type 1 in which the illumination device, the object, and the camera are horizontally arranged, and a type 2 in which these are vertically arranged (arrangement shown in FIG. 7). For both of these types, an appropriate lighting device is displayed according to the size of the field of view. The user will select one of these displayed lighting devices that matches the size of the field of view. Where φ is the diameter of the ring-type lighting device,
1 (lowercase letter L) (see FIG. 38) indicates the length of the linear lighting device.

When the user inputs that silhouette inspection is not possible, a question as to whether or not the target portion and the background are similar appears on the display device 41 (step 233).
). If the color of the target portion and the background are not similar, the differences to be noted are the color and the reflectance. Thus, the uniform light illumination screen shown in FIG. 38 is displayed (step 234).

On the screen for uniform light illumination, uniform light illumination (FIG. 8)
Are displayed, and the type of the illuminating device suitable for various visual field sizes in the uniform light illumination is displayed.

When the silhouette inspection is not possible and the color of the target portion and the background are similar, the difference to be noted is the direction of light reflection. Then, it is asked whether there is a projection or a hole in the focused portion (steps 235 and 237).

FIG. 39 shows a case where there is a projection at the target portion of the object.
(A screen for illumination focusing on the projections among the projections and depressions) is displayed (step 236). On this screen, the fact that parallel light illumination (FIG. 10) is appropriate and the type of illumination device used in this parallel light illumination are displayed for each size of the field of view.

If there is a hole in the target portion, an uneven hole illumination screen (for performing illumination focusing on the hole among the unevenness) shown in FIG. 40 is displayed (step 238). On this screen, depending on the purpose or location of the inspection, parallel light illumination or coaxial epi-illumination (FIG. 9) is displayed as appropriate, and an illuminating device suitable for each field of view size is displayed.

If there is no protrusion or hole in the target portion, the uneven edge illumination screen shown in FIG. 41 is displayed (step 239).
). On this screen, the fact that parallel light illumination is appropriate and the type of illumination device according to the size of the field of view are displayed.

For the purpose of the inspection shown in FIG. 31 for the presence / absence of parts and the absence of parts, whether the silhouette inspection is possible (step 241), the color of the parts and the base (the body on which the parts are mounted) are determined. If the colors are similar (step 24
3) are sequentially asked through the display device 41.

Referring specifically to the screens other than the illumination screen described with reference to FIG. 30, if the colors of the parts and the base are similar, the regular reflection utilizing illumination screen shown in FIG. 43 is displayed. (Step 245). Illumination using specular reflection includes collimated light illumination (FIG. 10) and dark field illumination (FIG. 11). The user can select a lighting device corresponding to the lighting method and the size of the field of view displayed on this screen.

The processing shown in FIGS. 32 to 36 is almost the same as that described above, and the contents of the processing can be understood from these drawings. The screen for uneven line illumination in Fig. 42, which has not been mentioned yet, is explained briefly.
Is displayed when the silhouette inspection is not possible and the direction of the flaw cannot be specified (NO in both steps 251 and 253). Scratches on the surface of the object (thin linear irregularities)
It has been shown that parallel light illumination is appropriate to properly image the image, and that a slit light guide or a straight tube fluorescent lamp should be selected according to the size of the field of view.

As described above, the predetermined question and the user's answer to it are repeated through the display device 41 and the input device 51 in accordance with the purpose of the test, and the optimum target portion to be focused on for the target test is obtained. The difference between the optical properties of the object and the background is determined, and an appropriate lighting device and lighting method are determined based on the result of the determination.

As a specific example, a case where the connector pin shown in FIG. 28 is inspected for detachment will be described. Since the inspection purpose corresponds to the presence / absence of parts and the missing item inspection, the process proceeds to the process shown in FIG. If the configuration is such that silhouette inspection can be performed on the inspection line, the transmitted light illumination screen shown in FIG. 37 is displayed. Since the field of view is about 30 mm, horizontal or vertical transmitted light illumination using a ring-type fluorescent lamp is appropriate. If the silhouette inspection cannot be performed, the process proceeds to the next determination, and since the colors of the parts and the base are not similar, the screen for uniform light illumination shown in FIG. 38 is displayed. In this example, the object is horizontally long (type 2), and the field of view is about 30 mm. Therefore, it can be concluded that it is appropriate to illuminate diagonally from above using a slit light guide.

In the above example, the process proceeds by the user interacting with the decision unit 56 through the display device 41 and the input device 51. The process may be advanced by the user inputting various data and information according to a selection flow or a guide book for selecting the above factors. Although the selection is made in the order of the factor and the order, for example, a procedure for first determining the “size of the visual field” may be used.

(5) Lighting Condition Determination Support Device (Part 4) As the last example of the lighting condition determination support device, the lighting device and the lighting method described above, and the optimum height or the optimum angle of the lighting device are determined. The device will be described.

The lighting condition determination support device 50 shown in FIG. 44 includes the above-described optimum height determination unit 54 and optimum angle determination unit.
55 and a lighting device and lighting method determining unit 56 are provided. Other configurations are the same as those shown in FIGS. 14, 23 and 27.

FIG. 45 shows a processing procedure executed in the device 50.

First, as described above, a lighting device and a lighting method suitable for the inspection purpose are determined through a dialog between the user and the determining unit 56 (step 311). The user displays 41
The most suitable lighting device and lighting method are selected from those displayed on the screen and input.

For the lighting device thus determined, it is determined whether to determine the optimum height or the optimum angle, and one of the determining units 54 and 55 is selected accordingly (step 312). ). Generally, it is necessary to determine an optimum height for a ring-shaped lighting device, and to determine an optimum angle for a linear lighting device. Necessary information is input according to the optimum height or angle to be determined, and the optimum height h _P or the optimum angle θ _P is determined according to the processing shown in FIG. 20 or the processing shown in FIG. 26 (step 313). , The result is displayed on the display device 41,
Alternatively, it is provided to a height or angle adjusting device (elevating mechanism or the like) (step 314).

FIG. 46 shows a processing procedure for determining an optimum height for each of a plurality of types (n types) of illumination devices when they are selected. The same processes as those shown in FIG. 20 are denoted by the same step numbers, and description thereof is omitted.

A counter i is provided. Initially, the counter i is initially set to i = 1 (step 321), and is incremented when there is an illumination device that does not satisfy the condition h0 ≦ h1 in step 204. (Step 323
). Input of various conditions after the (step 201), the type of i-th lighting device is input (step 322), whether determine optimum height h _P is determined for the inputted lighting device (step 204) . If it can be determined, the optimum height h _P is obtained, and the process is completed. If the optimum height cannot be determined for all types (n types) of lighting devices (step 324), a message to that effect is displayed and the process ends (step 207). In step 322, the radius R0 of the field of view having uniform illuminance of the illumination device may be input for each type of illumination device.

(6) Optimal lighting condition setting support device (Part 1:
Setting of Position of Illumination Device) In the above-described illumination condition determination support devices (Nos. 1 and 2), the optimal height and the optimal angle of the illumination device are determined based on the positional relationship between the illumination device and the inspection object. . Next, a description will be given of a device that determines the optimal position of the lighting device from the viewpoint of the quality of the image captured by the camera, or that assists the determination.

Also in this embodiment, the configuration of the image processing system shown in FIG. 12 is applied. The image processing device 40 captures a video signal representing the object OB captured by the television camera 20, and executes image processing (described later) to support setting of the optimum height of the lighting device 11. CRT display device
Reference numeral 41 displays a captured image of the object OB and a message for giving advice on the position setting of the lighting device 11.

FIG. 47 shows a circuit configuration of the image processing apparatus 40. The image processing device 40 includes an image input unit 61, an image memory 6
2, an image output unit 63, a microcomputer, a timing control unit 64, and the like.

The image input unit 61 includes an A / D converter for converting an analog video signal output from the television camera 20 into digital image data, and supplies the digital image data to the image memory 62. In addition, the image input unit 61 includes a histogram generation circuit that generates an image quality histogram representing image quality on the horizontal axis based on the input analog video signal in real time. The specific circuit configuration and operation of this histogram generation circuit will be described later.

The image memory 62 stores image data provided from the image input unit 61 in pixel units. The image output unit 63 includes a D / A converter for converting the image data read from the image memory 62 into an analog video signal, and supplies the converted one-screen analog video signal to the CRT display device 41. Thus, the object O photographed by the television camera 20
The image of B is also displayed on the display device 41.

The microcomputer has a CPU 65 which is a main body of control and calculation, and a ROM 66, a RAM 67, an I / O control unit 68 and the like are connected to the CPU 65 via a system bus. The CPU 65 decodes and executes programs stored in the ROM 66, and reads and writes various data to and from the RAM 67 to execute predetermined image processing. I /
A keyboard 69 is connected to the O control unit 68. A key input signal from the keyboard 69 passes through the I / O control unit 68 to the CP.
Sent to U65.

The timing control section 64 outputs a timing signal for controlling input / output operations of various data to the image input section 61, the image memory 62, and the image output section 63 under the control of the CPU 65.

FIG. 48 shows a specific example of the histogram generation circuit provided in the image input unit 61.

The CPU 65 sends an address / address via the CPU bus.
Controls the generator 80. Address generator 80
Generates necessary address signals, control signals, and timing signals and supplies them to the components. Image memory 62
The writing and reading are controlled by a control signal input via a control bus. Image memory 62 according to an address signal input via an address bus
The image data supplied from the data bus is written to and read from the address. It goes without saying that the address bus, control bus and data bus are also connected to other circuits. In the image memory 62,
Image data for one field (or one frame) is written.

As shown in FIG. 50, a window of 3 × 3 = 9 pixels is set in the image data of one field (or one frame) stored in the image memory 62, as shown in FIG. Let the pixel at the center of this window be P ₀ ,
Pixels around this center pixel P ₀ are counterclockwise P _1-
Given the sign of the P _8. The 3 × 3 window moves in the horizontal and vertical scanning directions over the entire screen area.

To set such a window,
Latch circuit 76, 75, 1H delay circuit (shift register)
78, latch circuits 70, 74, 1H delay circuit 71 and latch circuits 72, 73 are provided. Image data read out from image memory 62 in the order of horizontal and vertical scanning is supplied to these latch circuits and 1H delay circuits. The data is sequentially transferred in the above order. The 1H delay circuits 78 and 71 delay the input data by 1H (one horizontal scanning period) and output the same, and the latch circuits 76, 75, 70, 7
4, 72 and 73 output the input data with one pixel clock cycle delay.

The image data (pixel P ₇ ) read from the image memory 62 and the output image data (pixels P ₆ , P ₅ , P ₀ , P ₄ , P ₄ ) of the latch circuits 76, 75, 70, 74, 72, 73 are output. ₂ , P
₃ ) and the output image data (pixels P ₈ and P ₁ ) of the 1H delay circuits 78 and 71 are given to the non-conformity calculation unit 81. The output image data (center pixel P ₀ ) of the latch circuit 70 is also supplied to the non-compliance degree histogram generation unit 82 and the brightness histogram generation unit 83.

FIG. 49 shows an example of the configuration of the nonconformity calculation unit 81. The non-conforming calculation unit 81, the vertical, horizontal, fit of the window in the image data for each edge pattern eight extending obliquely (eight directions); (edge fitness edge resemblance) μ ₁ ~μ ₈ determined, MAX operation of these fit μ ₁ ~μ ₈ (selecting the largest one)
To determine the edge conformity μ. Then, * μ (The non-conformity is indicated by a bar above μ in the drawing.
In the specification, it will be expressed as * μ. ) = 1-μ to determine the degree of non-conformity (unlikely edge).

Eight subtraction circuits 91 to 98 are provided.
These subtraction circuits 91 to 98 are supplied with the image data of the above-described pixels P _{1 to} P ₈ in the window, respectively. Further, all of the subtraction circuit 91 to 98 is given image data of the central pixel P _0. Subtraction circuit 91 calculates the difference between the image data of the image data and the center pixel P ₀ of the pixel P ₁ (which is referred to as q _1). Similarly, the subtraction circuits 92 to 98
The difference between each and the center pixel image data P ₀ of the image data of ₂ to P ₈ is calculated (these and q ₂ ~q _8). Since the image data represents a monochrome image as described above (in the case of color image data, luminance data may be extracted), it indicates the brightness of each pixel. Therefore, the difference data q _{1 to} q ₈ is the central pixel P ₀
It represents the difference in brightness between the pixels P ₁ to P ₈ in its surroundings. FIG. 51 shows brightness patterns P _{0 to} P ₈ in the window.
(P _{0 to} P ₈ indicate image data in this case) shows how brightness patterns (brightness difference patterns) q _{0 to} q ₈ are generated based on the center pixel in the window. .

In this embodiment, eight types of edge patterns are represented by a fuzzy model expressed using a membership function. This fuzzy model is
52 (A)-(H). N is Negative, Z is ze
ro and P represent Positive, respectively, and the membership functions specified by these codes (which are called labels) are shown in FIG.

[0155] membership function N, as shown by the broken line, the negative -q _b smaller range which is a difference q brightness, the function value takes the value of the maximum value 1, the minimum is greater than 0 range It takes the values of 0, in the range of -q _b to 0 takes a value of from 1 changes linearly in response to the value to zero.

[0156] membership function P, the difference q of brightness takes a minimum value of 0 is the function value is less than zero range, takes a maximum value at a certain positive value q _b greater than the range, the range of 0~q _b Takes a value between 0 and 1 that varies linearly in response to that value.

[0157] membership function Z is a function value becomes the minimum is the value q _a larger range of the absolute value of the difference q brightness, the absolute value is the maximum value 1 at q _b lesser extent. Then, the absolute value is large and q _a smaller range than q _b of the difference q brightness, in response to the value, the function value is a middle value between 1 changes linearly to zero.

The fuzzy model shown in FIG. 52 (A) is bright in the lower right direction and dark in the upper left direction, and represents an oblique edge pattern on the upper right. FIG. 52 (B) shows an edge pattern that is bright to the right, dark to the left, and extends vertically. It can be easily understood that the fuzzy models shown in FIGS. 52 (C) to 52 (H) also represent other types of edge patterns. Fig. 52 (A) to Fig. 52 (H)
Are defined as directions 1 to 8.

Returning to FIG. 49, the above-described eight-direction edge
Eight edge matching degree operation circuits 111 to 111 corresponding to the pattern
118 are provided. Since these edge matching degree operation circuits have the same configuration except for the set membership function, the circuit 111 will be described.

Each of the edge matching degree calculation circuits 111 has a membership function circuit (hereinafter referred to as MFC: Membership Funct).
ion circuits) 101 to 108 and a MIN circuit 109 for selecting the minimum value of their outputs. Output data q _{1 to} q ₈ of each subtraction circuit 91 to 98 is MF
C101 to 108 are input.

The MFCs 101 to 108 have the right side shown in FIG.
A membership function Z, N or P is set corresponding to the fuzzy model of the edge pattern of direction 1 pointing upward by 45 degrees. Specifically, pixels P ₁ and P _1
5 MFC101 and 10 corresponding to the (brightness difference q ₁ of a q ₅₎
5 has the membership function Z as the pixels P ₂ , P ₃ , P ₄
MFC 102 corresponding to (brightness difference q ₂ , q ₃ , q ₄ )
, 103, 104 have a membership function N, and MFCs 106, 107, 108 corresponding to pixels P ₆ , P ₇ , P ₈ (brightness differences q ₆ , q ₇ , q ₈ ) have a membership function P. , Are set respectively.

The MFC 101 calculates and outputs the degree of conformity μ ₁₁ of the input data q _{1 to} the membership function Z set therein. When the MFC is configured by a memory that stores membership function data, it is only necessary reading the membership function data corresponding to the input data q _1. Just as MFC 102 -108 calculates and outputs the fitness μ ₁₂ ~μ ₁₈ of the input data q ₂ to q ₈ for the membership function N, Z, or P, which is set.

The MIN circuit 109 selects the smallest one of the fitness levels μ _{11 to} μ ₁₈ output from the MFCs 101 to 108 and outputs the selected fitness level as the fitness level μ ₁ of the edge fitness calculation circuit 111. As the value of the fitness mu ₁ is large (close to 1), if based on the brightness of the pixels P ₀ binarizes the data of other pixels in the window, a more appropriate binary image Is obtained.

The M of the other edge matching degree calculation circuits 112 to 118
FC has membership functions representing the edge patterns shown in FIGS. 52 (B) to 52 (H). All output data q ₁ to q ₈ of the subtraction circuit 91 to 98 is provided for each of these edges fit calculation circuit 112 to 118, fit these circuits 112 ~118 μ ₂ ~μ ₈
Is obtained.

The MAX circuit 85 is provided with the edge matching degree operation circuit 11
Select the largest of relevance μ ₁ ~μ ₈ to 1 to 118 are outputted, and outputs the goodness of fit mu. Since the highest degree of matching obtained when the direction (pixel pattern) of the edge in the window corresponds to the direction 1 to the direction 8 (pattern set in advance) is selected, the degree of matching μ is
This indicates the degree of conformity in the direction that matches the best of the directions 1 to 8.

The fitness μ output from the MAX circuit 85 is
It is provided to a subtraction circuit 87. As the other input of the subtraction circuit 87, data representing a logical value 1 output from the 1 generation circuit 86 is provided. The subtraction circuit 87 outputs a value * μ = 1−μ obtained by subtracting the degree of conformity μ from the value 1. This value * μ
Represents the degree of non-conformity (unlikely an edge) with the most appropriate edge pattern.

The above processing is repeatedly performed by scanning the window over all the pixels of the image data for one field (or one frame). Therefore, the degree of incompatibility * μ is obtained for all the pixels (for the window in which all the pixels are the center pixels).

The nonconformity * μ obtained by the nonconformity calculation unit 81
Is given to the non-compliance degree histogram generation unit 82. Also the non-conforming histogram generator 82, image data of the central pixel P ₀ (representing the brightness of the central pixel P ₀₎ is given as described above. Non-conformance degree histogram generator 82
Is calculated for each brightness represented by the image data of the central pixel P ₀ that caused the nonconformity * μ (in the case of 8-bit representation, for each of 256 steps) ) Cumulative addition to generate a non-conformity degree histogram. In this non-conformity degree histogram, the horizontal axis indicates brightness, and the vertical axis indicates the sum of non-conformity degrees.

[0169] brightness histogram generator 83, based on the image data of the reference pixel P ₀ given (brightness), each brightness, the number of reference pixels P ₀ with its brightness by accumulating, Generate a brightness histogram. The horizontal axis of this brightness histogram is brightness (for example, 256 levels), and the vertical axis is the number of pixels having that brightness.

Data representing the non-conformity degree histogram generated by the non-conformity degree histogram generation unit 82 and the brightness histogram generated by the brightness histogram generation unit 83, respectively, are given to the image quality histogram generation unit 84. The image quality histogram generation unit 84 calculates, for each brightness,
The cumulative addition value of the nonconformity in the nonconformity histogram is normalized by dividing by the number of corresponding pixels in the brightness histogram to calculate an average value of the nonconformity * μ. The image quality histogram shown in FIG. 54 is obtained by taking the average value of the nonconformity * μ on the vertical axis and the brightness on the horizontal axis. The average value of the nonconformity * μ on the vertical axis represents poor image quality as an evaluation standard of image quality.

In the image quality histogram shown in FIG. 54, the maximum value is Q _max , and the minimum value is Q _min . When the image quality data is binarized using the brightness k _{opt at} which the minimum value Q _min is obtained as a threshold value, the image quality is minimized and the best binary image data is obtained.

The CPU 65 reads the maximum value Q _max and the minimum value Q _min of poor image quality from the image quality histogram generated by the image quality histogram generating section 84. CPU 65
Reads the brightness k _opt corresponding to the minimum value Q _min . Thereafter, the CPU 65 searches the image quality histogram on the horizontal axis on the horizontal axis on the basis of the brightness k _opt in the bright and dark directions, and the value on the vertical axis is (1-a) × Q _max + a × Q first.
The width WB of the brightness exceeding the value of _min is calculated. Where a
Is a positive coefficient smaller than 1 and indicates tolerance.

The CPU 65 calculates the evaluation value V of the illumination at the set height h of the illumination device 11 according to the formula 5, and stores the calculation result in the RAM 67.

[0174]

V = (Q _max / Q _min ) WB Equation 5

In equation (5), Q _max / Q _min represents the depth of the valley in the image quality histogram. The larger this value is, the more clearly the edge-like portion and the non-edge-like portion are clearly separated. This means that the contrast is clear. Further, since WB is the width of the valley as described above, the larger the value is, the more the image quality does not change even if the illuminance of the illuminating device changes, that is, the higher the contrast variability.

The calculation of the evaluation value V is not limited to Equation 5, but can also be performed using Equations 6 to 10 below.

[0177]

V = Q _max / Q _min Equation 6

[0178]

V = WB Equation 7

[0179]

V = 1 / Q _min Equation 8

[0180]

V = WB / Q _min Equation 9

[0181]

V = (Q _max / Q _min ) × (WB / k _opt ) Equation 10

In particular, Equation 10 gives an evaluation value in consideration of the absolute value of the illuminance fluctuation of the light source and the width of the brightness that enables stable photographing.
In general, the degree of darkening due to the deterioration of the light source is proportional to the illuminance at that time. Therefore, the absolute value of the illuminance fluctuation due to the deterioration is smaller in a light source with a lower illuminance. On the other hand, the width WB is constant. Assuming that ( _Qmax / _Qmin ) WB is constant, _kopt should be small. The evaluation value V increases as k _opt decreases.

FIG. 55 shows the relationship between the installation height h of the lighting device 11 and the evaluation value V of the lighting. At a certain installation height h _opt , the evaluation value V of the lighting takes the maximum value V _max , and the evaluation value V monotonously decreases on both sides. That is, the evaluation value V is 1
It is a convex function with a number of peaks.

Therefore, the height of the lighting device 11 is equal to the evaluation value V
In consideration of the above, the optimum position can be set as follows.

An evaluation value V ₀ is obtained at a certain installation height h ₀ ,
Next, when a larger evaluation value V1 (V1> V0) is obtained when the height is increased to the installation height h1 (h1> h0), the height is further increased in order to set the optimum installation height _hopt. The height of the lighting device 11 may be adjusted in the direction of increasing h.

Conversely, when the evaluation value V0 is obtained at a certain installation height h0, and when the installation height is subsequently increased to h1 (h1> h0), a smaller evaluation value V1 (V1 <V0) is obtained. In order to set the optimal installation height h _opt , the height of the lighting device 11 may be adjusted in a direction to reduce the installation height h.

The CPU 65 of the image processing device 40
Is imaged by the camera 20, the image quality of the captured image is evaluated based on image data obtained by A / D conversion of a video signal output from the camera 20, and an instruction regarding the position setting of the lighting device 11 is given to the display device 41. The display assists the user in setting the lighting device 11 at an appropriate position.

FIGS. 56 and 57 show the processing procedure of such a CPU 65.

An initial screen including a message as shown in FIG. 58 is displayed on the CRT display device 41 (step 331).
). The user looks at this initial screen and loosens the stopper 34 to position the lighting device 11 at an appropriate height h0.
Tighten and fix. The user presses the setting key of the keyboard 69 when the setting of the lighting device 11 is completed.
YES.

The CPU 65 causes the television camera 20 to capture an image of the object, and causes the image input unit 61 to capture the obtained video signal (step 333). If necessary, the captured image is displayed on the display device 41.

The image input unit 61 converts the fetched video signal into an A /
After D conversion and conversion to image data, an image quality histogram is generated by the above-described method. The CPU 65 reads the maximum value Q _max , the minimum value Q _min , and the brightness k _opt corresponding to the minimum value Q _min of poor image quality from the image quality histogram generated by the image quality histogram generation unit 84 (step 334).
). Further, the CPU 65 determines the width WB of the brightness in the valley from the image quality histogram, and obtains the expression 5 or the expression 6 to the expression 10
Is calculated to calculate the illumination evaluation value V0, and the value is stored in the RAM 67 (step 335).

Subsequently, the CPU 65 causes the display device 20 to display a message as shown in FIG. 59 (step 336). The user looks at this display and manually operates the lighting device 11 to set the initial height h0.
It is moved to an appropriate higher height position h1 and fixed. Thereafter, the user operates the setting key, so that YES is determined in step 337.
Becomes

The imaging of the object, the input of image data, and the creation of an image quality histogram based on the input image data are performed again (step 338), and the evaluation value V1 of the illumination corresponding to the height h1 is obtained based on the new image quality histogram. Is calculated (steps 339 and 340).

The CPU 65 sets the evaluation value V0 of the previous (initial state)
Is compared with the current evaluation value V1 (step 341).
If 1> V0, the previous evaluation value V0 is replaced by the current evaluation value V1.
Is stored in the RAM 67 (step 342). Further, the message of FIG. 59 is displayed on the display device 41 to further increase the height of the lighting device 11 (step 343).

If V1.ltoreq.V0, the CPU 65 stores the current evaluation value V1 in the RAM 67 and updates the evaluation value (step 344), and FIG.
Is displayed on the display device 41 (step
345).

When the message shown in FIG. 59 is displayed, the user changes the lighting device 11 to the height position h1 higher than the previous time. When the message in FIG. 60 is displayed, the user changes the lighting device 11 to a lower height position h1 than the previous time.
Thereafter, the user operates the setting key of the keyboard 69, so that the determination in step 346 is YES and the process returns to step 338. The CPU 65 again takes an image of the object OB by the camera 20,
The creation of the image quality histogram and the calculation of the evaluation value are repeated.

While watching the message displayed on the display device 41 in this way, the user repeats the above procedure while adjusting the installation height of the illumination device 11 up and down and gradually reducing the amount of movement. Finally, the message shown in FIGS. 59 and 60 is switched only by slightly adjusting the height. The vicinity of the installation height at that time is the optimum installation height h _opt . Finally, the user operates the end key of the keyboard 69 (step 347), and the adjustment operation ends.

(7) Optimal lighting condition setting support device (Part 2:
Setting of Position of Illumination Device) The support device for the optimal position setting described above is provided by the user using the illumination device 11.
Is manually positioned. Next, an embodiment in which the user inputs the moving direction and the moving distance (or position) of the lighting device 11 from the keyboard to automatically position the lighting device 11 at the commanded position will be described.

An image processing system having an automatic elevating mechanism shown in FIG. 13 is used. As the configuration of the image processing device 40, the configuration shown in FIG. 47 is used. As shown by the chain line, a pulse motor 37 is connected to the I / O control unit 68. The drive of the motor 37 is controlled by a command from the CPU 65 (or by a command from the I / O control unit 68). Figure
The configurations of various circuits shown in FIGS. 48 and 49 are also used as they are.

FIGS. 61 and 62 show the processing procedure by the CPU 65. Also in this figure, the same components as those shown in FIGS. 56 and 57 are denoted by the same reference numerals, and redundant description will be avoided.

In FIG. 61, the processing of steps 331 to 336 is the same as that shown in FIG.

Referring to the message shown in FIG. 59, the user inputs the moving direction (up or down) and the moving distance of lighting device 11 from keyboard 69 (step 352), and operates the setting key (step 352). 337).

Then, based on the data representing the moving direction and the moving distance input from keyboard 69, I / O
The pulse motor 37 is driven via the control unit 68 (step 353). As a result, the illuminating device 11 moves to the specified direction in the specified direction and then stops there. Thereafter, imaging of the object OB, creation of an image quality histogram, and calculation of the evaluation value V1 are performed (steps 338 to 341).

The user looks at the message shown in FIG. 59 or FIG. 60 displayed on the display device 41 and inputs the moving direction and the moving distance of the lighting device 11 through the keyboard 69 according to the message (steps 354 and 355). Thereafter, if there is a setting key input, the process returns to step 353, and the illumination device 11 is moved by the drive control of the motor 37.

The user adjusts the installation height of the lighting device 11 up and down by inputting a command from the keyboard 69 while watching the message displayed on the display device 41 in this manner. When the adjustment is completed, the user operates the end key of the keyboard 69 to end the adjustment work (step
347).

In the above two embodiments, the moving direction of the lighting device 11 is instructed to the user by a visual display using the display device. However, the user is instructed by voice, a different warning sound according to the moving direction, or the like. You may instruct.

(8) Optimal lighting condition setting support device (Part 3:
Setting of Position of Lighting Device) An embodiment of a device for automatically setting the height position of the lighting device to an optimum height will be described in detail.

The image processing system shown in FIG. 13 is employed. The configuration of the image processing device shown in FIGS. 47 to 49 is used. In FIG. 47, a pulse motor 37 is connected to an I / O control unit 68.

The control processing procedure by the CPU 65 is shown in FIG. 63 and FIG. Here, the same processes as those shown in FIGS. 56 and 57 are denoted by the same reference numerals.

First, the rotation angle θ ₁ and the rotation direction of the pulse motor 37 are input from the keyboard 69 (step 36).
2). Instead of the rotation angle theta _1, it may be input to the moving distance. In this case, CPU 65 calculates the rotation angle theta ₁ of the motor 37 required to move the distance moved is input to the lighting device 11. The direction of rotation will generally be set up, but may be down.

Thereafter, when there is a start key input from the keyboard 69 (step 363), the object OB is imaged under illumination by the illumination device 11 positioned at the initial height h0, and an image quality histogram is created. , The evaluation value V0 is calculated (steps 333 to 335).

[0212] Next CPU65 is a pulse motor 37, the initial set angle theta ₁ is rotated to that input direction corresponding distance to raise or lower a lighting device 11 in step 362 (step 364). The height position of the lighting device 11 is h1. At this height position h1, imaging, creation of an image quality histogram, and calculation of the evaluation value V1 are performed (step 33).
9, 340).

[0213] The CPU 65 determines whether or not the absolute value of the difference between the previous evaluation value V0 and the current evaluation value V1 is equal to or smaller than a predetermined threshold TH (step 365). NO in this judgment
If so, proceed to 341.

Even if the result of this determination is YES, the angle θ
(The angle updated in step 367) is a predetermined threshold value θ
It is determined whether it is equal to or less than _th , and if NO, the process also proceeds to step 341.

In step 341, the evaluation values V0 and V1 are compared.

If [0216] V1> a V0, the process returns to step 364, the pulse motor 37 by the same angle theta ₁ to the previous lighting device 11 is driven to rotate in the same direction as the previous time is raised or lowered in the same direction. Thereafter, imaging and calculation of the evaluation value are performed again.

When V1 ≦ V0, the pulse motor 37
The rotation angle theta is set to half the value of the last angle theta ₁ of the rotating direction is set in the opposite direction (Step 36
7) Return to step 364. The pulse motor 37 is rotationally driven in the direction opposite to the previous one by the newly set angle θ to move the lighting device 11 in the opposite direction, and the image data is taken in again and the evaluation value is calculated.

As described above, every time it is determined that V1 ≦ V0, the rotation angle θ of the pulse motor 37 is sequentially set smaller and the same procedure is repeated while the rotation direction is reversed.
As a result, when the determinations in steps 365 and 366 are YES, the luminaire 11 has been set to the optimal installation height h _opt . This completes the height position adjustment operation of the lighting device.

(9) Edge Image Generating Apparatus An edge image can be generated using the above-mentioned edge-like fitness μ.

The configuration of the image processing device shown in FIG. 48 is adopted as an edge image generation device. Preferably, an image memory 88 for storing the degree of conformity μ for each pixel is further provided.
The nonconformity calculation unit shown in FIG. 49 is also used as it is.
The nonconformity * μ is unnecessary, and only the conformity μ may be output.

FIGS. 65 (A) to 65 (H) show other examples of the fuzzy models representing the edge patterns set in the edge matching degree calculation circuits 111 to 118, respectively. Here, the membership function Z is not used. The membership functions P and N have the shapes shown in FIG.

According to the fuzzy model shown in FIG. 65 (A), the MFC 101, 10
2 and 103 have membership functions P and MFCs 105 and 106
, 107 are set with the membership function N, respectively. A membership function that always outputs the maximum logical value 1 is set in the other MFCs 104 and 108 regardless of the value of the input data.

Similarly, membership functions according to FIGS. 65 (B) to 65 (H) are set in the other edge matching degree calculation circuits 112 to 118.

The fitness μ is calculated for each pixel for all image data constituting one field (or one frame) stored in the image memory 62, and
Stored in 88 corresponding locations. As a result, FIG.
The image representing the edge of the image as shown in FIG.
Is generated as shown in The fitness μ may be converted into binary data using an appropriate threshold, and then stored in the image memory 88.

(10) Optimal lighting condition setting support device (Part 4:
Mainly, setting of luminous intensity of lighting device) An embodiment of a device that sets the optimum lighting condition based on the image data of the object obtained by the imaging again or supports the setting will be described.

FIG. 67 shows the overall configuration of an automatic inspection apparatus having an illumination system in which the optimal illumination conditions have been set by the optimal illumination condition setting support device described in detail later. This automatic inspection device includes an illumination device 10, a camera 20, an image processing device 40, and a transport mechanism 45. The transport mechanism 45 is realized by a transport conveyor, and sequentially transports a large number of inspection objects OB to predetermined inspection positions.

The lighting device 10 and the camera 20 are arranged above the inspection position. The camera 20 is the object OB at the inspection position
Is imaged from directly above, and the image data obtained by the imaging is given to the image processing device 40.

The illumination device 10 is realized by, for example, a halogen lamp and illuminates the inspection object OB at the inspection position from obliquely above. The luminous intensity of the illuminating device 10 is set to an optimal value so that an optimal inspection of the object OB is guaranteed, as described later.

In the case where the inspection of the object OB by the image processing apparatus 40 is a sample inspection for determining whether it is a non-defective product or a defective product having chips or burrs, the accuracy of discriminating non-defective products from non-defective products is maximized The luminous intensity of the lighting device 10 is set so that This is the optimal lighting condition.

The lighting condition is not limited to the luminous intensity, but may be the set position or the set angle of the lighting device 10. Also, the optimal lighting conditions may be set in the image processing device 40,
It may be set in a controller incorporated in the lighting device 10. In any case, the illumination device 10 is controlled by the image processing device 40 or the controller so that the illumination device 10 performs illumination that satisfies the set optimal illumination condition.

The image processing device 40 binarizes the video signal representing the object OB given from the camera 20, and converts
Based on the process of measuring the characteristic amount such as the area of the object using the value image data and comparing the measured value with a predetermined reference value, it is determined whether the object OB is a non-defective item, or whether the object OB is missing or burred. It is determined whether the product is defective. The feature quantity serving as a basis for determination is not limited to the area, but may be a perimeter, a center of gravity, or the like.

FIGS. 68 and 69 show an apparatus for supporting the setting of the optimum lighting conditions. Illumination device 10, camera 20, image processing device 40, display device 41, and the like are denoted by the same reference numerals as those in the inspection device shown in FIG. 67 for simplification. This is because the inspection device and the optimum illumination condition setting support device can be used together. Of course, these two devices may be configured separately.

The camera 20 is located just above the observation position OV and at the same height as that in the automatic inspection apparatus.
The illumination device 10 is disposed obliquely above the observation position OV and at the same height position as that of the automatic inspection device at the same inclination angle. Preferably, the same camera and lighting device are used in the automatic inspection device and the support device.

The support device has a sample supply mechanism. This sample supply mechanism includes a rotary table 120, a table drive mechanism 121, and a detector 123.

The turntable 120 is disk-shaped, and has a non-defective sample OBa, a defective sample OBb with a chip, and a defective sample OB with burrs at the peripheral edge of the upper surface thereof.
c are positioned and fixed at equal angular intervals.

On the rotary table 120, outside the fixed position of each sample, an identification body 122a made of a metal plate,
122b and 122c are fixed. These identifiers 122
a, 122b, 122c are the samples OBa, OBb, OB
This is for causing the detector 123 to detect that c has reached the observation position OV. The installation height of the identifier 122a for the non-defective sample OBa, the installation height of the identifier 122b for the defective sample OBb having a chip, and the installation height of the identifier 122c for the defective sample OBc having a burr are mutually reciprocal. Is different.

The table driving mechanism 121 is provided with
Image processing device including motor connected to center of rotation
The rotary table 120 is intermittently rotated in response to the control signal from the controller 40, and each sample OBa, OBb, OBc is sequentially guided to the observation position OV in the field of view of the camera 120.

The detector 123 detects the non-defective sample OBa that has reached the observation position OV, the defective defective sample OBb having a chip, and the defective sample OBc having burrs. It is located outside.

The detector 123 has three proximity switches 123.
a, 123b, and 123c, and the first proximity switch 123a is an identifier for the non-defective sample OBa 1
At the same height position as the identification object 122a to detect 22a,
Further, the second proximity switch 123b is located at the same height position as the identification body 122b in order to detect the identification body 122b of the defective sample OBb having a chip, and the third proximity switch 123c is located at the same height as the defective sample OBc having burrs. Are installed at the same height position as the identification body 122c to detect the identification body 122c. Each proximity switch 123
The detection signals a, 123b, and 123c provide a detection signal to the image processing device 40 when detecting the identification objects 122a, 122b, and 122c, respectively.

FIG. 70 shows a circuit configuration of the image processing apparatus 40, which includes an A / D converter 131, a binarization processing section 132, and a display device.
41. Synchronous signal generator 133, CPU 130, ROM 135, R
It comprises an AM 134, an external storage device 136 and the like.

An A / D converter 131 converts an analog video signal output from the camera 20 into digital image data. The binarization processing section 132 binarizes the digital image data supplied from the A / D converter 131 with a predetermined threshold value. The display device 41 displays a binarized image represented by the binarized image data. Synchronous signal generator 133
Generates a vertical synchronizing signal, a horizontal synchronizing signal, and other timing signals, and supplies these signals to the A / D converter 131, the binarization processing unit 132, and the CPU.

The CPU 130 is a subject that executes control and calculation for searching for and determining the optimal lighting conditions, receives a detection signal from the detector 123, and
10. Control the operation of the table drive mechanism 121 and the like. RO
A program is stored in M135, and various data are temporarily stored in RAM134. The external storage device 136 is configured by a floppy disk device or the like, and stores the determined optimum lighting conditions of the lighting device 10 and the like.

FIG. 71 shows a processing procedure by the CPU 130 for determining the optimum lighting condition of the lighting device 10.

First, the number N _a of non-defective samples OBa, the number N _b of defective defective samples OBb, and the defective defective sample OBc arranged on the turntable 120 are described below.
The number N _c of is inputted from the keyboard, respectively (step 371). When a start command is given, the image processing device 40 gives a start signal to the table drive mechanism 121.
The rotary table 120 rotates (step 372).

When any of the samples reaches the observation position OV, the first proximity switch 123a is a non-defective sample OBa, and the second proximity switch 123b is a non-defective defective sample OBb. In the case of a certain defective sample OBc, the third proximity switch 123c detects the corresponding identification bodies 122a, 122b, 122c, respectively, and gives a detection signal to the image processing device 40 (step 173).
).

When the detection signal is input, the CPU 130 outputs a stop signal to stop the table driving mechanism 121,
The rotation of the turntable 120 is stopped (step 374). This positions the sample within the field of view of the camera 20 at the observation position OV.

In this embodiment, the luminous intensity of the lighting device 10 is optimized. The illumination device 10 is given an initial luminosity from the image processing device 40, and the illumination device 10 emits light at this initial luminosity,
Illuminating the sample. A sample at the observation position OV is captured by the camera 20 under illumination by the illumination device 10 (step 375). The output video signal of camera 20 is an image processing device
Enter 40.

This video signal is A / D converted and binarized. The CPU 130 measures the feature amount of the sample at the observation position OV using the binary image data (step 37).
6). In this embodiment, the feature amount is the area of the sample on the captured image. The measured features are stored in the RAM 134.

The RAM 134 is provided with areas for separately storing feature values for non-defective products, defective defective products, and defective products having burrs. The CPU 130 determines the observation position OV based on the detection signal input from the detector 123.
It is determined whether the sample is a good product, a defective product with a chip, or a defective product with burrs, and the measured feature amount is stored in a storage area for the determined type of sample.

Steps 372 to 376 are performed on a rotary table.
This is done sequentially for all the samples above 120.
The number of processes is the sum N of the number of samples input earlier.
_a + N _b + N reaches if the turntable to _c 120 will be one rotation (step 377).

When the feature amount (area) measurement is completed for all the samples on the rotary table 120, the RAM 134
Is divided by the number of samples N _a , N _b, or N _c , respectively, to calculate the average value S of the areas of the good-quality samples OBa.
_a , average value S of the area of defective defective sample OBb
_b, and the average value S _c of the area of the defective sample OBc with burr is determined (step 378).

Using these average values S _a , S _b , and S _c , an evaluation value S of the accuracy of discriminating a good product from a defective product is calculated by the following equation (step 379).

[0253]

S = −S _c −S _a | + | S _a −S _b | Equation 11

The larger the evaluation value S, the higher the accuracy with which a good product and a defective product can be determined.

The evaluation value calculation processing shown in FIG. 71 is repeated a plurality of times while changing the initial luminous intensity set in the lighting device 10. The initial luminous intensity is plotted on the horizontal axis, and the calculated evaluation value S is plotted on the vertical axis.
Similarly to the evaluation value V shown in FIG. 55, the evaluation value S is a convex function having one peak. Therefore, the initial luminous intensity giving the maximum value or an evaluation value near the maximum value indicates an illumination condition that can determine a good product and a defective product with the highest accuracy. Such an initial light intensity is determined by the CPU 130 or by the operator.

The optimal lighting conditions (optimal initial luminous intensity) determined in this way are stored in the external storage device 136 together with other necessary data. The optimal lighting conditions are used in the inspection device shown in FIG.

(11) Optimal lighting condition setting support device (part 5:
Mainly setting of luminous intensity of lighting device) In the above embodiment, the optimum lighting condition is set based on the area of the sample. Next, a description will be given of an apparatus for setting or supporting the setting of the optimum lighting condition based on the idea that the lighting condition that minimizes the variance of the feature amount of the sample is optimal.

The illumination system optimized on the basis of the above concept is used for an automatic measurement device that measures features such as the position of the center of gravity, the area, and the perimeter from the image of the object to be measured.
The configuration of this automatic measuring device is the same as that shown in FIG.

FIGS. 72 and 73 show the configuration of the optimum lighting condition setting support device. The same components as those shown in FIGS. 68 and 69 are denoted by the same reference numerals, and redundant description will be avoided.

[0260] The peripheral portion of the upper surface of the disc-shaped rotary table 120, a plurality of samples OB _s are positioned and fixed at equal angular intervals. On the outside of the fixed position of each sample OB _s identification member 122 made of metal plate are arranged at the same height. On the other hand, a detector 123 composed of a proximity switch is arranged outside the turntable 120 near the observation position OV and at the same height as the identification body 122.

The configuration of the image processing device 40 is the same as that shown in FIG. 70, and its illustration and description are omitted.

The processing procedure for determining the optimum illumination condition for measurement is basically the same as that shown in FIG. Hereinafter, points different from the processing in FIG. 71 will be mainly described. In this embodiment, the number of fixed samples OB _s the rotary table 120 on the first step 371 is entered.

[0263] rotating the rotary table 120, the rotary table 12 with any sample OB _s reaches the observation position OV
0 stops, and imaging of the sample OS _s is performed. A predetermined initial luminous intensity is set in the lighting device 10 by the image processing device 40.

The image processing device 40 processes image data taken from the camera 20, measures feature amounts such as the position of the center of gravity and the area, and stores the measured feature amounts.

The above operation is executed for all the samples on the turntable 120.

[0266] When the measurement of all the sample OB _s complete, the dispersion of the characteristic amounts for all samples measured are determined. The smaller the variance is, the smaller the variation of the feature amount is, and the higher the measurement accuracy is.

While the luminosity of the illumination device 10 is changed by the image processing device 40, similar measurement and variance calculation are performed. The luminous intensity that minimizes the obtained variance is searched, and the luminous intensity is determined as the optimal illumination condition.

In each of the above-described two embodiments, a rotary table is provided, but a belt conveyor as shown in FIG. 67 may be used for mounting and transporting a sample.

(12) Imaging Condition Determination Support Apparatus Finally, an apparatus for supporting selection of a lens included in the imaging system, determination of the installation distance of the camera with respect to the object, and the like will be described.

FIG. 74 shows the configuration of the imaging condition determination support device 50. The same components as those shown in FIG. 14 are denoted by the same reference numerals. Since the support device 50 can be realized by using the same image processing device as the support device shown in FIG. 14, the same reference numeral 50 is attached for convenience.

The imaging condition determination support device 50 includes a lens selection unit.
57 are provided. The lens selecting unit 57 guides a lens, a close-up ring, an installation distance, and the like suitable for the information according to information input by the user from the input device 51, for example, information on a desired field of view, a desired installation distance, and what is to be prioritized. .

With reference to FIG. 75, the relationship among various parameters of the imaging system will be described.

The installation distance of the camera 20 (the distance A between the object OB and the camera 20 is given by the following equation.

[0274]

A = (Ls (F + B−h−ΔH + d) / Lc) + B−h + d (12) where Ls: field size F: flange back B: total length of lens 22 h: first from lens tip Distance to principal point ΔH: Principal point interval d: Helicoid extension Lc: Effective screen of image sensor

The helicoid extension amount d is expressed by the following equation using the focal length fo of the lens 22 (combination lens or lens barrel).

[0276]

D = (Lc × fo / Ls) −FB + h + ΔH + fo Equation 13

The installation distance A of the camera is calculated from the expression 13 as Ls / Lc.
Can be derived and substituted into Equation 12 as follows.

[0278]

A = fo [(F + d + B−h−ΔH) / (F + d + B−h−ΔH−fo)] + d + B−h Formula 14

The field of view Ls is derived from Equations 13 and 14 as follows.

[0280]

Ls = Lc (A−D−B + h) / (F + d + B−h−ΔH) Equation 15

Since the flange back F and the effective screen Lc of the image pickup device are constants determined by the camera 20, they may be considered as fixed values. Since the total lens length B, the distance h from the lens selection to the first principal point, the principal point interval ΔH, and the actual focal length fo are constants determined by the lens 22, a set of these constants will be expressed as lens. The constant lens is uniquely determined by the type of lens.

Then, the equations (12) and (13) correspond to the function f ().
The following are concisely expressed by using the symbols.

[0283]

A = f (Ls, d, lens) Equation 16

[0284]

D = f (Ls, lens) Equation 17

By deriving d from Equation 14, the amount of helicoid extension is also expressed as follows.

[0286]

D = f (A, lens) Equation 18

Further, from Expression 15, the size Ls of the field of view is expressed as follows.

[0288]

Ls = f (A, d, lens) Equation 19

Equations 16 to 17 show the relationship among the camera installation distance A, the size Ls of the visual field, the helicoid extension amount d, and the lens constant lens. If at least two of the above four parameters are given using such a relational expression, the remaining two are determined.

A process of selecting an optimal lens type when a desired field size Ls ₀ and a desired installation distance A ₀ are given using the relationship of Expressions 16 to 19 will be described. It is assumed that there are N types of lenses, and the constant lens for these lenses is known in advance. Preferably, this lens constant lens is stored in a memory in advance for each type of lens. The user may input the lens constant lens each time.

FIG. 76 shows the processing in the case where the desired visual field Ls ₀ is prioritized.

First, the user inputs a desired visual field Ls ₀ and a desired installation distance A ₀ from the input device 51 (step 38).
0).

The lens constant lens of the first (n = 1) lens is read out from the memory or input through the input device 51, and using the desired field of view Ls ₀ and the n-th lens constant lens, the equation (17) is obtained. The amount of helicoid delivery d is calculated from the above (steps 381 to 383).

[0294] Since the helicoid feed amount d must be less than the maximum feed amount d _max helicoid was calculated helicoid movement amount d whether greater than the maximum feed amount d _max is determined (Sutebbu 384).

[0295] If, when the calculated helicoid movement amount d is larger than the maximum feed amount d _max is, since it is necessary to use a close-up ring thickness t which satisfies the inequality shown in the following equation 20, equation A close-up ring having a thickness satisfying 20 is selected (step 385). The data relating to the close-up ring is also preferably stored in a memory in advance.

[0296]

[Equation 20] t <d <(t + d _max ) Equation 20

Next, using the calculated helicoid extension amount d, the camera installation distance A is calculated for the lens according to Equation 16 (step 386).

The above processing is repeated for N types of lenses (steps 387 and 388), and the lens number n, close-up ring thickness t and installation distance A obtained in each processing are stored in the memory.

[0299] When the processing of the N types of lenses is completed, from among these lenses, lens number n ₀ with the closest installation distance A to the installation distance A ₀ of the desired is selected (step 289). If there is data on the thickness t of the close-up ring for the selected lens number n ₀ , it is also selected. The calculated camera installation distance A is actually the camera
20 is the distance to be set. Of course, the relationship of B + d <A must be established between the total length B of the lens 22, the helicoid extension amount d, and the camera installation distance A, and it is needless to say that such a lens needs to be selected. Absent.

The type of lens selected in this way, the thickness of the close-up ring, and the installation distance A are displayed on the marking device 41 as shown in FIG.

FIG. 77 shows a process for giving priority to a desired installation distance. The same processes as those shown in FIG. 76 are denoted by the same reference numerals, and redundant description will be avoided.

Using the desired installation distance A ₀ and the lens constant lens of the n-th lens, the amount of helicoid extension d is calculated according to equation (step 383). d>
close-up ring thickness t is calculated in the case of d _{ma x} (step 385).

Using the calculated helicoid extension d, the field of view Ls is calculated according to equation 19 (step 391).

The above processing is repeated for all types of lenses. Thereafter, the lens having the visual field Ls closest to the desired visual field Ls ₀ is selected (step 392).
), And the selection result is displayed as shown in FIG. 78 (step 393).

Finally, a process for obtaining a close-up ring and an installation distance for realizing a desired visual field with a designated lens will be described. No particular flow chart is shown for this process.

In this case, since the lens 22 is known, the desired helicoid extension d is calculated by substituting the desired field of view Ls ₀ and the designated lens constant lens into Equation 17. Since the helicoid extension d must be smaller than the maximum helicoid extension d _max of the lens, if the calculated helicoid extension d is greater than the maximum extension d _max , the thickness satisfying the inequality in Equation 20 is satisfied. A close-up ring of t is used.

Next, using the calculated helicoid extension amount d, the desired visual field Ls ₀ and the lens constant lens,
Thus, the installation distance A is obtained.

In this case as well, the relationship of B + d <A must be established between the total length B of the lens, the helicoid extension amount d, and the camera installation distance A. Therefore, when B+d> A, the designation is made. It means that the desired field of view cannot be realized with the lens that has been set.

As described above, when the desired visual field Ls ₀ and the desired installation distance A ₀ are input from the input device 51, the lens (focal length), the thickness of the close-up ring, and the installation distance are displayed on the CRT as recommended values. Displayed on the device. When a desired field of view and a lens are input, a suitable installation distance A is calculated and displayed.

[Brief description of the drawings]

FIG. 1 is a perspective view showing a ring fluorescent lamp.

FIG. 2 is a perspective view showing a straight tube fluorescent lamp.

FIG. 3 is a perspective view showing a ring light guide.

FIG. 4 is a perspective view showing a slit light guide.

FIG. 5 is a perspective view showing a straight light guide.

FIG. 6 is a perspective view showing a bifurcated light guide.

FIG. 7 shows transmitted light illumination.

FIG. 8 shows uniform light illumination.

FIG. 9 shows coaxial epi-illumination.

FIG. 10 illustrates collimated light illumination.

FIG. 11 illustrates dark field illumination.

FIG. 12 shows an overall configuration of an image processing system.

FIG. 13 illustrates another example of the image processing system.

FIG. 14 is a block diagram illustrating a configuration of a lighting condition determination support device suitable for supporting determination of an optimum height of the lighting device.

FIG. 15 shows a display example of the determined optimum height on the display device.

FIG. 16 is a view for explaining a method for determining an optimum installation height of the lighting device.

FIG. 17 is a view for explaining a method of determining an optimum installation height of the lighting device.

FIG. 18 shows an illumination device and an iso-illuminance curve of the illumination light.

FIG. 19 is a graph showing a relationship between a radius of a region illuminated with equal illuminance and an installation height of a lighting device.

FIG. 20 is a flowchart illustrating a procedure of an optimum height determination process.

FIG. 21 shows still another example of the image processing system.

FIG. 22 is a perspective view showing a shape of an object and a positional relationship between the object and a lighting device.

FIG. 23 is a block diagram showing a configuration of an illumination condition determination support device suitable for supporting an optimum angle determination.

FIG. 24 illustrates a method for determining an optimum installation angle of a lighting device.

FIG. 25 shows a state of incomplete diffuse reflection.

FIG. 26 is a flowchart illustrating a procedure of an optimum angle determination process.

FIG. 27 is a block diagram illustrating a configuration of a lighting condition determination support device suitable for supporting determination of a lighting device and a lighting method.

FIG. 28 is a perspective view showing the appearance of a connector as an example of an object.

FIG. 29 is a flowchart showing an overall procedure of a process of determining a lighting device and a lighting method.

FIG. 30 is a flowchart showing a processing procedure for determining a lighting device and a lighting method when the shape and dimensions are to be inspected.

FIG. 31 is a flowchart showing a processing procedure for determining a lighting device and a lighting method when the purpose is to inspect the presence or absence of a part or a missing part.

FIG. 32 is a flowchart showing a processing procedure for determining an illuminating device and an illuminating method for the purpose of flaw inspection.

FIG. 33 is a flowchart showing a processing procedure for determining a lighting device and a lighting method when the purpose is to inspect a color.

FIGS. 34A and 34B are flowcharts showing a processing procedure for determining an illuminating device and an illuminating method for the purpose of positioning.

FIG. 35 is a flowchart showing a processing procedure for determining an illuminating device and an illuminating method for the purpose of inspecting a foreign substance.

FIG. 36 is a flowchart showing a processing procedure for determining an illuminating device and an illuminating method for the purpose of inspecting characters and marks.
It is a chart.

FIG. 37 shows a display example of a screen for transmitted light illumination.

FIG. 38 shows a display example of a screen for uniform light illumination.

FIG. 39 shows a display example of an uneven projection illumination screen.

FIG. 40 shows a display example of an uneven hole illumination screen.

FIG. 41 shows a display example of an uneven edge illumination screen.

FIG. 42 shows a display example of an uneven line illumination screen.

FIG. 43 shows a display example of a regular reflection utilizing illumination screen.

FIG. 44 is a block diagram illustrating a configuration of a device suitable for determining a lighting device and a lighting method, and supporting determination of an optimum height or an optimum angle.

FIG. 45 is a flowchart illustrating the entirety of a lighting device, a lighting method, and a process of determining an optimum height or an optimum angle.

FIG. 46 is a flowchart showing a procedure for determining an optimum height when there are a plurality of lighting devices.

FIG. 47 is a block diagram illustrating a circuit configuration of an image processing device in the optimal lighting condition setting support device.

FIG. 48 is a block diagram illustrating a circuit configuration of a histogram generation circuit.

FIG. 49 is a block diagram illustrating a circuit configuration of a nonconformity calculation unit.

FIG. 50 shows a window set on an image.

FIG. 51 is for describing the operation of a subtraction circuit that converts brightness data into data representing a brightness difference.

FIGS. 52A to 52H show fuzzy models representing edge patterns.

FIG. 53 is a graph showing a membership function.

FIG. 54 is a graph showing an image quality histogram.

FIG. 55 is a graph showing the relationship between the installation height of the lighting device and the evaluation value of the lighting.

FIG. 56 is a flowchart showing a setting procedure of a height position of the lighting device.

FIG. 57 is a flowchart showing a setting procedure of a height position of the lighting device.

FIG. 58 shows a display example of a display screen for instructing height setting.

FIG. 59 shows a display example of a display screen for instructing height setting.

FIG. 60 shows a display example of a display screen for instructing height setting.

FIG. 61 is a flowchart showing another example of the height position setting processing of the lighting device.

FIG. 62 is a flowchart showing another example of the height position setting processing of the lighting device.

FIG. 63 is a flowchart showing yet another example of the height position setting processing of the lighting device.

FIG. 64 is a flowchart showing still another example of the height position setting processing of the lighting device.

FIGS. 65A to 65H show other fuzzy models representing edge patterns.

FIGS. 66A and 66B show how an edge image is generated from a gray image.

FIG. 67 shows an overall configuration of an automatic inspection device.

FIG. 68 shows the overall configuration of an optimal lighting condition setting support device.

FIG. 69 is a plan view showing an arrangement state of samples on the rotary table.

FIG. 70 is a block diagram illustrating a circuit configuration example of an image processing apparatus.

FIG. 71 is a flowchart showing a processing procedure for determining an optimum illumination condition of the illumination device.

FIG. 72 shows an overall configuration of an optimal lighting condition setting support device.

FIG. 73 is a plan view showing an arrangement state of samples on a rotary table.

FIG. 74 is a block diagram illustrating a configuration of an imaging condition determination support device.

FIG. 75 is for describing parameters of a camera, a lens, and the like.

FIG. 76 is a flowchart showing a processing procedure for supporting determination of an imaging condition.

FIG. 77 is a flowchart showing another processing procedure for supporting determination of an imaging condition.

FIG. 78 is a display example showing a selected imaging condition.

[Explanation of symbols]

 11 Illumination device 20 TV camera 30 Post 31 Supporting device 38 Lifting mechanism 41 Display device 61 Image input unit 65 CPU 68 I / O control unit 82 Non-conformity histogram generation unit 83 Brightness histogram generation unit 84 Image quality histogram generation unit

──────────────────────────────────────────────────続 き Continuation of the front page (31) Priority claim number Japanese Patent Application No. 4-171613 (32) Priority date June 4, 1992 (1992.6.4) (33) Priority claim country Japan (JP) (72) Inventor Shiro Fujieda 10 Omron Todo-cho, Ukyo-ku, Kyoto City, Kyoto Prefecture (56) References JP-A-62-194589 (JP, A) JP-A-1-168169 (JP, A) JP-A-2-69080 (JP, A) JP-A-2-187651 (JP, A) JP-A-3-22181 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06T 1 / 00,7 / 00

Claims (13)

(57) [Claims] An image pickup apparatus (20) for picking up an image of an object and outputting a video signal representing a picked-up image, and A / D conversion means (61) for converting the video signal output from the image pickup apparatus into digital image data. ), An image memory (62) for storing A / D converted digital image data for one screen, setting a window in the image data for one screen, and extracting pixel data for a plurality of pixels in the window Window means (70 to 76, 78) for generating data relating to the inclination of the image data in the window (91 to 98); means for calculating the degree of coincidence between the data relating to the inclination and a plurality of edge patterns, respectively (101-109, 111-11
8) means for calculating the evaluation value of the likelihood of the edge from the above matching degree (85), and means for controlling the calculation of the evaluation value of the likelihood of the edge so as to be performed for all the pixels of one screen (65, 80). An edge image generation device comprising: 2. The apparatus according to claim 1, further comprising an edge image memory (88) for storing the edge-likeness evaluation value. 3. The apparatus according to claim 1, further comprising means for level-discriminating the evaluation value of the likelihood of the edge with a predetermined threshold value to binarize the value. 4. An object is imaged by an imaging device (20),
This is a method of generating data representing an edge image of a captured image based on a video signal representing a captured image obtained as described above. A / D conversion of a video signal for one screen into digital image data Set a window in the image data, extract pixel data for a plurality of pixels in the window, generate data related to the slope of the image data in the window, and match the data related to the slope with a plurality of edge patterns. , Calculating the evaluation value of the likelihood of the edge from the degree of matching, calculating the evaluation value of the likelihood of the edge for all pixels of one screen, and obtaining edge image data composed of the above evaluation values. Image generation method. 5. A sample supply device (120, 120) for sequentially supplying a plurality of non-defective and defective samples to a predetermined observation position.
121), an illuminating device (10) that illuminates the sample at the observation position with variable illumination conditions (10), and images the non-defective and defective samples brought to the observation position by the supply device under illumination by the illuminating device. An imaging device that outputs a video signal representing a sampled image; a feature amount measurement unit that measures a feature amount of each sample based on a video signal output from the imaging device; Means (130) for calculating an evaluation value relating to the accuracy of discriminating between a good product and a defective product based on the feature value measured by the above-described method, and an evaluation value for each lighting condition obtained by changing the lighting condition of the lighting device. A lighting condition setting support device comprising: means (130) for searching for the lighting condition having the best evaluation value. 6. A detection device (12) for detecting the presence or absence of a sample supplied to the observation position and distinguishing between a good product and a defective product.
The device of claim 5, further comprising: 3, 123a, 123b, 123c). 7. The sample supply device according to claim 5, wherein said sample supply device comprises a rotary table (120) on which each sample is placed, and a driving device (121) for rotating said rotary table (120). The described device. 8. An illuminating device (10) for illuminating a predetermined observation position, the illumination condition of which is variable, and a plurality of non-defective and defective samples are sequentially supplied to the observation position. In addition, each sample of the defective product is imaged under the illumination by the above-mentioned lighting device, and the characteristic amount of each sample is measured based on a video signal obtained by the imaging. A lighting condition setting support method for calculating an evaluation value for the lighting device, obtaining an evaluation value for each lighting condition while changing the lighting condition of the lighting device, and searching for a lighting condition having the best evaluation value. 9. A sample supply device (120, 121) for sequentially supplying a plurality of samples to a predetermined observation position, an illumination device (10) for illuminating a sample at the observation position, and an illumination condition variable, An imaging device (20) for imaging each sample brought to the position under the illumination of the illumination device and outputting a video signal representing the sampled image; A feature quantity measuring means (130) for measuring a feature quantity; a means (130) for calculating a variance of the feature quantity measured by the feature quantity measuring means; and a lighting condition obtained by changing lighting conditions of the lighting device. Based on the variance of
A lighting condition setting support device comprising: means (130) for searching for lighting conditions that minimize variance. 10. The device according to claim 9, further comprising a detection device (123) for detecting a sample supplied to the observation position. 11. The apparatus according to claim 9, wherein said sample supply device comprises a rotary table (120) on which each sample is placed, and a driving device (121) for rotating said rotary table (120). The described device. 12. An illuminating device (10) for illuminating a predetermined observation position, the illumination condition of which is variable, is arranged, a plurality of samples are sequentially supplied to the observation position, and each sample supplied to the observation position is illuminated by the illumination device. Imaging under the illumination by the camera, measuring the feature amount of each sample based on the video signal obtained by the imaging, calculating the variance of the measured feature amount of all the samples, and changing the lighting conditions of the lighting device while changing the lighting conditions. A lighting condition setting support method that finds the variance for each condition and searches for the lighting condition that minimizes the variance. 13. In order to determine photographing conditions of a camera having a lens and a necessary close-up ring which can be detachably mounted, information on a field of view of the camera, information on a distance of the camera from an object to be photographed, information on a lens, and information on a close-up ring are provided. An input device for inputting at least two pieces of information, an arithmetic unit for calculating remaining information using at least two pieces of information input from the input device, and the residue represented by a calculation result by the arithmetic unit And a display device for displaying the information of the photographing condition.