JP6917780B2 - Image inspection equipment - Google Patents

Description

The present invention relates to an image inspection apparatus that inspects a defect of an inspection object by using an image obtained by capturing an image of the inspection object.

Conventionally, the surface of the inspection object is irradiated with light from the illumination unit, the reflected light is received by the imaging unit, the brightness distribution of the received light is analyzed, and defects such as scratches existing on the inspection object are present. An image inspection device for inspecting light is known (see, for example, Patent Documents 1 to 3).

As an image inspection apparatus of this type, an image inspection apparatus to which the so-called deflation metric principle disclosed in Patent Document 1 is applied is known. This image inspection device includes an illumination unit that irradiates an inspection object with a pattern light whose illuminance changes periodically, and an image pickup unit that images an inspection object that is irradiated with the pattern light. A plurality of pattern lights whose phases of the illuminance distribution are shifted are sequentially illuminated by the illuminating unit, and each time the pattern light is irradiated, the inspection object is imaged by the imaging unit, and the inspection object is imaged based on the obtained plurality of luminance images. Generates phase data showing the shape of. An inspection image is generated based on the obtained phase data, and by using this inspection image, a defect inspection related to the shape of the inspection object can be performed.

Further, Patent Document 2 measures a three-dimensional shape by using a phase shift method.

Further, in addition to the plurality of pattern lights whose phases are shifted, a plurality of pattern lights having different wavelengths may be sequentially irradiated to image an inspection object to generate an inspection image.

Further, as another image inspection device, an image inspection device using the so-called photometric stereo method disclosed in Patent Document 3 is also known. This image inspection device inspects three or more lighting units for illuminating an object to be inspected from different directions, a lighting control unit that sequentially lights each lighting unit, and a timing at which each lighting unit irradiates light. It is equipped with an imaging unit that images an object from a certain direction. Then, a normal vector can be calculated from the pixel values of each of the plurality of captured images by using the photometric stereo method, and an inspection image can be generated based on the normal vector.

Furthermore, high dynamic composition (HDR), which realizes a wide dynamic range by capturing multiple images with different brightness and then synthesizing them, is known, and it is also possible to generate an inspection image using this HDR. It is done.

U.S. Pat. No. 6,100,990 JP-A-2002-257528 Japanese Unexamined Patent Publication No. 2015-232486

As described above, as a method of obtaining an inspection image, a method of irradiating a plurality of phase-shifted pattern lights and taking an image each time, and a method of irradiating a plurality of pattern lights having different wavelengths and taking an image each time. , There are a method of irradiating light from different directions three times or more and taking an image each time, a method of taking a plurality of images having different brightness and synthesizing them, and the like. It is common in that.

In such a method in which the number of illuminations and the number of imagings are a plurality of times, the following problems may occur. For example, when an inspection image is obtained using a line camera, the phase is changed in the X direction in one line, which is the imaging range of the line camera, in order to execute the deflection methylation processing of Patent Document 1. It is necessary to irradiate multiple pattern lights and multiple pattern lights whose phases are changed in the Y direction, but in order to meet the demand for higher resolution in recent years, the screen size is 8K (8192 columns x 1024 rows). Attempt to do so, the number of times of lighting of 1024 times × pattern light becomes the number of times of lighting required to obtain an inspection image, and the number of times of lighting is at least several thousand times.

Here, the illumination operation of the illumination unit and the imaging operation of the imaging unit are synchronized only at the start of imaging of the inspection image by using an encoder pulse as in Patent Document 2, for example. Assuming that light emission and imaging are performed independently for each, the deviation between the illumination timing and the imaging timing accumulates in proportion to the number of imaging times. That is, in making a circuit, it is inevitable that a timing shift of about 10 ns occurs per illumination and imaging. For example, if light emission and imaging are repeated thousands of times in order to obtain an inspection image, Eventually, a timing shift of about several tens of μs will occur. This can be an obstacle to further speeding up inspections.

On the other hand, it is conceivable to keep the lighting unit continuously lit. If the illumination unit is continuously lit, the problem that the illumination timing and the imaging timing deviate from each other is fundamentally solved, but there is a problem that the illumination becomes dark. Therefore, it is necessary to increase the brightness by lighting the flash only at the moment of imaging, instead of continuously lighting the lighting unit. Further, since the light emission pattern cannot be switched by the means of continuous lighting, it is not possible to realize the deflation metrics processing or the multispectral synthesis processing.

The present invention has been made in view of this point, and an object of the present invention is a deviation between the illumination timing and the imaging timing even when the number of illuminations and the number of imagings for obtaining an inspection image are large. The purpose is to realize a further speedup of inspection by making it extremely small.

In order to achieve the above object, in the present invention, in an image inspection device that inspects a defect of an inspection object by using an image of an inspection object moving in one direction, a pattern light having a periodic illuminance distribution is used. It has a pattern light illuminance unit for irradiating the inspection object and a plurality of light receiving elements arranged in a line, and can be installed so that the arrangement direction of the light receiving elements is orthogonal to the moving direction of the inspection object. Generates a plurality of pattern lights in which the phase of the illuminance distribution is shifted to at least one of the arrangement direction of the light receiving elements and the movement direction of the inspection target, and sequentially irradiates the inspection target. Then, based on the pattern light illumination unit, the image control unit that controls the image pickup unit, and the principle of deflection metric so that the inspection object is imaged at the timing when the pattern light is irradiated to obtain a plurality of illuminance images. An image generation unit that generates phase data on the surface of the inspection object from the plurality of illuminance images captured by the imaging unit, and generates an inspection image showing the shape of the inspection object based on the phase data, and the above. It is provided with an inspection unit that executes defect inspection of the inspection object based on the inspection image generated by the image generation unit, and the imaging control unit is configured to be able to receive one encoder pulse signal from the outside. When the encoder pulse signal is received, at least one of the illumination condition of the pattern light illumination unit and the imaging condition of the imaging unit is changed to generate a plurality of brightness images with respect to the imaging unit. It can be configured to sequentially transmit the imaging trigger signals of.

According to this configuration, when the imaging control unit receives the encoder pulse signal, a plurality of imaging trigger signals are sequentially transmitted to the imaging unit. When the image pickup unit receives the image pickup trigger signal, the image pickup unit starts image pickup, and the illumination unit is controlled by the image pickup control unit, whereby a plurality of luminance images are generated. Since a plurality of imaging trigger signals are sequentially transmitted to the imaging unit, the deviation of the imaging timing is eliminated each time the imaging trigger signal is received. Therefore, even if the number of illuminations and the number of imagings increase when a plurality of luminance images are generated, the deviation between the imaging timing and the illumination timing can be maintained in an extremely small state.

Further, when the image pickup control unit receives one encoder pulse signal from the outside, the image pickup control unit is configured to sequentially transmit a plurality of illumination trigger signals to the pattern light illumination unit in synchronization with the image pickup trigger signal. Can be done.

Further, the image pickup control unit can be configured to receive a single encoder pulse signal from the outside and sequentially transmit a plurality of image pickup trigger signals to the image pickup unit in synchronization with the illumination trigger signal.

Further, the pattern light illumination unit is configured to irradiate the inspection object with pattern light having a periodic illuminance distribution, and the imaging control unit sets the phase of the illuminance distribution to the arrangement direction of the light receiving elements and the inspection target. A plurality of pattern lights shifted in at least one direction of the moving direction of the object are generated and sequentially irradiated to the inspection object, and the inspection object is imaged at the timing when the pattern light is irradiated to obtain a plurality of luminance images. As described above, the pattern light illumination unit and the image pickup unit can be configured to be controlled.

Further, the pattern light illumination unit is configured to irradiate the inspection object with light from different directions, and the imaging control unit sequentially irradiates the inspection object with light from different directions. The pattern light illumination unit and the image pickup unit can be configured to control the pattern light illumination unit and the image pickup unit so as to image an inspection object at the timing of irradiation of the light and obtain a plurality of luminance images.

Further, the pattern light illumination unit is configured to irradiate the inspection object with light having a different wavelength, and the imaging control unit irradiates the inspection object with light having a different wavelength, and the timing at which the light is irradiated. The pattern light illumination unit and the image pickup unit can be controlled so as to obtain a plurality of luminance images by imaging the inspection target object.

Further, the image pickup control unit is configured to control the pattern light illumination unit and the image pickup unit so as to obtain a plurality of luminance images having different brightness, and the image generation unit has a plurality of luminances having different brightness. The images can be combined to generate an inspection image having a dynamic range wider than the dynamic range of each luminance image.

Further, a lighting condition storage unit for storing the lighting conditions of the pattern light illumination unit is incorporated in the pattern light illumination unit, and when the pattern light illumination unit receives the illumination trigger signal, the illumination condition storage unit stores the illumination conditions. The inspection object can be illuminated according to the stored illumination conditions.

According to the present invention, since a plurality of imaging trigger signals are sequentially transmitted to the imaging unit, the illumination timing and the imaging timing can be determined even when the number of illuminations and the number of imagings for obtaining an inspection image are large. The deviation can be made extremely small, which can further speed up the inspection.

It is a figure which shows typically the operation state of the image inspection apparatus which concerns on Embodiment 1. FIG. It is a block diagram of the image inspection apparatus which concerns on Embodiment 1. FIG. It is a front view of the light emitting part of the pattern light illumination part. It is a figure which shows the 1st light emitting diode row. It is a figure which shows the 2nd light emitting diode row. It is a figure which shows the irradiation state of four kinds of Y direction pattern light which changed the phase of an illuminance distribution by 90 °. It is a figure which shows the irradiation state of four kinds of X-direction pattern light which changed the phase of an illuminance distribution by 90 °. It is a figure which shows the positional relationship between the image pickup part and the pattern light illumination part when the work is a columnar member. It is a figure which shows the positional relationship between the image pickup part, and the pattern light illumination part when the work is a member having translucency. It is a figure which shows the interface for selection of a lighting installation method. It is a figure which shows the interface for confirming the top and bottom of a camera. It is a figure which shows the interface for camera posture confirmation. It is a figure which shows the interface for selecting the work moving direction. It is a figure which shows the interface for selection of an illumination direction. It is a figure which shows the interface for selecting the cable pull-out direction. It is a figure which shows the 1st pattern light and the 2nd pattern light which irradiate when acquiring the positional relationship information of a light receiving element and a pattern light illumination part. It is a figure which shows the 1st image and the 2nd image at the time of arranging the pattern light illumination part so that a signal line comes out to the left. It is a figure which shows the 1st image and the 2nd image at the time of arranging the pattern light illumination part so that a signal line comes out upward. It is a figure which shows the 1st image and the 2nd image at the time of arranging the pattern light illumination part so that a signal line comes out downward. It is a figure which shows the 1st image and the 2nd image at the time of arranging the pattern light illumination part so that a signal line comes out to the right. It is a figure explaining the procedure of acquiring the positional relationship information of a light receiving element and a pattern light illumination part when an area camera is used. It is a figure which shows the image pickup part and the pattern light illumination part which concerns on other form of the information acquisition part. It is a figure which shows the interface for direction selection. It is a flowchart which shows the generation procedure of the inspection image. It is a figure which shows typically the procedure of generating the inspection image which applied the principle of deflation metrics. It is a figure which shows the example of the actual luminance image. It is a figure which shows typically the concept of simple defect extraction. It is a figure which shows the interface for defect extraction. It is a figure corresponding to FIG. 25 when a part of the inspection image is enlarged and displayed. It is a figure corresponding to FIG. 25 in the case of performing forced extraction of a defective part. It is a figure which shows the example which displayed the histogram and the pin at the time of setting a threshold value. It is a figure which shows the example of the inspection image. It is a table which shows the suitability of a filter for each inspection image. It is a figure which shows the interface for parameter setting of a filtering process. It is a figure which shows the change of the inspection image before and after the filter processing. It is a flowchart when the background selection is performed a plurality of times. It is a figure which shows the example of the inspection image in the case where the positional relationship between the pattern light illumination unit and the image pickup unit is correct and incorrect. It is a time chart which shows the relationship between the image pickup trigger signal and the illumination trigger signal, and an encoder pulse. It is a figure which shows a part of the circuit structure of the current control part of the pattern light illumination part. It is a figure explaining the current value control method by the current control part. It is a block diagram of the image inspection apparatus which concerns on Embodiment 2. FIG. It is a block diagram of the image inspection apparatus which concerns on Embodiment 3. It is a flowchart which shows the edge generation method. This is an inspection image showing a state in which the pointer is placed on the edge. It is an inspection image which shows the state which displayed the initial coordinates. It is a figure which shows typically the method of forming a segment and finding an edge position. It is a figure which shows typically the method of projecting the pixel in a segment, and finding the edge position. It is a figure which shows the state which executed the connection process and specified the contour. It is a figure which shows the connection processing method when two contours are close to each other. It is a figure which shows another connection processing method when two contours are close to each other. It is a figure which shows the processing method when the curvature of the contour is small. It is a figure which shows the case where the inspection object is a bead or the like. It is a flowchart which shows the contour identification method of both sides. It is a figure which shows typically the method of obtaining the control point P0 in the contour specifying method on both sides. It is a figure which shows typically the method of forming a segment and finding an edge position in the contour specifying method on both sides. It is a block diagram of the image inspection apparatus which concerns on Embodiment 4. FIG. It is a figure explaining the procedure which makes a part of a work into an inspection target area. It is a figure explaining the procedure which makes almost all of the workpiece an inspection target area. It is a figure explaining the procedure of taking a plurality of partial inspection images and setting an inspection target area. It is a figure corresponding to FIG. 2 which concerns on modification 1 of embodiment. FIG. 2 is a view corresponding to FIG. 2 according to a modified example 2 of the embodiment. FIG. 2 is a view corresponding to FIG. 2 according to a modified example 3 of the embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. It should be noted that the following description of the preferred embodiment is essentially merely an example and is not intended to limit the present invention, its application or its use.

(Embodiment 1)
(Overall configuration of image inspection device 1)
FIG. 1 schematically shows an operating state of the image inspection apparatus 1 according to the embodiment of the present invention. The image inspection device 1 is configured to inspect defects in the work W using an image obtained by capturing an image of the work W (object to be inspected) moving in one direction. Specifically, a plurality of pattern lights are applied to the work. It includes at least a pattern light illumination unit 2 for executing pattern light illumination to irradiate W, an image pickup unit 3, a control unit 4, a display unit 5, a keyboard 6, and a mouse 7, but other than these. It can also be equipped with equipment. For example, an external control device 8 including a programmable logic controller (PLC) or the like is connected to the control unit 4. The external control device 8 may form a part of the image inspection device 1. Further, the external control device 8 does not have to be a component of the image inspection device 1.

FIG. 1 shows a case where a plurality of works W are transported in the direction indicated by the white arrow in FIG. 1 in a state of being placed on the upper surface of the transport belt conveyor B. Work W is an object to be inspected. The external control device 8 is a device for sequence control of the transport belt conveyor B and the image inspection device 1, and a general-purpose PLC can be used.

In the description of this embodiment, the transport direction of the work W by the transport belt conveyor B (moving direction of the work W) is the Y direction, and the direction orthogonal to the Y direction in the plan view of the transport belt conveyor B is the X direction. The direction orthogonal to the X direction and the Y direction (the direction orthogonal to the upper surface of the transport belt conveyor B) is defined as the Z direction, but this is defined only for convenience of explanation.

The image inspection device 1 can be used for visual inspection of the work W, that is, for inspecting the surface of the work W for defects such as scratches, stains, and dents. During its operation, the image inspection device 1 receives an inspection start trigger signal that defines the start timing of defect inspection from the external control device 8 via a signal line. The image inspection device 1 performs imaging, illumination, and the like of the work W based on the inspection start trigger signal, and after a predetermined process, obtains an inspection image. After that, the inspection result is transmitted to the external control device 8 via the signal line. As described above, during the operation of the image inspection device 1, the input of the inspection start trigger signal and the output of the inspection result are repeatedly performed between the image inspection device 1 and the external control device 8 via the signal line. As described above, the input of the inspection start trigger signal and the output of the inspection result may be performed via the signal line between the image inspection device 1 and the external control device 8, or other signals (not shown). It may be done via a line. For example, the sensor for detecting the arrival of the work W and the image inspection device 1 may be directly connected, and the inspection start trigger signal may be input from the sensor to the image inspection device 1.

In addition to being configured with dedicated hardware, the image inspection device 1 may be configured by installing software on a general-purpose device, for example, installing an image inspection program on a general-purpose or dedicated computer. In the following example, a case where the image inspection program is installed on a dedicated computer in which hardware such as a graphic board is specialized for image inspection processing will be described.

(Structure of pattern light illumination unit 2)
The pattern light illumination unit 2 is for irradiating the work W with pattern light having a periodic illuminance distribution, for example, in a plurality of light emitting diodes, a liquid crystal panel, an organic EL panel, a digital micromirror device (DMD), or the like. It can be configured and can be simply called an illumination unit. Although the liquid crystal panel, the organic EL panel, and the DMD are not shown, those having a conventionally known structure can be used. The pattern light illumination unit 2 is connected to the control unit 4 via a signal line 100a, and can be installed apart from the image pickup unit 3 and the control unit 4.

When a plurality of light emitting diodes are used for the pattern light illumination unit 2, the plurality of light emitting diodes can be arranged in a dot matrix to generate pattern light having a periodic illuminance distribution by controlling the current value. In the case of a liquid crystal panel and an organic EL panel, by controlling each panel, the light emitted from each panel can be made into a pattern light having a periodic illuminance distribution. In the case of a digital micromirror device, it is possible to generate and irradiate a pattern light having a periodic illuminance distribution by controlling the built-in micromirror surface. The configuration of the pattern light illumination unit 2 is not limited to that described above, and any device or device capable of generating pattern light having a periodic illuminance distribution can be used.

Hereinafter, a case where a plurality of light emitting diodes are used for the pattern light illumination unit 2 will be described in detail. In this case, the pattern light illumination unit 2 irradiates from the light emitting unit 23 (shown in FIG. 3) formed by mounting a plurality of light emitting diodes 20 and 21 arranged two-dimensionally on the substrate 23a and the light emitting diodes 20 and 21. It includes a diffusing member 24 (shown in FIGS. 1, 5A, and 5B) for diffusing the generated light. The light emitting unit 23 has a substantially rectangular shape as a whole. The light emitting surface of the light emitting unit 23 is a virtual division line J1 extending in the Y direction through the central portion of the light emitting unit 23 in the X direction and a virtual division line extending in the X direction through the central portion of the light emitting unit 23 in the Y direction. It can be virtually divided into the first to fourth regions S1 to S4 by J2. The first to fourth regions S1 to S4 are all substantially rectangular shapes having the same size. Each area S1 to S4 can also be called a block, and one block consists of 12 LEDs × 12 rows × 2 directions.

In the first region S1, as shown in FIG. 4A, there are 12 rows of first light emitting diode rows A1 to A12 composed of a plurality of light emitting diodes 20 arranged in series at equal intervals in the Y direction and connected in series. As shown in 4B, 12 rows of second light emitting diode rows B1 to B12 composed of a plurality of light emitting diodes 21 arranged in series at equal intervals in the X direction are formed. As shown in FIG. 3, the first light emitting diode rows A1 to A12 in the first region S1 are arranged in the X direction and are arranged parallel to each other. Electric power is individually supplied to the first light emitting diode rows A1 to A12.

As shown in FIG. 3, the second light emitting diode rows B1 to B12 in the first region S1 are arranged in the Y direction and are arranged parallel to each other. Electric power is individually supplied to the second light emitting diode rows B1 to B12.

The first light emitting diode rows A1 to A12 and the second light emitting diode rows B1 to B12 are arranged so as to intersect each other in the first region S1. The light emitting diodes 20 forming the first light emitting diode rows A1 to A12 and the light emitting diodes 21 forming the second light emitting diode rows B1 to B12 are arranged so as not to overlap with each other in the light irradiation direction. As a result, the plurality of light emitting diodes 20 and 21 are arranged in a dot matrix in the first region S1.

Since 12 light emitting diodes 20 and 21 are connected in series to form each row A1 to A12 and B1 to B12, the control line is provided in each row without providing a control line for each of the light emitting diodes 20 and 21. Only one set is needed. Further, in each region of the second region S2, the third region S3, and the fourth region S4, the first light emitting diode rows A1 to A12 and the second light emitting diode rows B1 to B12 intersect with each other in the same manner as in the first region S1. Is located in. That is, in this embodiment, while increasing the number of the light emitting diodes 20 and 21 and arranging the light sources of the light emitting unit 23 precisely, the light emitting diode train to be controlled is the first light emitting diode train A1 to the first region S1. A12 and second light emitting diode rows B1 to B12, first light emitting diode rows A1 to A12 and second light emitting diode rows B1 to B12 in the second region S2, and first light emitting diode rows A1 to A12 in the third region S3. It is sufficient to control a total of 96 rows (12 rows × 2 × 4 regions) of the second light emitting diode rows B1 to B12, the first light emitting diode rows A1 to A12 in the fourth region S4, and the second light emitting diode rows B1 to B12. ..

The second light emitting diode rows B1 to B12 in the first region S1 and the second light emitting diode rows B1 to B12 in the second region S2 are arranged so as to be aligned on the same straight line extending in the X direction, respectively. The second light emitting diode rows B1 to B12 in the third region S3 and the second light emitting diode rows B1 to B12 in the fourth region S4 are arranged so as to be aligned on the same straight line extending in the X direction, respectively.

Further, the first light emitting diode rows A1 to A12 in the first region S1 and the first light emitting diode rows A1 to A12 in the third region S2 are arranged so as to be aligned on the same straight line extending in the Y direction, respectively. .. The first light emitting diode rows A1 to A12 in the second region S2 and the first light emitting diode rows A1 to A12 in the fourth region S4 are arranged so as to be aligned on the same straight line extending in the Y direction, respectively.

The diffusion member 24 shown in FIGS. 5A and 5B is made of a translucent plate material arranged so as to cover all of the light emitting diodes 20 and 21 of the first to fourth regions S1 to S4. A well-known plate material can be used as the plate material constituting the diffusion member 24, and the incident light can be diffused and emitted.

By individually controlling the current values flowing through the first light emitting diode rows A1 to A12 and the second light emitting diode rows B1 to B12 in the first to fourth regions S1 to S4, the illuminance can be increased in the Y direction as shown in FIG. 5A. It is possible to generate a Y-direction pattern light that changes and has a uniform illuminance in the X direction, and an X-direction pattern light that changes in the X direction and has a uniform illuminance in the Y direction as shown in FIG. 5B. .. Since the light emitting diodes 20 and 21 are arranged in a dot matrix as described above, the light sources are arranged in a dot shape in the light emitting unit 23. However, by providing the diffusion member 24, each light emitting diode is provided. The light of 20 and 21 is diffused and irradiated to the outside, and as shown in each figure, it is possible to obtain a pattern light in which the illuminance gradually changes when viewed from the outside.

Further, as the control of the light emitting diodes 20 and 21, for example, there are PWM control and control by the current value (current value control), but the current value control is particularly preferable. By controlling the current value, it is possible to realize a quick illuminance change that fully utilizes the speed of the response speed of the light emitting diodes 20 and 21, for example, the time interval for switching one pattern light to another pattern light is 2μ. It can be as short as a second.

By passing a current having the same current value through all the light emitting diodes 20 and 21, it is possible to irradiate light having a uniform illuminance distribution in the plane. When the current values flowing through all the light emitting diodes 20 and 21 are changed in the same manner, the light emitting state can be changed from the dark surface emitting state to the bright surface emitting state.

In the case of the Y-direction pattern light shown in FIG. 5A, since the light and darkness changes in the Y direction, it can be expressed as a pattern light in which the striped pattern is repeated in the Y direction. When generating the Y-direction pattern light, by shifting the phase of the illuminance distribution in the Y direction, it is possible to generate a plurality of Y-direction pattern lights having different phases of the illuminance distribution. The illuminance distribution of the Y-direction pattern light can also be represented by a waveform similar to the sin waveform. In this case, for example, the phase is changed by 90 °, and the Y-direction pattern light in the case of 0 ° and the Y-direction in the case of 90 °. It is possible to generate a pattern light, a Y-direction pattern light in the case of 180 °, and a Y-direction pattern light in the case of 270 °.

Further, in the case of the X-direction pattern light shown in FIG. 5B, since the light and darkness changes in the X direction, it can be expressed as a pattern light in which the striped pattern is repeated in the X direction. When generating the X-direction pattern light, by shifting the phase of the illuminance distribution in the X direction, it is possible to generate a plurality of X-direction pattern lights having different phases of the illuminance distribution. The illuminance distribution of the X-direction pattern light can also be represented by a waveform similar to the sin waveform. In this case, for example, the phase is changed by 90 °, and the X-direction pattern light in the case of 0 ° and the X-direction in the case of 90 °. It is possible to generate a pattern light, an X-direction pattern light in the case of 180 °, and an X-direction pattern light in the case of 270 °. That is, the pattern light illumination unit 2 can illuminate the work W in different illumination modes.

When performing the deflection metering process described later, the pattern light to irradiate the work W can be not only a sine waveform but also a pattern light such as a triangular wave.

Further, the length of one side of the light emitting unit 23 can be set to, for example, 100 mm. 24 light emitting diodes 20 and 21 are arranged vertically and 24 horizontally inside a 100 mm square. By controlling the current values of the light emitting diodes 20 and 21, high-speed and high-gradation pattern light can be generated. The length of one side of the light emitting unit 23, the number of the light emitting diodes 20 and 21, the number of rows, and the number of regions of the light emitting unit 23 are examples, and are not particularly limited. When the illuminance distribution of the pattern light can be represented by a waveform similar to the sine waveform, the length of one side of the light emitting unit 23 can be one wavelength of the sine waveform.

(Structure of Imaging Unit 3)
As shown in FIG. 2, the imaging unit 3 has a plurality of light receiving elements 3a arranged in a line, and the arrangement direction of the light receiving elements 3a is orthogonal to the moving direction (Y direction) of the work W (X direction). ), And a condensing optical system 32. The light receiving element 3a is an image sensor including an image sensor such as a CCD (charge-coupled device) or a CMOS (complementary metal oxide semiconductor) that converts the intensity of light obtained through the condensing optical system 32 into an electric signal. The condensing system optical system 32 is an optical system for condensing light incident from the outside, and typically has one or more optical lenses.

The image pickup unit 3 is connected to the control unit 4 via a signal line 100b different from the signal line 100a of the pattern light illumination unit 2, and is installed away from the pattern light illumination unit 2 and the control unit 4. You can do it. That is, the image pickup unit 3 is configured to be able to be installed independently of the pattern light illumination unit 2. The signal line 100b of the imaging unit 3 may be branched from the signal line 100a of the pattern light illumination unit 2.

An imaging trigger signal transmitted from the control unit 41 of the control unit 4 is input to the imaging unit 3 via the signal line 100b. The imaging unit 3 is configured to execute imaging each time it receives an imaging trigger signal. The imaging trigger signal will be described later.

As shown in FIG. 1, in the case of a flat work W, the pattern light emitted from the pattern light illuminating unit 2 toward the surface of the work W is reflected by the surface of the work W and is reflected by the condensing system of the imaging unit 3. The positional relationship between the image pickup unit 3 and the pattern light illumination unit 2 can be set so as to be incident on the optical system 32.

Further, as shown in FIG. 6A, even when the work W is a columnar member, the pattern light emitted from the pattern light illuminating unit 2 toward the surface of the work W is reflected on the surface of the work W. The positional relationship between the image pickup unit 3 and the pattern light illumination unit 2 can be set so as to be incident on the condensing optical system 32 of the image pickup unit 3.

Further, as shown in FIG. 6B, when the work W is a transparent member such as a transparent film or a sheet, the pattern light emitted from the pattern light illuminating unit 2 toward the surface of the work W is the work. The positional relationship between the image pickup unit 3 and the pattern light illumination unit 2 can be set so as to pass through W and enter the condensing optical system 32 of the image pickup unit 3.

In any of the above cases, the pattern light illumination unit 2 and the image pickup unit 3 are arranged so that the specular reflection component reflected on the surface of the work W is incident on the condensing system optical system 32 of the image pickup unit 3.

The positional relationship between the image pickup unit 3 and the pattern light illumination unit 2 shown in FIG. 1 and the positional relationship between the image pickup unit 3 and the pattern light illumination unit 2 shown in FIG. 6A are each irradiated from the pattern light illumination unit 2 and the work W. This is the positional relationship of the reflected light reception that the light receiving element 3a of the line camera 31 receives the pattern light reflected by.

On the other hand, the positional relationship between the image pickup unit 3 and the pattern light illumination unit 2 shown in FIG. 6B is such that the light receiving element 3a of the line camera 31 receives the pattern light emitted from the pattern light illumination unit 2 and transmitted through the inspection object. It is the positional relationship of light reception. In this case, the number of times the pattern light is reflected is 0.

In the positional relationship of receiving reflected light, the number of times of reflection of the pattern light is 1, and in the positional relationship of receiving transmitted light, the number of times of reflection of the pattern light is 0. Further, instead of the line camera 31, an area camera (a camera in which the light receiving elements are arranged so as to be arranged in the X direction and the Y direction) can be used for the image pickup unit 3, and in the case of this area camera, a form of coaxial illumination is also available. It is possible. In this case, the number of reflections is two because the reflection by the half mirror is added even though the reflected light is installed. By having the user input the number of times the pattern light is reflected, it is possible to determine whether or not there is a positional relationship of receiving reflected light and whether or not there is a positional relationship of receiving transmitted light.

As shown in FIG. 2, the image pickup unit 3 is provided with the LED pointer 34. The LED pointer 34 is for irradiating the pointer light in the axial direction of the condensing optical system 32, and is configured to be able to switch between an irradiated state and a non-irradiated state by a control signal from the control unit 4. ing. As an example of the irradiation form of the pointer light, for example, as shown in FIG. 8, a form for confirming the vertical direction of the line camera 31 can be used. In FIG. 8, the upper side and the lower side of the line camera 31 are irradiated with pointer light, respectively, the upper side is continuously lit and the lower side is blinking. The pointer light may be in any form, and for example, a character or the like may be used as the pointer light. The direction orthogonal to the arrangement direction of the light receiving elements 3a when the line camera 31 is viewed from the light receiving side can be the vertical direction of the line camera 31.

Further, as shown in FIG. 2, the imaging unit 3 is provided with a gravitational acceleration sensor 35 that detects gravitational acceleration. The gravitational acceleration sensor 35 is for grasping the posture of the imaging unit 3 (the angle formed by the axial direction of the condensing optical system 32 and the horizontal plane, the angle formed by the axial direction of the condensing optical system 32 and the vertical plane, etc.). Therefore, the posture of the imaging unit 3 can be obtained in substantially real time according to a conventionally known method by using the output value of the gravitational acceleration sensor 35.

(Structure of display unit 5)
The display unit 5 is composed of, for example, an organic EL display, a liquid crystal display, or the like. The display unit 5 is connected to the control unit 4, and is, for example, an image captured by the imaging unit 3, various images generated based on the image captured by the imaging unit 3, defect inspection results of the work W, and an interface for operation. , Interface for various settings, setting values, etc. can be displayed. The display unit 5 can display a plurality of inspection images and the like at one time.

By using the display unit 5 as a touch panel, the display unit 5 can be provided with various information input functions.

(Structure of keyboard 6 and mouse 7)
The keyboard 6 and the mouse 7 are conventionally known computer operation devices, and are connected to an input information receiving unit 44 (shown in FIG. 2) of the control unit 4. By operating the keyboard 6 or the mouse 7, various information can be input to the control unit 4, and an image or the like displayed on the display unit 5 can be selected.

Specifically, by operating the keyboard 6 or the mouse 7, the moving direction information regarding the moving direction of the work W with respect to the arrangement direction of the light receiving element 3a of the line camera 31, the light receiving element 3a of the line camera 31 and the pattern light illuminating unit 2 When the positional relationship information related to the positional relationship of the above is input to the input information receiving unit 44, the input information receiving unit 44 can receive the input moving direction information and the positional relationship information. The keyboard 6, the mouse 7, and the input information receiving unit 44 are the information acquisition unit 9 of the present invention.

The movement direction information is the movement direction of the work W illustrated in FIG. 1, and can be either the direction of the white arrow or the direction opposite to the white arrow. The movement direction information includes the vertical direction of the line camera 31 and the moving direction of the work W when the direction orthogonal to the arrangement direction of the light receiving elements 3a when the line camera 31 is viewed from the light receiving side is the vertical direction of the line camera 31. And can be included. Further, the positional relationship information includes the positional relationship of the above-mentioned reflected light reception and the positional relationship of the transmitted light reception.

Instead of the keyboard 6 and the mouse 7, or in addition to the keyboard 6 and the mouse 7, for example, a computer operation device such as a voice input device and a touch operation panel can be used.

(Configuration of control unit 4)
The control unit 4 is a unit for controlling each part of the image inspection device 1, and can be composed of a CPU, an MPU, a system LSI, a DSP, dedicated hardware, and the like. The control unit 4 is equipped with various functions as described later, and these may be realized by a logic circuit or may be realized by executing software.

As shown in FIG. 2, the control unit 4 includes a filter processing unit 40, a control unit 41, an image generation unit 42, a UI generation unit 43, an input information reception unit 44, an image analysis unit 45, and an operation reception. It has a unit 46, a display control unit 47, a position designation reception unit 48, and a setting unit 49. A storage device 10 is connected to the control unit 4. The storage device 10 may be a part of a component of the image inspection device 1, or may be a device different from the image inspection device 1. The storage device 10 can be composed of a semiconductor memory, a hard disk, or the like. Further, the storage device 10 may be provided with a reading device capable of reading information stored in various storage media such as a CD-ROM and a DVD-ROM.

(Structure of UI generation unit 43)
The UI generation unit 43 is a part for generating various interfaces to be displayed on the display unit 5, and the various interfaces generated by the UI generation unit 43 are output from the control unit 4 to the display unit 5 and are output to the display unit 5. Is displayed in.

The interfaces generated by the UI generation unit 43 are, for example, the lighting installation method selection interface 50 shown in FIG. 7, the camera vertical confirmation interface 51 shown in FIG. 8, the camera posture confirmation interface 52 shown in FIG. 9, and FIG. There is a work moving direction selection interface 53 and an illumination direction selection interface 54 shown in FIG. The interfaces 50 to 54 are displayed on the display unit 5 when the image inspection device 1 is set before the image inspection device 1 is operated.

The illumination installation method selection interface 50 shown in FIG. 7 is an interface for inputting positional relationship information between the light receiving element 3a of the line camera 31 and the pattern light illumination unit 2, and includes the light receiving element 3a and the pattern light illumination unit 2. It is configured so that it is possible to input whether the positional relationship of is the positional relationship of receiving reflected light or the positional relationship of receiving transmitted light. The lighting installation method selection interface 50 displays a reflected light installation illustration 50a for schematically showing the positional relationship of reflected light reception and a transmitted light installation illustration 50b for schematically showing the positional relationship of transmitted light reception. Has been done. The user sees the reflected light installation illustration 50a and the transmitted light installation illustration 50b displayed on the display unit 5, and has the same positional relationship between the light receiving element 3a installed at the actual site and the pattern light illumination unit 2. Since the illustrations having a positional relationship may be selected with the mouse 7 or the like, input errors are less likely to occur. The input contents of the user's input operation are received by the input information receiving unit 44. This makes it possible to input the positional relationship information easily and without mistakes.

The camera vertical confirmation interface 51 shown in FIG. 8 is an interface for confirming the vertical direction of the line camera 31. When the LED pointer 34 of the image pickup unit 3 irradiates the pointer light, the illustration 50a including the pointer light is displayed so that the user can know that the blinking side is the lower side of the line camera 31. There is. In addition, it is described in sentences which side is the lower side of the line camera 31.

The camera posture confirmation interface 52 shown in FIG. 9 is also an interface for confirming the vertical direction of the line camera 31. The camera posture confirmation interface 52 incorporates a posture display area 52a that schematically displays the current posture of the image pickup unit 3 obtained by the gravitational acceleration sensor 35 of the image pickup unit 3. The posture display area 52a includes a rear view 52b showing the inclination of the imaging unit 3 with respect to the horizontal plane when viewed from the back, a side view 52c showing the inclination of the imaging unit 3 with respect to the vertical plane when viewed from the side, and an imaging unit. Top view 52d showing the posture when 3 is viewed from the top is displayed. The user can confirm the vertical direction of the line camera 31 by looking at each of FIGS. 52b, 52c, and 52d displayed in the posture display area 52a.

The work moving direction selection interface 53 shown in FIG. 10 is an interface for selecting and inputting the moving direction of the work W. The camera vertical confirmation interface 51 shown in FIG. 8 can be incorporated into the work moving direction selection interface 53. The work moving direction selection interface 53 includes an upward movement illustration 53a showing how the work W moves from the lower side (bottom side) of the line camera 31 to the upper side, and the work moving direction selection interface 53 from the upper side to the lower side of the line camera 31. A downward movement illustration 53b showing how the work W is moving is displayed. The user looks at the camera up / down confirmation interface 51 displayed on the display unit 5 to grasp the vertical direction of the line camera 31, and then looks at the upward movement illustration 53a and the downward movement illustration 53b to work. The moving direction of W with respect to the line camera 31 can be selected and input.

The illumination direction selection interface 54 shown in FIG. 11 is an interface for selecting and inputting the positional relationship between the line camera 31 and the pattern light illumination unit 2. The illumination direction selection interface 54 can be displayed on the premise that the user knows the orientation of the line camera 31 (which is the upper side or the lower side), and the pattern light illumination for the line camera 31. It is configured so that the orientation of the part 2 can be selected from a plurality of illustrations.

That is, the illumination direction selection interface 54 includes a first illustration 54a showing a state in which the pattern light illumination unit 2 is arranged so that the signal line 100a exits to the left with respect to the line camera 31, and the pattern light illumination unit 2. The second illustration 54b showing a state in which the signal line 100a is arranged so as to come out upward with respect to the line camera 31, and the pattern light illumination unit 2 are arranged so that the signal line 100a comes out downward with respect to the line camera 31. A third illustration 54c showing the above-mentioned state and a fourth illustration 54d showing a state in which the pattern light illumination unit 2 is arranged so that the signal line 100a comes out to the right side with respect to the line camera 31 are displayed. The user looks at the first to fourth illustrations 54a to 54d displayed on the display unit 5 and selects the positional relationship between the line camera 31 installed at the actual site and the pattern light illumination unit 2. Can be entered.

The interface generated by the UI generation unit 43 can include the cable pull-out direction selection interface 55 shown in FIG. The "cable" is a signal line 100a connected to the pattern light illumination unit 2. The cable pull-out direction selection interface 55 can be configured to include a so-called pull-down menu button. The cable pull-out direction selection interface 55 can be displayed on the premise that the user knows the orientation of the line camera 31 (whether the upper side or the lower side is), and the pattern with respect to the line camera 31 can be displayed. It is configured so that the direction of the light illumination unit 2 can be selected from a plurality of options.

That is, the pull-down menu 55a of the cable pull-out direction selection interface 55 includes four options of "reference direction", "clockwise 90 °", "clockwise 180 °", and "clockwise 270 °". (In FIG. 12, only the "reference direction" is displayed). Here, the "reference direction" is a state in which the pattern light illumination unit 2 is arranged so that the signal line 100a is emitted in the direction (left direction) shown in the first illustration 54a of the illumination direction selection interface 54 shown in FIG. This is the option to select when. "90 ° clockwise" is an option to be selected when the pattern light illumination unit 2 is arranged so that the signal line 100a is output in the upward direction. “Clockwise 180 °” is an option to be selected when the pattern light illumination unit 2 is arranged so that the signal line 100a appears in the right direction. “Clockwise 270 °” is an option to be selected when the pattern light illumination unit 2 is arranged so that the signal line 100a is emitted downward. The expression of each option is an example, and is not limited to the above expression.

(Structure of image analysis unit 45)
The image analysis unit 45 shown in FIG. 2 analyzes a plurality of images captured by the image pickup unit 3 to acquire positional relationship information regarding the positional relationship between the light receiving element 3a of the line camera 31 and the pattern light illumination unit 2. This is a part for the purpose, and can be a part of the information acquisition unit 9.

Specifically, the image analysis unit 45 has a first pattern having a periodic illuminance distribution in one direction as a first step when setting the image inspection device 1 performed before operating the image inspection device 1. To generate a first image of an irradiated surface irradiated with light and a second image of an irradiated surface irradiated with a second pattern light having a periodic illuminance distribution in a direction orthogonal to the one direction. In addition, the pattern light illumination unit 2 and the image pickup unit 3 are controlled. The irradiation surface may be a work W or another member.

The first pattern light in the first step can be an X-direction pattern light whose illuminance changes in the X direction as shown on the left side of FIG. 13, and is set as a pattern light which becomes darker toward the right side of FIG. The second pattern light can be a Y-direction pattern light whose illuminance changes in the Y direction as shown on the right side of FIG. 13, and is set as a pattern light which becomes darker toward the lower side of FIG.

First, the X-direction pattern light on the left side of FIG. 13 is irradiated to an arbitrary irradiation surface, and the irradiation surface irradiated with the X-direction pattern light is imaged by the imaging unit 3 to obtain a first image. As shown in FIG. 14, it is assumed that the first image is an image with a gradation that becomes darker toward the right side. After that, the Y-direction pattern light on the right side of FIG. 13 is irradiated to an arbitrary irradiation surface, and the irradiation surface irradiated with the Y-direction pattern light is imaged by the imaging unit 3 to obtain a second image. It is assumed that the second image is an image without gradation as shown in FIG. Since the arrangement direction of the light receiving elements 3a is the X direction, when the first image and the second image shown in FIG. 14 are obtained, the signal line 100a is directed to the left as shown in the lower side of FIG. It can be estimated that the pattern light illumination unit 2 is arranged in. This relationship may be defined in advance.

When the first image is an image without gradation as shown in FIG. 15 and the second image is an image with gradation that becomes darker toward the right side, the signal line 100a is as shown in the lower side of FIG. It can be estimated that the pattern light illumination unit 2 is arranged so as to come out in the upward direction.

When the first image is an image without gradation as shown in FIG. 16 and the second image is an image with gradation that becomes darker toward the left side, the signal line 100a is as shown in the lower side of FIG. It can be estimated that the pattern light illumination unit 2 is arranged so as to come out downward.

As shown in FIG. 17, when the first image is an image with gradation that becomes darker toward the left side and the second image is an image without gradation, the signal line 100a is as shown in the lower side of FIG. It can be estimated that the pattern light illumination unit 2 is arranged so as to come out to the right.

That is, the image analysis unit 45 executes a step of analyzing the first image and the second image captured by the image pickup unit 3, and as a result, the first image and the second image are images with gradation or without gradation. The positional relationship information between the light receiving element 3a and the pattern light illuminating unit 2 is obtained by obtaining whether the image is an image and, in the case of an image with gradation, which side is darker (or brighter). Can be obtained.

Further, instead of the line camera 31, an area camera (a camera in which the light receiving elements are arranged so as to be arranged in the X direction and the Y direction) can be used for the image pickup unit 3. In the case of this area camera, the image analysis unit 45 can acquire the positional relationship information between the area camera and the pattern light illumination unit 2 by analyzing one image captured by the image pickup unit 3.

Specifically, as shown in FIG. 18A, the pattern light illuminating unit 2 irradiates an arbitrary irradiation surface with a pattern light capable of displaying the alphabet "R", and the irradiation surface irradiated with the pattern light is an imaging unit. Image at 3. As a result, when the image of "R" shown at the top of FIG. 18 is obtained, it can be estimated that the pattern light illumination unit 2 is arranged so that the signal line 100a comes out to the left. When the second image of "R" shown in FIG. 18 is obtained, it can be estimated that the pattern light illumination unit 2 is arranged so that the signal line 100a comes out in the upward direction. When the third image of "R" shown in FIG. 18 is obtained, it can be estimated that the pattern light illumination unit 2 is arranged so that the signal line 100a comes out to the right. When the image of "R" shown at the bottom of FIG. 18 is obtained, it can be estimated that the pattern light illumination unit 2 is arranged so that the signal line 100a comes out downward. The pattern light may indicate something other than "R", and may be, for example, a character, a symbol, a figure, a combination thereof, or the like.

(Other forms of information acquisition unit)
As shown in FIG. 19, a predetermined mark 3c is provided in the image pickup unit 3, and the pattern light illumination unit 2 is provided with a left side symbol A, a lower side symbol B, a right side symbol C, and an upper side symbol D indicating four directions, respectively. Set up. The UI generation unit 43 generates the direction selection interface 56 shown in FIG. The generated direction selection interface 56 is displayed on the display unit 5. The direction selection interface 56 is provided with a pull-down menu 56a on which the left side symbol A, the lower side symbol B, the right side symbol C, and the upper side symbol D can be selected.

After visually confirming to which side the pattern light illumination unit 2 the mark 3c of the image pickup unit 3 is located, the user points to the mark 3c of the image pickup unit 3 by the pull-down menu 56a of the direction selection interface 56. Select the symbols A to D in the direction. As a result, the positional relationship information between the light receiving element 3a of the line camera 31 and the pattern light illuminating unit 2 can be input.

(Caution function)
By using the illumination direction selection interface of FIG. 11, the cable lead-out direction selection interface of FIG. 12, and the direction selection interface of FIG. 20, the user can obtain the positional relationship information between the light receiving element 3a and the pattern light illumination unit 2. You can enter it. However, there are cases where the user makes an erroneous input operation or erroneously recognizes the vertical direction of the line camera 31 in the first place, and it is conceivable that the input result of the user may differ from the actual installation situation.

In response to this, the image inspection device 1 can be provided with a warning function. The attention-calling function allows the user to position the light receiving element 3a and the pattern light illumination unit 2 by using the illumination direction selection interface of FIG. 11, the cable pull-out direction selection interface of FIG. 12, and the direction selection interface of FIG. By performing the analysis shown in FIGS. 13 to 17 after inputting the relationship information or performing the analysis shown in FIG. 18 in the case of an area camera, the input positional relationship information matches the actual installation situation. If the input positional relationship information does not match the actual installation status as a result of the determination, for example, a message is displayed on the display unit 5 to prompt the user to confirm and input again. It is a function.

(Structure of control unit 41)
The control unit 41 shown in FIG. 2 is an image pickup control unit for controlling the pattern light illumination unit 2 and the image pickup unit 3. Specifically, a plurality of pattern lights in which the phase of the illuminance distribution is shifted in at least one direction of the arrangement direction of the light receiving element 3a and the movement direction of the work W are generated and sequentially irradiated to the work W, and the pattern light is irradiated. The pattern light illumination unit 2 and the image pickup unit 3 are controlled so as to obtain a plurality of luminance images by imaging a region including at least the irradiation surface of the pattern light in the work W at the same timing. When obtaining a plurality of luminance images, both the illumination conditions of the pattern light illumination unit 2 and the imaging conditions of the image pickup unit 3 may be changed, or only one of them may be changed.

The illumination conditions of the pattern light illumination unit 2 can include, for example, the type of pattern light to be generated, the order in which the pattern light is generated when a plurality of pattern lights are generated, the brightness of the pattern light (light intensity), and the like. .. Examples of the types of pattern light to be generated include at least four Y-direction pattern lights shown in FIG. 5A and four X-direction pattern lights shown in FIG. 5B, but in addition to these, pattern lights having different wavelengths and the like can be generated. Can be included. The order of generation of the pattern light is as shown in FIG. 5A: the Y-direction pattern light at 0 °, the Y-direction pattern light at 90 °, the Y-direction pattern light at 180 °, and the Y-direction at 270 °. The order of the pattern light is set, and then the X-direction pattern light at 0 °, the X-direction pattern light at 90 °, the X-direction pattern light at 180 °, and the X-direction pattern at 270 ° as shown in FIG. 5B. Can be light.

The illumination conditions of the pattern light illumination unit 2 can be stored in the illumination condition storage unit 10a provided in the storage device 10. Further, the pattern light illumination unit 2 can incorporate an arithmetic unit (field programmable gate array: FPGA) 2a having an illumination condition storage unit 2b capable of storing the illumination conditions of the pattern light illumination unit 2. It is not limited to the FPGA, and a calculation unit other than the FPGA and a lighting condition storage unit 2b can be incorporated.

The imaging condition of the imaging unit 3 includes, for example, at least one of a gain at the time of imaging and an exposure time (shutter speed). The control unit 41 can arbitrarily change the gain and exposure time of the image pickup unit 3, and the image pickup unit 3 executes imaging with the gain and exposure time set by the control unit 41.

(Transmission of trigger signal)
The control unit 41 is configured to be able to receive the encoder pulse signal when the encoder pulse signal is input from the outside. The encoder pulse signal is a pulse signal generated outside the control unit 41, for example, a pulse signal output from a rotary encoder provided on a rotating shaft of a conveyor belt conveyor B for transportation, a pulse signal output from an external control device 8, or the like. Is.

The control unit 4 has a trigger signal transmission unit 4b. The trigger signal transmission unit 4b may form a part of the control unit 41, or may be configured separately from the control unit 41. When the control unit 41 receives one encoder pulse signal from the outside, it triggers to generate a plurality of luminance images by changing at least one of the illumination condition of the pattern light illumination unit 2 and the imaging condition of the imaging unit 3. The signal transmission unit 4b is configured to sequentially transmit a plurality of imaging trigger signals to the imaging unit 3 and to transmit a plurality of illumination trigger signals to the pattern light illumination unit 2. It is also possible to configure so that a plurality of imaging trigger signals are sequentially transmitted only to the imaging unit 3.

In this embodiment, as described above, the pattern light illumination unit 2 sequentially irradiates eight types of pattern light. Therefore, as shown in FIG. 35, the trigger signal transmission unit 4b is an encoder pulse signal EP1 from the outside. Is received, the imaging trigger signal is transmitted eight times (C1 to C8). After that, when the encoder pulse signal EP2 is received, the imaging trigger signal is transmitted eight times in the same manner.

Further, when the trigger signal transmission unit 4b receives one encoder pulse signal from the outside, the trigger signal transmission unit 4b is configured to sequentially transmit a plurality of illumination trigger signals to the pattern light illumination unit 2 in synchronization with the image pickup trigger signal. There is. Since it is necessary to sequentially irradiate the pattern light illumination unit 2 with eight types of pattern light, when the trigger signal transmission unit 4b receives one encoder pulse signal EP1 from the outside, the illumination trigger signal is transmitted eight times (L1). ~ L8) Send. After that, when the encoder pulse signal EP2 is received, the illumination trigger signal is transmitted eight times in the same manner.

Further, the control unit 41 may be configured to receive a single encoder pulse signal from the outside and sequentially transmit a plurality of image pickup trigger signals to the image pickup unit 3 in synchronization with the illumination trigger signal.

(Generation of pattern light by pattern light illumination unit)
The pattern light illumination unit 2 generates a plurality of pattern lights by controlling the current value flowing through the light emitting diodes 20 and 21 based on the illumination trigger signal transmitted from the trigger signal transmission unit 4b. Since the pattern light has a periodic illuminance distribution, the current values to be passed through the light emitting diodes 20 and 21 are set according to the pattern light to be generated so that the illuminance distribution can be obtained. The current values flowing through the light emitting diodes 20 and 21 correspond to the illumination conditions, and can be stored in the illumination condition storage unit 10a of the storage device 10.

The illumination condition can be illumination setting data including illumination timing information (lighting time, lighting interval), illumination intensity information, and illumination pattern information of the light emitting diodes 20 and 21 of the pattern light illumination unit 2. All of the lighting setting data can display the user interface for lighting setting on the display unit 5 and accept the adjustment by the user.

The lighting timing information is the lighting time in which the light emitting diodes 20 and 21 are turned on after receiving the lighting trigger signal, and the light emitting diodes 20 and 21 that are turned on first when the light emitting diodes 20 and 21 to be turned on are switched. It consists of a lighting interval (interval) from turning off the light to turning on the next light emitting diodes 20 and 21. Further, the illumination intensity information is information indicating the illumination intensity (brightness) of the light emitting diodes 20 and 21, and specific examples thereof include current values. Further, the illumination pattern information is information that defines a sin wavy pattern, and includes, for example, a period and a phase shift amount (how much a phase shift is performed in one phase shift). The illumination pattern does not have to be a sine wave, and may be, for example, a rectangular wave.

The lighting setting data may be created by input means such as a keyboard 6 and a mouse 7 for the lighting condition storage unit 10a.

The illumination condition storage unit 10a may be provided in the pattern light illumination unit 2, may be provided in the image pickup unit 3, or may be provided in any component of the image inspection device 1.

The control unit 41 temporarily reads the lighting conditions stored in the lighting condition storage unit 10a of the storage device 10, and outputs the lighting conditions stored in the lighting condition storage unit 10a to the arithmetic unit 2a of the pattern light lighting unit 2. , Can be stored in the lighting condition storage unit 2b included in the arithmetic device 2a. As a result, when the pattern light illumination unit 2 receives the illumination trigger signal from the outside, the work W can be illuminated according to the illumination conditions stored in the illumination condition storage unit 2b, so that the illumination conditions are stored in the illumination condition storage device 10. The generation speed of the pattern light of the pattern light illumination unit 2 can be increased as compared with the case of reading from the unit 10a.

As shown in FIG. 36, the pattern light illumination unit 2 incorporates a current control unit 2c that controls the current value flowing through the light emitting diodes 20 and 21. The current control unit 2c is configured to be able to receive the illumination trigger signal, and when it receives the illumination trigger signal, it controls the current value to be passed through the light emitting diodes 20 and 21 according to the illumination conditions stored in the illumination condition storage unit 2b.

A current control unit 2c for controlling the first light emitting diode trains A1 to A12 and a current control unit 2c for controlling the second light emitting diode trains B1 to B12 are provided, and both current control units 2c have the same configuration. .. The current control unit 2c shown in FIG. 36 controls the first light emitting diode train A1, and the current control unit 2c includes one D / A converter 2d, a plurality of sample hold circuits 2e, and a plurality of voltages /. It is provided with a current conversion circuit 2f. The sample hold circuit 2e and the voltage / current conversion circuit 2f are provided in the same number as the number (12) of the light emitting diodes 20 constituting the first light emitting diode train A1. Twelve sample hold circuits 2e are connected to the D / A converter 2d. Further, a voltage / current conversion circuit 2f is connected to the output side of each sample hold circuit 2e. Each light emitting diode 20 is connected to the output side of the voltage / current conversion circuit 2f.

When switching the pattern, the D / A converter 2d is configured to sequentially output the control voltages for A1 to A12 as shown in FIG. 37. The 12 sample hold circuits 2e sample the voltage of the waveform signal output from the D / A converter 2d at a predetermined sample timing based on the sampling signal input from the outside. After the first sample hold circuit in FIG. 37 samples, the second sample hold circuit samples, and the voltage is sequentially sampled in this way.

The voltage value sampled by the first sample hold circuit is converted into a current value corresponding to the voltage value by the voltage / current conversion circuit 2f, and the current of the current value flows through the light emitting diode A1. In this example, the voltage value sampled by the first sample hold circuit is the lowest, and the voltage value sampled by the second and third sample hold circuits becomes higher in this order. That is, since the first current value is the lowest and the current value increases in the order of the second and third current values, the brightness of the light emitting diodes A1 to A12 changes according to the current value. Patterned light with a periodic illuminance distribution is generated. By changing the waveform signal output from the D / A converter 2d, the illuminance distribution of the pattern light changes, and the illuminance distribution of the pattern light can be represented by a waveform similar to the sine waveform described above, or a triangular wave. It can also be represented by a waveform similar to a sine wave or a waveform similar to a square wave. Further, by setting the current value so that the brightness of all the light emitting diodes 20 is the same, it is possible to illuminate the entire surface with a uniform illuminance.

In this embodiment, the output voltage of the D / A converter 2d is used as the control voltage of each light emitting diode 20, and the timing is specified by the sample hold circuit 2e for each light emitting diode 20 to sample, so that one D / A converter is used. The amount of light of the plurality of light emitting diodes 20 can be controlled by 2d. As a result, the number of high-speed, high-gradation D / A converters 2d, which are generally large in size, can be reduced, and pattern light can be generated by the sample hold circuit 2e, which is small in size. Therefore, the substrate area can be reduced and the cost can be reduced. Can be reduced.

Since the light emitting diodes 21 in each of the second light emitting diode rows B1 to B12 are arranged in the Y direction, the X direction pattern light can be generated by controlling the current value of the second light emitting diode rows B1 to B12. Further, since the light emitting diodes 20 in each of the first light emitting diode rows A1 to A12 are arranged in the X direction, the Y direction pattern light can be generated by controlling the current value of the first light emitting diode rows A1 to A12.

Further, the wavelengths of the four Y-direction pattern lights shown in FIG. 5A and the wavelengths of the four X-direction pattern lights shown in FIG. 5B may be different. For example, by irradiating red pattern light from the light emitting diodes 20 of the first light emitting diode rows A1 to A12 and irradiating blue pattern light from the light emitting diodes 21 of the second light emitting diode rows B1 to B12, the Y direction pattern light is emitted. And the wavelength of the X-direction pattern light can be changed. In this case, if the work W is simultaneously irradiated with the Y-direction pattern light and the X-direction pattern light and the image pickup unit 3 provided with the color camera captures the image, the raw image obtained by the irradiation of the Y-direction pattern light and the X-direction pattern light can be captured. A raw image by irradiation can be obtained at the same time, and the time required for imaging can be shortened.

(Period of illuminance distribution of pattern light)
When the illuminance distribution of the pattern light can be represented by a waveform similar to a sine waveform, the period of the sine waveform can be set to, for example, 100 mm. By setting the period to about 100 mm, even if the surface of the work W having low specular reflectance (not a perfect diffuse reflector) is imaged, the stacking process of the shape described later can be performed, and the work W can be supported. The range can be widened. Further, by setting the period of the sine waveform to about 100 mm, the attenuation of the lens of the condensing optical system 32 of the imaging unit 3 due to the MTF is suppressed. However, when inspecting the coated surface or the like, it may be preferable to shorten the cycle, so the cycle is not limited to 100 mm. By shortening the cycle, it is possible to reduce the influence of the base when inspecting the coated surface, and it becomes easier to detect a subtle difference in the surface gradient.

(Generation of inspection image by deflection metric)
In this embodiment, the imaging unit 3 images the work W at the timing of being illuminated by the pattern light illumination unit 2 to obtain a plurality of luminance images, and the phase displacement measurement is performed based on the plurality of luminance images captured by the imaging unit 3. The phase map calculated using the principle of the method (Phase Measuring Deflectometry: PMD, hereinafter referred to as "deflectometry") is processed to generate multiple inspection images capable of detecting different types of defects. , The work W can be inspected using the inspection image. The step of obtaining a plurality of luminance images can be executed by the control unit 41 outputting the illumination trigger signal and the image pickup trigger signal to control the pattern light illumination unit 2 and the image pickup unit 3 as described above. The step of generating an inspection image can be performed by the image generation unit 42 shown in FIG.

The image generation unit 42 generates at least an inspection image showing the reflection state of the surface of the work W and an inspection image showing the shape of the work W based on the plurality of luminance images captured by the imaging unit 3. Examples of the inspection image showing the reflection state of the surface of the work W include a specular reflection component image, a diffuse reflection component image, and the like. Examples of the inspection image showing the shape of the work W include a shape image, a depth contour image, a gloss ratio image, and the like.

Hereinafter, the generation of the inspection image will be described in detail according to the flowchart shown in FIG. 21 and the schematic diagram shown in FIG. 22. In step SA1 of the flowchart, the control unit 41 controls the pattern light illumination unit 2 and the image pickup unit 3 to obtain a plurality of luminance images (raw images). The luminance image obtained in step SA1 is an intermediate image, which is an image that is temporarily stored in the storage device 10 or the like but is not displayed on the display unit 5. The luminance image is an image obtained by imaging a work W illuminated with different illumination modes (illumination by a plurality of patterns of light).

In step SA1, the control unit 41 outputs a lighting trigger signal to the pattern light lighting unit 2. When the pattern light illumination unit 2 receives the illumination trigger signal, it generates pattern light according to the illumination conditions stored in the illumination condition storage unit 2a and illuminates the work W. The illumination conditions at this time are the Y-direction pattern light at 0 °, the Y-direction pattern light at 90 °, the Y-direction pattern light at 180 °, and the Y-direction pattern light at 270 ° as shown in FIG. 5A. , X-direction pattern light at 0 °, X-direction pattern light at 90 °, X-direction pattern light at 180 °, and X-direction pattern light at 270 °, as shown in FIG. 5B. It is a condition. Therefore, the pattern light illumination unit 2 has a total of eight types of pattern light (four types in the X direction and four types in the Y direction) in which the phase of the illuminance distribution is shifted in both the arrangement direction of the light receiving element 3a and the movement direction of the work W. ) Is generated and the work W is sequentially irradiated.

At the same time, the control unit 41 outputs an image pickup trigger signal to the image pickup unit 3. When the image pickup unit 3 receives the image pickup trigger signal, the image pickup unit 3 images the work W each time the pattern light is irradiated. When the screen size is 8K (8192 columns x 1024 rows), 1024 times (number of rows) x number of illuminations of pattern light (8 times) is the number of illuminations required to obtain one luminance image, and the same number is used. It is the number of images taken.

In this embodiment, a plurality of imaging trigger signals (specifically, the same number of imaging trigger signals as the number of rows of pixels of one luminance image) are sequentially transmitted in order to obtain one luminance image, so that imaging is performed. It is possible to eliminate the deviation of the imaging timing each time the trigger signal is received. Since the illumination trigger signal is similarly transmitted to the pattern light illumination unit 2 and the illumination trigger signal and the image pickup trigger signal are synchronized, the deviation of the illumination timing can be eliminated. Therefore, even if the number of illuminations and the number of imagings increase when a plurality of luminance images are generated, the deviation between the imaging timing and the illumination timing can be maintained in an extremely small state.

Since the number of times the pattern light is irradiated at the time of imaging is eight, eight luminance images can be obtained. As shown in FIG. 22, four luminance images (X) captured when irradiating four types of X-direction pattern light and four luminance images (Y) captured when irradiating four types of Y-direction pattern light. ) And is obtained.

FIG. 23 shows eight luminance images obtained by actually imaging the metal work W. The images on the left side of FIG. 23 are, in order from the top, an X-direction pattern light at 0 °, an X-direction pattern light at 90 °, an X-direction pattern light at 180 °, and an X-direction pattern at 270 °. It is a luminance image obtained by irradiating each with light. The images on the right side of FIG. 23 are, in order from the top, a Y-direction pattern light at 0 °, a Y-direction pattern light at 90 °, a Y-direction pattern light at 180 °, and a Y-direction pattern at 270 °. It is a luminance image obtained by irradiating each with light. It should be noted that only the X-direction pattern light or only the Y-direction pattern light may be irradiated.

After that, the process proceeds to step SA2 of the flowchart shown in FIG. 21, and the eight luminance images obtained in step SA1 are subjected to deflation metrics processing. This can be done by the image generation unit 42. By imaging the work W irradiated with the pattern light, each pixel value of the obtained luminance image includes a specular reflection component and a diffuse reflection component (+ environmental component). Since the phase of the illuminance distribution of the pattern light is shifted by 90 ° (π / 2) in the X direction and the image is taken four times, four types of pixel values can be obtained by irradiating the pattern light in the X direction. The pixel value of the brightness image captured the first time is I _{1 , the pixel value of the brightness image captured the second time is I 2} , the pixel value of the brightness image captured the third time is I ₃ , and the fourth time. Assuming that the pixel value of the luminance image captured in is I ₄ , the pixel values I _{1 to} I ₄ are given by the following equation 1.

In Equation 1, Rd is a diffuse reflection component, Rs is a specular reflection component, and φs is a specular angle (phase). The diffuse reflection component, the specular reflection component, and the specular reflection angle are phase data, and phase data for one line can be obtained based on a plurality of line images captured by the imaging unit 3.

Similarly, since the phase of the illuminance distribution of the pattern light is shifted by 90 ° (π / 2) in the Y direction and the image is taken four times, four types of pixel values can be obtained by irradiating the pattern light in the Y direction. _{Pixel values I 1 to} I ₄ can be expressed by the above equation 1 also in the Y direction.

(Specular reflection component image)
The specular reflection component image is given by the following equation 2. In Equation 2, the diffusion component is excluded by the difference between the opposite phases. As shown in FIG. 22, it is obtained in the X direction and the Y direction, respectively, and a specular reflection component image is obtained by synthesizing them. FIG. 29 shows an example of a specular reflection component image. The normal image in FIG. 29 is an image including a specular reflection component and a diffuse reflection component.

(Specular reflection angle)
The specular reflection angle is given by the following equation 3. The angle is calculated with tanθ = sinθ / cosθ based on the specular reflection component with a π / 2 deviation.

(Average image)
The average image contains a diffusion component and an environmental component, and is given by the following equation 4. The specular reflection component is eliminated by adding the opposite phases.

(Diffuse reflection component image)
The diffuse reflection component image is given by the following equation 5. An example of a diffuse reflection component image is shown in FIG.

In step SA3 in the flowchart of FIG. 21, contrast correction is performed on the diffuse reflection component image. Further, in step SA4, contrast correction is performed on the specular reflection component image. Each contrast correction can be a linear correction. For example, the ROI is corrected so that the average is the median. In the case of 8 bits, the median value may be 128 levels. As a result, the corrected diffusion component and the corrected specular reflection component can be obtained.

Further, in step SA5, the difference between the phase component and the reference phase is taken. In this step SA5, the difference is acquired with respect to the phase of the reference plane. For example, the user specifies a spherical shape, a cylindrical shape, a flat shape, or the like as the reference plane, and obtains the difference from these. Alternatively, it may be extracted with a free curved surface. The corrected phase (difference) is obtained in the X direction, and the corrected phase is also obtained in the Y direction. Corresponds to the phase image shown in FIG. An example of the corrected phase image is shown in FIG.

The diffuse reflection component image, specular reflection component image, and reference phase difference image are output images.

In step SA6, a layered image and a depth contour image are obtained based on the reference phase difference image obtained in step SA5. The layered image is an image in which 1/2 reduction is repeated. As a result, images of layered phases are obtained in the X direction and the Y direction, respectively.

On the other hand, the depth contour image is an intermediate image in which a portion having a large phase difference is emphasized, and is a concept different from curvature. The depth contour image can be obtained at a higher speed than the shape image obtained by stacking the shapes, and has advantages such that line scratches on the work W are extremely easy to see and the contour can be easily extracted. An example of a depth contour image is shown in FIG.

Next, in step SA7, the shapes are stacked on the layered phase images to generate the shape images. A shape image can be obtained by performing a stacking calculation using the Gustav-Jacobi method or the like for the specular reflection angles in the X and Y directions. The shape image is an output image. An example of the shape image is shown in FIG.

In general, there are many cases where the shape is restored by unwrapping and then triangular distance measurement, but in this embodiment, unwapping is avoided and the local differential values are accumulated and calculated by the Gustav-Jacobi method. As a result, the shape is restored without relying on triangular distance measurement. As a shape restoration method, a known method can be appropriately used. A method that does not use triangular ranging is preferable. It is also possible to use a hierarchical method in which reduced images are held in multiple stages. It is also possible to use a method having a difference between the reduced image and the normal image.

Furthermore, the feature size can be set as a parameter. The feature size is a parameter for setting the size of the scratch to be detected according to the purpose and type of inspection. For example, when the parameter value of the feature size is 1, the finest scratches can be detected, and when this value is increased, large scratches can be detected. As a result, when the feature size is increased, a larger scratch is easily detected, and the unevenness of the work W surface becomes clear.

In step SA8, simple defect extraction is performed. Details of simple defect extraction will be described later. After performing simple defect extraction in step SA8, a defect extraction image displaying the extracted defects is output.

In addition, the image generation unit 42 can also generate a gloss ratio image. The gloss ratio image is an image showing the ratio of the specular reflection component and the diffuse reflection component. In the gloss ratio image, pixels having different ratios of the specular reflection component and the diffuse reflection component are emphasized. An example of a gloss ratio image is shown in FIG.

(Structure of filter processing unit 40)
The filter processing unit 40 shown in FIG. 2 has a filter processing setting unit 40a capable of individually setting filter processing application and filter processing non-application for each inspection image generated by the image generation unit 42, and a filter processing setting. It has a filter processing execution unit 40b that executes a filter processing on the inspection image set to be filtered by the unit 40a. The type of filter to be applied is not particularly limited, but for example, it can be selected according to the type of the inspection image, and examples thereof include a shading correction filter and a smoothing filter.

As shown in FIG. 30, the shading correction filter is not suitable for shape images, gloss ratio images, depth contour images and phase images. The reason is that the shape image is not an image of brightness but an image showing unevenness, so that the effect of the shading correction filter cannot be obtained. Further, since the gloss ratio image is an image obtained by obtaining the ratio, it is an image in which shading has disappeared in the first place. Further, the depth contour image and the phase image are corrected in consideration of the phase, and there is no point in applying the shading correction filter. On the other hand, a shading correction filter is suitable for a specular reflection component image, a diffuse reflection component image, and a normal image.

The median filter in FIG. 30 is a smoothing filter and is not suitable for a shape image, but is suitable for a gloss ratio image, a depth contour image, a phase image, a specular reflection component image, a diffuse reflection component image, and a normal image. ..

The filter processing setting unit 40a is configured to obtain information on the type of the inspection image and automatically set the application of the filter processing and the non-application of the filter processing according to the inspection image. Information on the type of inspection image can be obtained from the image generation unit 42. As shown in FIG. 30, the filter processing setting unit 40a automatically applies the shading correction filter to the normal reflection component image, the diffuse reflection component image, and the normal image, and the shape image, the gloss ratio image, and the depth contour image. And it does not automatically apply to phase images. Further, the filter processing setting unit 40a does not automatically apply the median filter to the shape image, but automatically applies to the gloss ratio image, the depth contour image, the phase image, the specular reflection component image, the diffuse reflection component image, and the normal image. Applies to.

That is, by investigating the effectiveness of the filter processing in advance and presetting the inspection image to which the filter processing is applied in the image inspection device 1, the filter processing setting unit 40a responds to the type of the inspection image. The application and non-application of at least one of the shading correction filter and the smoothing filter can be automatically set. The preset contents may be changed by the user.

Since the filter processing execution unit 40b can execute the filter processing according to the preset contents, the user selects the application / non-application of the filter processing for each inspection image and the type of the filter processing when the filter processing is applied. It is not necessary and the operability is improved. Further, since the filter processing setting unit 40a is configured to set the filter processing only for the inspection image used for the inspection, the filter processing is not executed for the inspection image not used for the inspection. ..

The type of filter processing to be preset may be a filter that can investigate the effectiveness in advance and a filter that shows the presence or absence of the effect by the inspection image.

An inspection image can be selected from the display image selection pull-down menu 58e of the defect extraction interface 58 shown in FIG. 25. Selected as an inspection image by the display image selection pull-down menu 58e, and selected as an inspection image (normal reflection component image, diffuse reflection component image, etc.) after the filter processing execution by the filter processing execution unit 40b. An inspection image (shape image, etc.) that has been processed and has not been filtered by the filter processing execution unit 40b can be displayed on the display unit 5 at the same time. Even if the inspection image is suitable for filtering, the filtering execution unit 40b does not filter the inspection image (inspection image that is not displayed on the display unit 5) that is not selected as the inspection image. Can be done. Even if the filtering process is not performed at this point, the filtering processing execution unit 40b may perform the filtering process later at the stage of being selected as the inspection image.

It should be noted that, on the premise that the filter processing application and the filter processing non-application are automatically set according to the type of the inspection image, the user disables the automatically set filter processing or automatically sets the filter processing. The user may be able to change the type of filtering performed. For example, the filter processing setting unit 40a can be configured so that the user accepts an operation of not applying the filter processing to the inspection image determined that the filter processing is unnecessary. Specific examples of the operation include operations such as displaying an interface on the display unit 5 and incorporating various buttons and pull-down menus in the interface.

When the filter processing execution unit 40b executes the filter processing on a plurality of inspection images, the filter processing execution unit 40b performs the filter processing that is executed on the inspection images for the plurality of inspection images that have been filtered. It may be configured so that it can be non-executed all at once. When the user performs an operation of not applying the filtering process to the inspection image, the filtering process can be automatically disabled for each inspection image. It is also possible to configure to accept an operation of undoing the filtering process after collectively disabling the filtering process, that is, an operation of executing the filtering process on a plurality of inspection images.

When performing each filter processing, as shown in FIG. 31, various parameters such as the strength of the filter processing can be set. The filter processing execution unit 40b is configured so that the parameters of the filter processing can be collectively set and changed for the inspection image set to be filtered by the filter processing setting unit 40a. When a parameter is set or changed in the parameter setting interface shown in FIG. 31, the operation is reflected by the filter processing execution unit 40b on another inspection image set to be filtered.

Further, it is not necessary to preset the inspection image to which the filtering process is applied in the image inspection device 1. In this case, the user operates various buttons, pull-down menus, etc. built into the interface to set application / non-application of the filter processing individually or collectively for each inspection image. .. Various buttons and pull-down menus built into the interface are the filter processing setting section.

FIG. 32 is a diagram showing an inspection image before filtering and an inspection image after filtering. The filtering process is applied to the original image 1 (specular reflection component image) displayed in the first image display area 58a and the original image 2 (diffuse reflection component image) displayed in the second image display area 58b. ing. As shown in FIG. 32, the shading of the specular reflection component image and the diffuse reflection component image is corrected so that the defective portion can be easily seen. On the other hand, since the filter process is not applied to the original image 3 (shape image) displayed in the third image display area 58c, the defective portion related to the shape remains easy to see.

(Simple defect extraction)
In the simple defect extraction, as shown in FIG. 24, the concept of simple defect extraction is shown in an image 1 in which line scratches (defects) of the work W are easily inspected, an image 2 in which stains (defects) in the work W are easily inspected, and an image 2 of the work W. This is a step of generating images 3 for which dents (defects) can be easily inspected, and then displaying the defects appearing in images 1 to 3 on the defect extraction image. The defect extraction image is an image that can extract and display the defects appearing in the images 1 to 3, and is called a composite defect image because the extracted defects can be combined and displayed in one image. You can also.

For example, the specular reflection component image is an image in which it is easy to confirm stains that dull the specular reflection, scratches that have no shape change but dull specular reflection, or scratches that do not return specular reflection due to the shape change. The diffuse reflection image is an image in which it is easy to confirm the state of the texture on the surface of the work W (specifically, characters on the printed matter, dark stains, etc.). The shape image is an image in which phase changes are accumulated while looking at surrounding pixels according to the feature size. Here, if the feature size of the shape image is set large, it is possible to capture unevenness that is relatively shallow and has a large area even in the shape change. On the other hand, if the feature size of the shape image is set small, line scratches and scratches with a small area can be captured. In addition, defects that are difficult to appear in the shape image (for example, fine scratches and deep scratches) tend to appear easily in the specular reflection component image. Further, in the depth contour image, a reference plane is calculated and the deviation from the reference plane is imaged. From the depth contour image, line scratches and scratches with a small area can be captured. That is, the image generation unit 42 can generate a plurality of inspection images capable of detecting different types of defects based on the plurality of luminance images captured by the image pickup unit 3. Further, the image generated by the image generation unit 42 is not limited to the above-mentioned image.

The specular reflection component image, diffuse reflection image, shape image, depth contour image, gloss ratio image, defect extraction image, etc. generated by the image generation unit 42 are inspection images and can be displayed on the display unit 5. .. The display unit 5 may display only one of the images generated by the image generation unit 42, or may display an arbitrary plurality of images. It is preferable that the image to be displayed on the display unit 5 can be selected by the user.

For example, the UI generation unit 43 can generate the defect extraction interface 58 shown in FIG. 25, and the defect extraction interface 58 can be displayed on the display unit 5. The defect extraction interface 58 includes a first image display area 58a for displaying the original image 1, a second image display area 58b for displaying the original image 2, and a third image display area 58c for displaying the original image 3. A fourth image display area 58d for displaying the defect-extracted image is incorporated. The number of image display areas is not limited to four, and may be three or less, or five or more. In FIG. 25, a normal reflection component image is selected as the original image 1, a diffuse reflection component image is selected as the original image 2, and a shape image is selected as the original image 3, but the image generated by the image generation unit 42 is selected. If so, any image may be displayed in any area.

The user can also switch the images displayed in the first to third image display areas 58a to 58c to other images. When switching the displayed image, for example, the display area is selected and activated by the operation of the mouse 7, and the display image selection pull-down menu 58e incorporated in the defect extraction interface 58 is used to select and activate the displayed image. You can make a choice. The display image selection pull-down menu 58e is an image selection unit that selects at least one inspection image to be used for inspection from a plurality of inspection images generated by the image generation unit 42.

In the pull-down menu 58e for selecting a display image, a specular reflection component image, a diffuse reflection image, a shape image, a depth contour image, a gloss ratio image, and the like are displayed as options. The image generation unit 42 generates a defect extraction image using the selected inspection image. The display image selection pull-down menu 58e is an image selection reception unit that can accept selection of an inspection image for generating a defect extraction image from a plurality of inspection images generated by the image generation unit 42. ..

The image inspection device 1 is configured so that a part of the inspection image displayed on the display unit 5 can be designated and enlarged. For example, when the user wants to enlarge the area B1 indicated by the broken line on the image displayed in the first image display area 58a of FIG. 25, the pointer of the mouse 7 is placed on the start point PA and the end point PB is reached as it is. When the drag operation is performed with the mouse 7, a rectangular area having a diagonal line connecting the start point PA and the end point PB is specified, and this area becomes the area B1. This operation corresponds to an enlargement operation for designating and enlarging a part of an area of one inspection image, and this enlargement operation is accepted by the operation reception unit 46 of the control unit 4. The shape of the region specified by the enlargement operation may be, for example, a circular shape, a shape surrounded by a free curve, or the like. Further, the enlargement operation may be performed by any operation of the mouse 7 or the keyboard 6.

Upon receiving the enlargement operation by the user, the display control unit 47 of the control unit 4 enlarges the image area B1 displayed in the first image display area 58a and enlarges the first image display area as shown in FIG. 26. Display on 58a. Further, the display control unit 47 displays the first image display area in the other inspection images (images displayed in the second to fourth image display areas 58b to 58d) displayed on the display unit 5 shown in FIG. 25. Each area B2 corresponding to the area B1 designated by the image displayed on the 58a is enlarged at the same magnification in conjunction with the above-mentioned enlargement operation and displayed on the display unit 5 (see FIG. 26). The expansion of the region B1 and the expansion of the region B2 can be performed at the same time, but may be accompanied by a slight time delay. The enlargement operation may be performed on any of the inspection images displayed in the first to fourth image display areas 58a to 58d, and no matter which inspection image is enlarged, another inspection image may be performed. Can be enlarged at the same magnification and displayed on the display unit 5 in conjunction with the enlargement operation.

Further, the operation reception unit 46 may be configured to be able to accept a reduction operation for reducing one inspection image among the inspection images displayed on the display unit 5. The reduction operation can be performed by, for example, a click operation of the mouse 7 or a key operation of the keyboard 6. The display control unit 47 may be configured to reduce the other inspection image displayed on the display unit 5 at the same magnification in conjunction with the reduction operation and display it on the display unit 5.

Further, the operation reception unit 46 may be configured to be able to accept a scroll operation for scrolling one inspection image among the inspection images displayed on the display unit 5. The scroll operation can be performed by, for example, a click operation of the mouse 7 or a key operation of the keyboard 7. The display control unit 46 may be configured to scroll another inspection image displayed on the display unit 5 in conjunction with the scroll operation. The scrolling direction can be any direction of the display unit 5 in the vertical direction, the horizontal direction, and the diagonal direction.

The defect extraction interface 58 shown in FIG. 25 incorporates a detail display button 58f for selecting whether or not to perform detailed display. When the detail display button 58f is operated to display the details, the threshold value setting area 58i is displayed as shown in FIG. Each threshold value of the original images 1 to 3 can be displayed in the threshold value setting area 58i. The threshold value is a value for defining a pixel value range to be a non-defective portion. For example, when the brightness level of each pixel is set to 256 levels from 0 to 255, the pixel value range to be a non-defect portion can be defined by a specific numerical value from 0 to 255. Black may be a non-defect part, and white may be a non-defect part.

(Background selection method)
There are two ways to set the threshold. The first is the method of selecting the background. The defect extraction interface 58 shown in FIG. 25 is provided with a background selection button 58g, and after pressing the background selection button 58g, the background of each inspection image (original images 1 to 3) is selected. The background selection can be performed on the original images 1 to 3 enlarged and displayed after the enlargement operation. The background is a portion other than the defective portion of the work W, that is, a non-defective portion of the work W.

As a method of selecting the background, for example, there are a method of placing the pointer of the mouse 7 on the non-defective portion of the work W in the original images 1 to 3 and clicking, or a method of operating the touch panel. This makes it possible to specify the position (X coordinate, Y coordinate) of the non-defective portion of the work W. The position of the non-defect portion can be specified a plurality of times for any one image among the original images 1 to 3. Further, after designating the position of the non-defect portion for any one image (first inspection image) among the original images 1 to 3, the position of the non-defect portion is also specified for another image. Can be specified. When the background is selected, the pixel value range to be a non-defect portion changes in the direction of widening.

This will be described based on the flowchart shown in FIG. 33. When the user specifies the position of the non-defective portion of the work W (first time) (step SB1), the first designation of the position of the non-defective portion of the control unit 4 is performed by the control unit 4. It is accepted by the position designation reception unit 48. The position information of the non-defective portion (first non-defective portion position information) received by the position designation reception unit 48 is stored in the position information storage unit 10b of the storage device 10 as a coordinate format.

The setting unit 49 of the control unit 4 reads the position information of the non-defective unit received by the position designation reception unit 48 from the position information storage unit 10b, and obtains the pixel value of the position of the non-defect unit received by the position designation reception unit 48. .. Specifically, not only the pixel at the position of the non-defect portion (clicked pixel) but also the pixel value (9 pixel values) of the 3 × 3 area around the pixel is acquired, and the maximum value and the minimum value are acquired. And ask. For example, if the maximum value is 200, the margin is three times the sensitivity of background selection (for example, 5), and 200 + 5 × 3 = 215 is the maximum value. Further, for example, if the minimum value is 80, 80-5 × 3 = 65 is set as the minimum value. In this case, the pixel value range to be the non-defect portion is 215 to 65. Based on the pixel value of the position of the non-defect portion, the setting unit 49 sets the pixel value range to be the non-defect portion by a plurality of inspection images (including the first inspection image) displayed on the display unit 5. ) Is automatically set (step SC1). The above sensitivity and margin are examples.

After that, when the user specifies the position of the non-defective portion of the work W (second time) (step SB2), as the internal processing of the control unit 4, the second designation of the position of the non-defective portion is performed by the control unit. It is accepted by the position designation reception unit 48 of 4. The position information of the non-defective portion (second non-defective portion position information) received by the position designation reception unit 48 is similarly stored in the position information storage unit 10b of the storage device 10.

The setting unit 49 of the control unit 4 reads the position information of the second non-defective unit received by the position designation reception unit 48 from the position information storage unit 10b, and receives the position information of the second non-defective unit received by the position designation reception unit 48. Get the pixel value of the position. Based on the pixel value of the position of the second non-defect portion, the setting unit 49 sets the pixel value range to be the non-defect portion by a plurality of inspection images displayed on the display unit 5 (for the first inspection). Automatically set for (including images) (step SC2).

When the user specifies the position of the non-defect portion of the work W (third time) (step SB3), the non-defect portion should be designated based on the pixel value of the position of the non-defect portion of the third time. The pixel value range is automatically set for a plurality of inspection images (including the first inspection image) displayed on the display unit 5 (step SC3). That is, the position designation receiving unit 48 can receive the designation of the position of the non-defective portion a plurality of times.

Further, the inspection image (second inspection image) that is not displayed on the display unit 5 and has not been filtered is selected by the display image selection pull-down menu 58e of the defect extraction interface 58 shown in FIG. Then, the image displayed on the display unit 5 is switched (step SB4). When an inspection image that is not displayed on the display unit 5 and has not been filtered is selected, the setting unit 49 sets a plurality of positions indicating the positions of the non-defective parts stored in the position information storage unit 10b. The information (position information stored in steps SB1 to 3) is read out, and the pixel value corresponding to the position information in the inspection image is referred to. After that, the pixel value range to be a non-defect portion is updated (step SC4). The filter processing unit 40 filters the inspection image. The background can be selected for the inspection image after the switching (step SB5), and the position information (second non-defective part position information) specified at this time is stored in the position information storage unit 10b.

Further, after the inspection image that is not displayed on the display unit 5 and has not been filtered is selected as the second inspection image, another image is selected as the third inspection image. There is (step SB6). When the third inspection image is selected, the setting unit 49 sets a plurality of position information indicating the positions of the non-defective parts stored in the position information storage unit 10b (position information stored in steps SB1 to SB5). Is read out, and the pixel value range to be a non-defect portion is updated by referring to the pixel value corresponding to the position information in the third inspection image. Although not shown, the background can be selected for the third inspection image, and the position information (third non-defective part position information) specified at this time is stored in the position information storage unit 10b. The filter processing unit 40 performs filter processing on the third inspection image.

That is, when the third inspection image is selected, the setting unit 49 indicates the position of the non-defective portion in the second inspection image stored in the position information storage unit 10b (second non-defect). The part position information) can be read out, and the pixel value range to be a non-defective part can be updated by referring to the pixel value corresponding to the position information in the second inspection image. Further, the setting unit 49 indicates the position of the non-defective portion in the first inspection image stored in the position information storage unit 10b when the third inspection image is selected (the position of the first non-defective portion). Information) and position information (second non-defective part position information) indicating the position of the non-defective part in the second inspection image are read out, and corresponding to the first and second non-defect part position information in the third inspection image. It may be configured to update the pixel value range to be a non-defective part by referring to the pixel value to be processed. The background selection history can be saved, for example, only 10 times, and the oldest ones can be deleted after that.

The reason for selecting the background instead of the defective part of the work W is as follows. That is, if the method of designating the defective parts by clicking them one by one on the screen of the display unit 5, it is necessary to set a plurality of defective conditions, which complicates the operation. In addition to having to prepare a sample of each defective part, the user cannot deal with unknown defective parts, and the user feels stress in the work of designating small defective parts such as line scratches. Is also possible. Therefore, although the background selection is preferable in many cases, the method of specifying the defective portion described below may also be effective, and in the present embodiment, the method of selecting the background and the method of specifying the defective portion are used. Allows people to choose.

(Forced extraction method)
The second method of setting the threshold value is a method of designating a defective portion (forced extraction method). For example, when the difference between the defective portion and the non-defective portion is not so large, it may be desired to set the threshold value by selecting the defective portion, and in this case, the forced extraction method is effective.

The defect extraction interface 58 shown in FIG. 25 is provided with a forced extraction button 58h, and after pressing the forced extraction button 58h, a defective portion of each inspection image is selected (shown in FIG. 27). The forced extraction is to extract the defective portion of the work W. Forced extraction can be performed on each enlarged inspection image. The flow of forced extraction is almost the same as the case of background selection shown in FIG. 33.

As a method of forced extraction, for example, there are a method of placing the pointer of the mouse 7 on the defective portion of the work W in the original images 1 to 3 and clicking the pointer, and a method of operating the touch panel. This makes it possible to specify the position (X coordinate, Y coordinate) of the defective portion of the work W. The position of the defective portion can be specified a plurality of times for any one image among the original images 1 to 3. Further, after designating the position of the defective portion for any one of the original images 1 to 3 (the image for the first inspection), the position of the defective portion is also specified for another image. It can be carried out. When forced extraction is performed, the pixel value range to be a non-defect portion changes in the direction of narrowing. In addition, forced extraction and background selection can be performed on one original image.

The designation of the position of the defective portion of the work W is accepted by the position designation receiving unit 48 of the control unit 4 shown in FIG. The position designation receiving unit 48 can receive the designation of the position of the defective portion a plurality of times. The position information (coordinates) of the defective portion received by the position designation reception unit 48 is stored in the position information storage unit 10b of the storage device 10.

The setting unit 49 of the control unit 4 reads the position information of the defective portion received by the position designation receiving unit 48 from the position information storage unit 10b, and obtains the pixel value of the position of the defective portion received by the position designation receiving unit 48. In the case of forced extraction, unlike the case of background selection, the pixel value of the pixel (clicked pixel) at the position of the defect portion is acquired, and the setting unit 49 is non-defect based on the pixel value at the position of the defect portion. The pixel value range to be a unit is automatically set for a plurality of inspection images (including the first inspection image) displayed on the display unit 5. At this time, the pixel value range to be the non-defect portion is set so as not to include the pixel value at the position of the defect portion. However, if the changed pixel value range does not become a valid range, the upper limit value is set to 0 and the lower limit value is set to 255.

Further, when an inspection image (second inspection image) that is not displayed on the display unit 5 and has not been filtered is selected, the setting unit 49 has a defect stored in the position information storage unit 10b. A plurality of position information indicating the position of the portion is read out, and the pixel value corresponding to the position information in the second inspection image is referred to. The pixel value range to be a non-defective part in the second inspection image is updated. The filter processing unit 40 filters the second inspection image. Forced extraction can also be performed on the second inspection image, and the position information specified at this time is stored in the position information storage unit 10b.

Further, when an inspection image that is not displayed on the display unit 5 and has not been filtered is selected as the second inspection image, and then another image is selected as the third inspection image. In addition, the setting unit 49 reads out a plurality of position information indicating the position of the defective part stored in the position information storage unit 10b, and refers to the pixel value corresponding to the position information in the third inspection image as the non-defect part. Update the pixel value range to be. The filter processing unit 40 performs filter processing on the third inspection image.

Further, the setting unit 49 provides position information (second defect portion position) indicating the position of the defect portion in the second inspection image stored in the position information storage unit 10b when the third inspection image is selected. Information) can be read out, and the pixel value range to be a non-defect portion can be updated by referring to the pixel values corresponding to the position information in the second inspection image.

Further, the setting unit 49 indicates the position of the defective portion in the first inspection image stored in the position information storage unit 10b when the third inspection image is selected (first defective portion position information). And the position information (second defect portion position information) indicating the position of the defect portion in the second inspection image is read out, and the pixel values corresponding to the first and second defect portion position information in the third inspection image are referred to. It may be configured to update the pixel value range to be a non-defective portion.

The image generation unit 42 sets a region included in all of the pixel value ranges to be non-defective portions set for a plurality of inspection images by the setting unit 49 as a non-defect region, and includes the region in any of the pixel value ranges. A defect extraction image is generated in which the non-existent region is used as the defect region. This defect-extracted image can be displayed in the fourth image display area 58d of FIGS. 26 and 27. Further, when the background selection or forced extraction is performed, the pixel value range to be the non-defective part is changed, but when the position designation reception unit 48 accepts the designation of the position of the non-defect part or the defective part, the image generation unit 42 can automatically update the non-defect region and the defect region in the defect extraction image.

The display control unit 47 enlarges the area corresponding to the area specified by another inspection image in the defect extraction image displayed on the display unit 5 at the same magnification in conjunction with the enlargement operation, and displays the area on the display unit 5. It can be displayed. Further, the display control unit 47 reduces the area corresponding to the area designated by another inspection image in the defect extraction image displayed on the display unit 5 at the same magnification in conjunction with the reduction operation, and the display unit 47. It can be displayed in 5. Further, the display control unit 47 can scroll the defect extraction image displayed on the display unit 5 in conjunction with the scroll operation.

It is also possible to perform background selection after performing forced extraction, or to perform forced extraction after performing background selection.

(Histogram and pins)
Further, as shown in FIG. 28, the histogram and the pin can be displayed when the threshold value is set. For example, when an arbitrary position of the original image 1 displayed in the first image display area 58a is clicked with the mouse 7, a pin marked with "A" is displayed. Then, in the threshold value setting area 58i, the same pin with "A" is displayed in the place corresponding to the pixel value of the clicked point of the original image 1, and the pixel value of the clicked point of the original image 1 is displayed. You can know. If you click another point, the pin labeled "B" will be displayed, and you can know the pixel value at that position. The pixel values of the A pin and the B pin can be known separately. Three or more pins can be displayed. When the pin is displayed in the original image 1, the pin is displayed at the corresponding position of the other inspection images in the same manner. When the inspection image is switched, the pin can be displayed at the corresponding position of the inspection image after the switching.

The "pin" is an example of indicating that the user has specified an arbitrary position of the inspection image, and the display shape does not have to be pin-shaped, for example, an arrow or a flag. You may.

Further, in the threshold value setting area 58i, a histogram which is a frequency distribution of pixel values can also be displayed. The histogram can be generated by the histogram generation unit 42a of the control unit 4 shown in FIG. The histogram generation unit 42a generates a histogram of the region including the position of the non-defect portion or the defect portion received by the position designation reception unit 48.

In the threshold value setting area 58i, two threshold value display lines 58j indicating the upper limit and the lower limit of the pixel value range to be the non-defective portion are displayed. While looking at the histogram displayed in the threshold value setting area 58i, for example, the user drags the threshold value display line 58j with the mouse 7 to set the pixel value range and the defective part to be non-defective parts. At least one of the power pixel value ranges can be arbitrarily changed. This drag operation can be accepted by the operation reception unit 46. That is, on the histogram generated by the histogram generation unit 42a, it is possible to accept at least one change of the pixel value range to be the non-defective part and the pixel value range to be the defective part. The setting unit 49 sets a pixel value range to be a non-defect unit for a plurality of inspection images based on the changed pixel value range received by the operation reception unit 46.

The method of setting the threshold value is not limited to the above-mentioned method, and various methods can be used. For example, the threshold value can be set by the brightness ratio of the original image 1 and the original image 2 instead of the simple pixel value range. Further, the threshold value can be set by combining a plurality of conditions, and in this case, the logical operation (OR or NOT) may be customized by the user.

It also uses the distances used in statistics, such as the Mahalanobis distance, to determine the brightness of the processed image and generates a gray image in a statistically rational way, depending on the degree of deviation from the threshold. You can also do it.

Further, the operation of selecting the background and the operation of forcibly extracting can be a method of using the minimum value and the maximum value in the predetermined area, but even the pixel value in the predetermined area greatly deviates from the other pixel values. The pixel value may be removed for calculation. Further, the standard deviation of the pixel values within a predetermined region can be used to enable robust operation even if there is a value such as an outlier. The predetermined area may be a rectangular area, or area segmentation may be used. As a result, it is possible to acquire a large number of pixels that are similar to some extent, so that it becomes easy to automatically set the threshold value.

Further, clustering may be used to automatically determine the pixel value range of the non-defective portion. This eliminates the need for multiple clicks. Further, an appropriate threshold value may be set by designating an area including a defect portion and an area without a defect portion and finding an outlier corresponding to the defect from them.

It is also possible to compare the background selection and the distribution of forced extraction to determine a threshold value suitable for separating the defective portion and the non-defective portion. It is also possible to compare the distributions of background selection and forced extraction to indicate whether the two groups are separable. It is also possible to compare the distribution of background selection and forced extraction, determine which image type is most suitable for separating the two groups, and set the threshold value for that image type.

Further, when the background is selected or the forced extraction is performed, the display of the inspection image may be binarized or displayed in gray scale. In the case of binarization, the non-defective part can be set within the threshold range of the inspection image, and the other defective part can be set.

(Control of the pattern light illumination unit by the control unit 41)
In this embodiment, as shown in FIG. 23, the deflection metric processing is performed after obtaining eight luminance images, but the algorithm of the deflation metric processing is the irradiation of the pattern light emitted from the pattern light illuminating unit 2. It is configured after the order is tentatively specified. In this embodiment, the irradiation order of the pattern light on the algorithm is the Y-direction pattern light when 0 °, the Y-direction pattern light when 90 °, and the Y-direction pattern light when 180 ° are shown in FIG. 5A. Y-direction pattern light at 270 °, X-direction pattern light at 0 ° shown in FIG. 5B, X-direction pattern light at 90 °, X-direction pattern light at 180 °, X at 270 ° The algorithm is configured assuming that the order of direction pattern light is.

Here, when the image inspection device 1 is in operation, the moving work W is imaged by the image pickup unit 3, but at the time of this image pickup, the positional relationship between the pattern light illumination unit 2 and the image pickup unit 3 is not in a predetermined state. Can be considered. In particular, in this embodiment, since the pattern light illumination unit 2 and the image pickup unit 3 are separate bodies, the user can freely install the pattern light illumination unit 2 and the image pickup unit 3, so that the pattern light can be used. It is assumed that it is likely that the shift direction of the phase of the illuminance distribution is set to be different from the direction specified by the algorithm.

For example, in the algorithm, the phase shift direction of the illuminance distribution of the pattern light is specified as the X direction from the first irradiation to the fourth irradiation, but in the actual installation state, it is the fourth from the first irradiation. If the phase shifts in the Y direction until the irradiation of, it is possible to obtain a plurality of luminance images by repeating illumination and imaging, but the inspection is performed based on the luminance images captured under inappropriate conditions. Image will be generated. For this reason, the dents existing in the work W are less likely to appear in the inspection image, the dents that should be displayed in black are displayed in white on the contrary, or the dents are displayed in a state where white and black are mixed even though they are all dents. There are times.

Specifically, as shown in FIG. 34, when the phase shift direction of the illuminance distribution of the pattern light is the direction specified by the algorithm, the recessed defect portion is shown on the left side in the figure. Is displayed on the inspection image so that is black. However, as shown in the central portion in the figure, when the orientations of the pattern light imaging unit 2 and the imaging unit 3 differ by 180 ° from the direction assumed by the algorithm, a dented defect portion is formed. It is displayed in white. Therefore, the user may mistakenly recognize it as a convex defect portion. Further, as shown on the right side in the figure, when the positional relationship between the pattern light imaging unit 2 and the imaging unit 3 is actually received by reflected light but is set to receive transmitted light, a white portion is formed. And the black part will be displayed together.

On the other hand, in this embodiment, the information acquisition unit 9 provides information on the moving direction of the work W with respect to the arrangement direction of the light receiving element 3a of the line camera 31 of the imaging unit 3, the light receiving element 3a, and the pattern light illumination unit. Obtaining the positional relationship information related to the positional relationship with 2, the control unit 41 obtains the moving direction information and the positional relationship information acquired by the information acquisition unit 9, and the phase of the pattern light emitted by the pattern light illumination unit 2. The shift direction can be determined.

That is, even if the pattern light illumination unit 2 and the image pickup unit 3 are installed so that the shift direction of the phase of the illuminance distribution is different from the direction specified by the algorithm, the movement direction information and the positional relationship information Based on the above, the phase shift direction of the pattern light emitted by the pattern light illumination unit 2 can be reset to the direction specified by the algorithm.

Therefore, it is not necessary to re-install the pattern light illumination unit 2 and the image pickup unit 3. Further, there is a case where the pattern light illumination unit 2 and the image pickup unit 3 can be installed only so that the shift direction of the phase of the illuminance distribution is different from the direction specified by the algorithm due to the restriction of the installation location and the wiring. However, it is possible to perform image inspection by applying the principle of deflectometry.

That is, when the pattern light illumination unit 2 and the image pickup unit 3 are installed so that the shift direction of the phase of the illuminance distribution is different from the direction specified by the algorithm, an appropriate inspection image is obtained. Therefore, the irradiation order of the pattern light can be changed by controlling the first light emitting diode rows A1 to A12 and the second light emitting diode rows B1 to B12 so that an appropriate inspection image can be obtained.

(Generation of inspection image by image generation unit 42)
The image generation unit 42 can also generate an inspection image related to the shape of the work W according to the movement direction information and the positional relationship information acquired by the information acquisition unit 9. As a result, when the pattern light illumination unit 2 and the image pickup unit 3 are installed so that the shift direction of the phase of the illuminance distribution is different from the direction specified by the algorithm, the pattern of the pattern light illumination unit 2 An appropriate inspection image can be obtained by the generation method of the image generation unit 42 without changing the light irradiation order.

By obtaining the movement direction information and the positional relationship information, it is possible to grasp how the pattern light is irradiated to the work W, and in what direction the phase shift direction of the pattern light is. Even so, the image generation unit 42 can handle the image and generate an inspection image related to the shape of the work W so as to match the shift direction of the phase of the pattern light specified by the algorithm. For example, in the algorithm, the shift direction of the phase of the illuminance distribution of the pattern light is specified as the X direction from the first irradiation to the fourth irradiation, and is specified as the Y direction from the fifth irradiation to the eighth irradiation. However, in the actual installation state, the phase shifts in the Y direction from the first irradiation to the fourth irradiation, and the phase shifts in the X direction from the fifth irradiation to the eighth irradiation. In this case, the pattern light irradiation order is not changed, and the image is taken as it is, and after the imaging, the luminance image obtained by the fifth irradiation to the eighth irradiation is obtained by the first irradiation to the fourth irradiation according to the algorithm. The luminance image obtained by the first irradiation to the fourth illumination is treated as the luminance image obtained by the fifth irradiation to the eighth irradiation according to the algorithm. In this way, by exchanging the actually captured luminance images and processing them, it is possible to perform an image inspection applying the principle of deflection metrics without changing the installation state of the pattern light illumination unit 2 and the image pickup unit 3. can.

(Structure of inspection department)
As shown in FIG. 2, the control unit 4 is provided with an inspection unit 4a. The inspection unit 4a executes a defect inspection of the work W based on the inspection image generated by the image generation unit 42. For example, it can be determined whether or not the work W has a non-defective portion based on the above threshold value, and if it is determined that the work W has a non-defect portion, it can be notified.

(HDR function)
The image inspection apparatus 1 can incorporate a high dynamic range imaging (HDR) function. In the HDR function, the control unit 41 controls the pattern light illumination unit 2 and the image pickup unit 3 so as to obtain a plurality of luminance images having different brightness. Examples of the luminance image obtained by the HDR function include a plurality of luminance images having different exposure times, a plurality of luminance images having different emission intensities of the pattern light illumination unit 2, and a plurality of luminance images having different exposure times and emission intensities. .. The image generation unit 42 can generate an inspection image having a dynamic range wider than the dynamic range of each luminance image by, for example, synthesizing three luminance images having different brightness. As the HDR synthesis method, a well-known method can be used.

In the HDR function, it is necessary to obtain a plurality of luminance images, and in this case as well, as described in the section of transmitting the trigger signal, when one encoder pulse signal from the outside is received, the illumination condition of the pattern light illumination unit 2 is obtained. A plurality of imaging trigger signals are sequentially transmitted to the pattern light illumination unit 2 and the imaging unit 3 so that at least one of the imaging conditions of the imaging unit 3 is changed to generate a plurality of luminance images. be able to.

(Multi-spectral lighting)
The pattern light illumination unit 2 may be configured to enable multispectral illumination. Multispectral illumination is to irradiate the work W with light having different wavelengths at different timings, and is suitable for inspecting color unevenness, stains, etc. of printed matter (inspection object). For example, the pattern light illumination unit 2 can be configured so that the work W can be irradiated with yellow, blue, and red in order. Specifically, the pattern light illumination unit 2 having LEDs of many colors can also be used. Alternatively, the pattern light illumination unit 2 may be configured by a liquid crystal panel, an organic EL panel, or the like.

The imaging unit 3 images the work W at the timing when light is irradiated to obtain a plurality of luminance images. The image generation unit 42 can obtain an inspection image by synthesizing a plurality of luminance images. Here, the light can also include ultraviolet rays and infrared rays.

(Action and effect of the embodiment)
As described above, according to this embodiment, a plurality of pattern lights are sequentially irradiated to the work W, and each time the pattern light is irradiated, the work W is imaged to generate a plurality of brightness images, and a plurality of brightness images are generated. Since phase data indicating the shape of the work W is generated based on the image and an inspection image related to the shape of the work W is generated, image inspection can be performed by applying the principle of reflectance.

Even if the pattern light illumination unit 2 and the image pickup unit 3 are installed so that the shift direction of the phase of the illuminance distribution is different from the direction assumed by the algorithm, the work with respect to the arrangement direction of the light receiving element 3a. According to the movement direction information regarding the movement direction of W and the positional relationship information regarding the positional relationship between the light receiving element 3a and the pattern light illumination unit 2, the phase shift direction of the pattern light emitted by the pattern light illumination unit 2 is set to an appropriate direction. Can be.

Further, even if the pattern light illumination unit 2 and the image pickup unit 3 are installed so that the shift direction of the phase of the illuminance distribution is different from the direction assumed by the algorithm, the movement direction information and the positional relationship information are described above. Therefore, the image generation unit 42 can generate an inspection image related to the shape of the work W.

Further, a plurality of inspection images can be displayed on the display unit 5 at the same time. When the user looks at a certain inspection image and specifies the position of the non-defect part or the defective part of the work W while displaying a plurality of inspection images on the display unit 5 at the same time, the specified result is accepted. Based on the pixel value of the position of the non-defect portion or the defective portion, the pixel value range to be the non-defect portion can be set for a plurality of inspection images. As a result, the presence or absence of defects can be determined in a plurality of inspection images without the user designating the positions of the non-defective portion or the defective portion with respect to all the inspection images.

In addition, a region included in all of the pixel value ranges set for a plurality of inspection images to be non-defective parts is defined as a non-defect region, and a region not included in any of the pixel value ranges is defined as a defective region. Since the defect extraction image can be generated, the user can detect different types of defects simply by looking at the defect extraction image.

Further, when a user performs an operation of designating a part area of one inspection image and enlarging it while displaying a plurality of inspection images on the display unit 5 at the same time, the other inspection images are supported. The area to be displayed is enlarged at the same magnification and displayed on the display unit. As a result, a plurality of inspection images can be enlarged and displayed in the same manner with a simple operation. The user can see each inspection image in which the same area is enlarged and displayed at the same magnification, and can specify the position of the non-defective portion or the defective portion of the work W in a certain inspection image.

Further, when generating a plurality of inspection images capable of detecting different types of defects, the user simultaneously views the filtered inspection image and the unfiltered inspection image. be able to.
(Embodiment 2)
In the first embodiment, the function of generating an inspection image by performing the deflection metric processing on the obtained luminance image has been described, but the image inspection apparatus 1 according to the present invention has this deflation metric processing function. In addition, for example, it is possible to have a function of generating an inspection image by using a photometric stereo method.

Hereinafter, in the case of generating an inspection image by using the photometric stereo method, the same parts as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted with reference to FIG. 38, and the different parts will be described in detail. explain.

The image inspection device 1 according to the second embodiment can be configured in the same manner as the image inspection device disclosed in, for example, Japanese Patent Application Laid-Open No. 2015-232486. That is, the image inspection device 1 includes an imaging unit 3 that images the work W from a certain direction, an illumination unit 200 for illuminating the work W from three or more different illumination directions, and a control unit 400. It includes at least a display unit 5, a keyboard 6, and a mouse 7 similar to those in the first embodiment.

The illumination unit 200 is configured to irradiate the work W with light from different directions, and controls the first to fourth light emitting units 201 to 204 and the first to fourth light emitting units 201 to 204. It has a lighting control unit 205 and the like. The illumination unit 200 is a portion that executes multi-directional illumination that irradiates the work W with light from different directions. The first to fourth light emitting units 201 to 204 are arranged so as to surround the work W at intervals from each other. Light emitting diodes, incandescent bulbs, fluorescent lamps, and the like can be used for the first to fourth light emitting units 201 to 204. Further, the first to fourth light emitting units 201 to 204 may be separate bodies or may be integrated.

The control unit 400 includes a control unit 401, a normal vector calculation unit 402, a contour image generation unit 403, a texture drawing image generation unit 404, and a trigger signal transmission unit 405. The control unit 401 is configured to be able to receive the encoder pulse signal when the encoder pulse signal is input from the outside. When the control unit 401 receives one encoder pulse signal from the outside, a trigger signal transmission is performed so that at least one of the illumination condition of the illumination unit 200 and the imaging condition of the imaging unit 3 is changed to generate a plurality of luminance images. The unit 405 is configured to sequentially transmit a plurality of imaging trigger signals to the imaging unit 3. Further, when the control unit 401 receives one encoder pulse signal from the outside, the trigger signal transmission unit 405 is configured to sequentially transmit a plurality of illumination trigger signals to the illumination unit 200. In this embodiment, since the first to fourth light emitting units 201 to 204 are sequentially turned on, the lighting trigger signal is transmitted four times. The imaging trigger signal is transmitted four times in synchronization with the transmission of the illumination trigger signal.

For example, when the lighting unit 200 receives the first lighting trigger signal, the lighting control unit 205 lights only the first light emitting unit 201. At this time, the imaging unit 3 receives the imaging trigger signal and images the work W at the timing when the light is irradiated. When the illumination unit 200 receives the second illumination trigger signal, the illumination control unit 205 lights only the second light emitting unit 202, and at this time, the imaging unit 3 images the work W. In this way, four luminance images can be obtained. The number of lights is not limited to four, but can be any number as long as the number of lights is three or more and the work W can be illuminated from different directions.

The normal vector calculation unit 402 calculates the normal vector for the surface of the work W of each pixel by using the pixel values for each pixel that are in a corresponding relationship between the plurality of luminance images captured by the image pickup unit 3. The contour image generation unit 403 performs differential processing on the calculated normal vector of each pixel in the X and Y directions to generate a contour image showing the contour of the inclination of the surface of the work W. The texture drawing image generation unit 404 calculates the albedo of each pixel in the same number as the normal vector from the calculated normal vector of each pixel, and shows a pattern in which the tilted state of the surface of the work W is removed from the albedo. Generate a texture drawing image.

Although not shown, also in the second embodiment, as in the first embodiment, for each inspection image displayed on the display unit 5, a non-defective portion or a defective portion of the inspection object in the inspection image is used. Based on the pixel value of the position of the reception unit that can accept the designation of the position, the non-defective part or the defective part received by the reception unit, the pixel value range to be the non-defect part is set for a plurality of inspection images. It has a setting unit. Further, among the inspection images displayed on the display unit 5, an operation reception unit capable of accepting an enlargement operation for designating and enlarging a part of an area of one inspection image, and another displayed on the display unit 5. It also includes a display control unit that enlarges the area corresponding to the area designated by the above-mentioned one inspection image in the inspection image at the same magnification in conjunction with the above-mentioned enlargement operation and displays it on the above-mentioned display unit. Of course, it is possible to perform a reduction operation and a scroll operation instead of the enlargement operation. Furthermore, the filter processing unit (not shown) is the same as that in the first embodiment.

(Action and effect of the embodiment)
According to the second embodiment, a plurality of inspection images capable of detecting different types of defects can be obtained by using the photometric stereo method, and defect inspection is performed using the obtained inspection images. be able to.
(Embodiment 3)
FIG. 39 is a block diagram of the image inspection apparatus 1 according to the third embodiment of the present invention. This embodiment 3 is different from the first embodiment in that the image inspection apparatus 1 is provided with the edge generation unit 500, but the other parts are the same as those of the first embodiment, and thus the same as the first embodiment. The parts will be given the same sign, the explanation will be omitted, and the different parts will be explained in detail.

When inspecting the work W, for example, there are cases where it is desired to inspect a defect in the contour of a complicated shape, or when it is desired to inspect a defect in a specific area having a complicated shape. It is necessary to extract the contour accurately on the image. The edge generation unit 500 is a part for generating the outline of the work W on the screen displayed on the display unit 5, and can be a part of the control unit 4.

The contour of the work W is also referred to as an edge, and the edge is not only the outer line of the work W, but also the peripheral edge of a hole or a recess formed in the work W, the edge of the groove formed in the work W, and the work W. In some cases, the outline of the bead provided, the outline of the seal member provided on the work W, the edge of the sealant applied to the work W, etc., and basically any edge is displayed on the screen in this embodiment. It can be generated and confirmed above.

The edge generation unit 500 has a position reception unit 501 that accepts the designation of an arbitrary position that is an initial position for edge detection of the work W on the inspection image displayed on the display unit 5, and an initial position reception unit 501 that receives the designation. The edge tangent line is a direction substantially orthogonal to the slope direction of the edge detected by the edge detection unit 502 and the edge detection unit 502 that detects the position or the initial edge position in the vicinity of the initial position and specifies the slope direction of the edge. A search area setting unit 503 that sets a search area, which is an area for searching an edge at a position separated by a predetermined distance in the edge tangential direction from the detected edge, and a search area set by the search area setting unit 503, assuming a direction. It is provided with a local threshold value setting unit 504 for setting a local threshold value for edge position detection effective in the search area based on the pixel value in the search region, and an edge connecting unit 505.

The edge detection unit 502 detects the edge position in the search area based on the pixel value in the search area and the local threshold value for edge position detection, and executes the edge detection process for specifying the gradient direction of the edge. Then, the edge detection process is configured to be repeatedly executed for the search area sequentially set by the search area setting unit 503. The edge connecting portion 505 is configured to specify the contour of the work W by executing a connecting process for connecting the edge positions sequentially detected in each search area.

In this embodiment, the edge detection unit 502 detects the edge position and specifies the slope direction of the edge. However, this is a process for assuming the edge tangential direction, and the slope direction of the edge is determined. When the edge tangential direction is assumed without specifying, it is not necessary to specify the slope direction of the edge.

Hereinafter, a specific edge generation method by the edge generation unit 500 will be described with reference to the flowchart shown in FIG. 40. In step SD1 of this flowchart, i = 1 is set. After that, the process proceeds to step SD2, in which the user specifies an arbitrary position as the initial position for edge detection of the work W on the inspection image displayed on the display unit 5 by clicking the mouse 7. do. The method of specifying the initial position is not limited to the operation of the mouse 7, and may be the operation of another input device.

For example, on the inspection image shown in FIG. 41, a pointer (indicated by a white arrow) is placed on a line recognized by the user as an edge, and the mouse 7 is clicked and operated. Since it is considered that the user specifies a part where the edge is easy to understand on the inspection image, the initial position can be accurately specified by having the user point out the initial position.

The click-operated position is accepted by the position reception unit 501. Since the clicked position may be slightly off the edge, the edge detection unit 502 has the coordinates with the strongest edge strength in the region S0 in the predetermined range including the clicked position as shown in FIG. 42. The coordinates of the portion where the difference between the pixel values is large are obtained, and the coordinates corresponding to the position are obtained and set as the initial coordinates P0 (step SD3). As a result, even if the clicked position is slightly off the edge, the position can be corrected and the initial coordinates can be placed on the edge. Further, in step SD3, the gradient direction of the edge (the direction in which the pixel value changes) is also specified, and as shown on the left side of FIG. 43, the gradient vector n that passes through the initial coordinates P0 and indicates the gradient direction of the edge is set. Ask. This is done by the edge detection unit 502. If no edge exceeding the threshold value is found, the process ends.

In step SD4 following step SD3, a tangent vector r (tangent vector at the initial coordinate P0) substantially orthogonal to the gradient vector n is obtained, and this tangent vector r is used as a traveling vector (unit vector). The direction of the travel vector can be assumed to be the edge tangential direction.

After that, the process proceeds to step SD5, and as shown on the right side of FIG. 43, the coordinates (point Si) of the positions separated by the progress vector r × L (predetermined distance) are obtained. In step SD6, the segment S1 centered on the point Si is formed parallel to the traveling vector r. This segment S1 corresponds to a search area which is an area for searching for edges. This is done by the search area setting unit 503.

After forming the segment S1, in step SD7, the pixels in the segment S1 shown on the left side of FIG. 44 are projected (degenerate) in the direction orthogonal to the edge gradient direction, so that the pixels are one-dimensional as shown in the center of the figure. The pixel value is set, and this one-dimensional pixel value is obtained as edge information. The edge position is obtained from the obtained edge information. Since it is binarization within the segment S1, the threshold value for binarization can be determined in a narrow area. The threshold value obtained in the segment S1 is the local threshold value for edge position detection. The local threshold value for edge position detection can be obtained by the Otsu method (discriminant analysis method). This is done by the local threshold setting unit 504. Further, for example, a local threshold value for detecting the edge position may be set with respect to the differential value of the one-dimensional pixel value, and the edge position may be obtained.

In step SD8, it is determined whether or not the edge position has been obtained. If NO is determined in step SD8, the process returns to step SD5 through steps SD16 and 17. If the edge position cannot be found even after trying multiple times, the process ends.

On the other hand, if YES is determined in step SD8 and the edge position is obtained, the process proceeds to step SD9, and the obtained edge position is set as Pi as shown on the right side of FIG. 43. In step SD10, the progress vector is updated to the unit vector from Pi-1 to Pi.

After that, in step SD11, a segment is formed in parallel with the progress vector centering on the point Pi obtained in step SD9. In step SD12, the segment formed in step SD11 is degenerate to binarize, edge information is obtained, and the edge position is obtained from this edge information. In step SD13, the edge position obtained in step SD12 is set as Pi. In step SD14, the progress vector is updated to the unit vector from Pi-1 to Pi. In step SD15, i = i + 1. Then, the process proceeds to step SD5.

That is, as shown in FIG. 45A, a plurality of edge positions (points P0 to P10 in this example) can be obtained by repeatedly executing the edge detection process for the search areas that are sequentially set. After obtaining a plurality of edge positions, the edge connecting portion 505 connects the edge positions with a line in the obtained order (P0 → P1 → P2 ... P10 → P0), thereby contouring the work W. Can be identified. At this time, it may be determined whether or not the distance between P0 and P10 is equal to or less than a predetermined distance, and if it is equal to or less than a predetermined distance, a method of connecting P0 and P10 may be adopted.

When connected from P10 to P0, the closed area flag can be given and the process can be terminated. P0 is the start point and the end point of the contour. In some cases, the contour is not a closed area. The outline of the work W identified in this way can be displayed in a state of being synthesized on the inspection screen in any color other than white and black.

FIG. 45B shows a case where the contour is already specified by the white dots (edge positions), and then P0 to P3 are sequentially obtained as the edge positions. After obtaining P3, it is preferable to stop the process so as not to intersect the contour already specified. Further, after obtaining P3, the process may be terminated by connecting to P0 instead of the white dot, whereby stray can be avoided. Further, it may be a closed region composed of only the edge positions indicated by the white dots in FIG. 45B. The white dots on the drawing are shown only for convenience of explanation, and the color and form can be set arbitrarily, but the already obtained edge position and the edge position obtained after that are It is preferable to display in different colors and forms.

FIG. 45C shows a case where the contour is already specified by the white point (edge position) and then P0 and P1 are obtained as the edge positions. If the contour already specified exists in the segment S1 formed after obtaining P1, the process can be stopped. If P0 exists in the segment S1, it may be connected to P0 without stopping the processing.

FIG. 46 is a diagram showing a processing method when the curvature of the contour is small. As shown in the upper part of FIG. 46, after the edge position P1 is obtained, the segment S1 is formed at a distance L by a predetermined distance in the traveling vector direction in order to obtain the next edge position. If the curvature of the contour is small, the contour is located outside the segment S1, and the edge position may not be detected.

As described above, when the edge position cannot be detected in the segment S1, the search area setting unit 503 is configured to shorten the predetermined distance L and reset the segment S1. As shown in the second figure from the top of FIG. 46, the predetermined distance L can be set to, for example, L / 2. The edge detection unit 502 detects the edge position in the segment S1 reset by shortening the predetermined distance L, and specifies the gradient direction of the edge. If an edge cannot be detected in the segment S1 even if the predetermined distance is set to L / 2, the predetermined distance is further shortened to form the segment S1.

In the second figure from the top of FIG. 46, when the pixels in the segment S1 are projected, the edges are projected diagonally with respect to the contour, so that the edges may be blurred and the exact edges may be difficult to understand. On the other hand, as shown at the bottom of FIG. 46, the edge detection unit 502 sets the edge position detected in the segment S1 reset by the search area setting unit 503 as the temporary edge position P2 and sets the search area. The unit 503 sets a segment S1'(reset search area) extending in a direction orthogonal to the vector t connecting the temporary edge position P2 and the edge position P1 detected immediately before the temporary edge position P2. As a result, the edges become clear when the pixels in the segment S1'are projected. The edge detection unit 502 is configured to detect the edge position in the segment S1', specify the gradient direction of the edge, and replace the detected edge position with the temporary edge position P2.

Further, in an inspection object as shown in FIG. 47, for example, a bead, a sealant, an O-ring, etc., the inspection object has a predetermined width C and has contours on both sides in the width direction. In the present embodiment, the contours on both sides in the width direction can be specified by applying the method described above. Hereinafter, the contour identification method on both sides will be described in detail based on the flowchart of FIG. 48.

In step SE1 of the flowchart of FIG. 48, i = 1 is set. Then, the process proceeds to step SE2. In step SE2, the user designates an arbitrary position as the initial position for edge detection of the inspection object on the inspection image displayed on the display unit 5 by clicking the mouse 7. For example, on the inspection image shown in FIG. 47 (partially shown), a pointer (indicated by a white arrow) is placed at a position recognized by the user as the center of the inspection object in the width direction. , Click the mouse 7. The clicked position is accepted by the position reception unit 501.

In step SE3, the coordinates with strong edge strength are obtained within the region S0 (shown in FIG. 49) in the predetermined range including the clicked position, and the first edge position E0 is set. In step SE4, the gradient vector n is traced to obtain the corresponding edge position E0'. The direction of the gradient vector n'at the edge position E0'is opposite to that of the gradient vector n. The intermediate position between the edge position E0 and the edge position E0'is set as the initial coordinate (control point) P0.

In step SE5, the progress vector r is calculated by averaging the gradient vector n at the edge position E0 and the gradient vector n'at the edge position E0'. In step SE6, as shown in FIG. 50, the coordinates (point Si) of the positions separated by the traveling vector r × L (predetermined distance) are obtained. In step SE7, the segment S1 centered on the point Si is formed parallel to the traveling vector r.

In step SE8, one-dimensional pixel values are obtained as edge information by projecting (degenerating) the pixels in the segment S1 in a direction orthogonal to the edge gradient direction, and edge positions Ei and Ei'are obtained from this edge information. In step SE9, it is determined whether or not the edge position has been obtained. If NO is determined in step SE9, the process returns to step SE6 through steps SE17 and 18. If the edge position cannot be found even after trying multiple times, the process ends.

On the other hand, if YES is determined in step SE9 and the edge positions Ei and Ei'are obtained, the process proceeds to step SE10 and the intermediate position between the edge positions Ei and Ei'is set as Pi. In step SE11, the progress vector is updated to the unit vector from Pi-1 to Pi. After that, in step SE12, a segment is formed in parallel with the traveling vector centering on the point Pi obtained in step SE10. In step SE13, the segment formed in step SE12 is degenerate to binarize, edge information is obtained, and the edge position is obtained from this edge information. In step SE14, the edge position obtained in step SE13 is set as Pi. In step SE15, the progress vector is updated to the unit vector from Pi-1 to Pi. In step SE16, i = i + 1. Then, the process proceeds to step SE6. Therefore, by repeatedly executing the edge detection process on both sides in the width direction of the inspection target for the search areas that are sequentially set, it is possible to obtain a plurality of edge positions on both sides. After determining the edge position, the edge connecting portion 505 connects the edge positions with a line, so that the contours on both sides can be specified.

The edge can be generated at the time of setting the image inspection device 1, and the generated edge can be stored in the storage device 10 as a reference model line. After setting the image inspection device 1, when operating the image inspection device 1, the inspection object is imaged to obtain an inspection image, the position of the inspection object in the inspection image is corrected, and then the inspection image is displayed. The edge of the inspection target object is detected, and the detected edge is compared with the reference model line, and when the difference is greater than or equal to a predetermined value, it can be identified as a defective portion. This can be performed by the inspection unit 4a.

(Action and effect of the embodiment)
According to the third embodiment, when the user specifies the edge of the work W or the position in the vicinity thereof on the inspection image displayed on the display unit 5, the search area is set and the threshold value in the search area is set. , The detection of the edge position in the search area is repeated, and a plurality of edge positions can be sequentially obtained along the contour of the work W, and the obtained edge positions are connected to automatically obtain the contour of the work W. Can be identified. Therefore, the contour can be specified with high accuracy even with a simple operation, and the inspection accuracy can be improved.
(Embodiment 4)
FIG. 51 is a block diagram of the image inspection apparatus 1 according to the fourth embodiment of the present invention. This embodiment 4 is different from the first embodiment in that the image inspection apparatus 1 is provided with the edge generation unit 300, but the other parts are the same as those of the first embodiment, and thus the same as the first embodiment. The parts will be given the same sign and the explanation will be omitted, and the different parts will be explained in detail.

The edge generation unit 300 has an edge detection area setting unit 301 for setting a plurality of edge detection areas at different positions on the inspection image displayed on the display unit 5, and a plurality of edge points in each edge detection area. It is formed by a contour segment forming unit 302 and a contour segment forming unit 302 that form a contour segment forming a part of the contour of an inspection object for each edge detection region by detecting and connecting adjacent edge points. A connection process for connecting the end of one contour segment and the end of the other contour segment closest to the end is executed, and the connection process is repeated until there are no open ends of the contour segment to form a closed region. It includes a closed region forming portion 303 to be formed.

That is, the upper side of FIG. 52 shows an inspection image obtained by capturing the work W, and the user can see the inspection image and, as shown in the center of FIG. 52, the edge detection region at a different position on the inspection image. Set F1 and F2. The edge detection areas F1 and F2 can be formed by surrounding a desired portion of the work W by, for example, an input operation of the mouse 7 or the keyboard 6 by the user, and the portion surrounded by this operation is set as an edge detection area setting unit. 301 is set as the edge detection areas F1 and F2.

In the edge detection region F1, edge points P1, P2, P3, ... Are detected as shown by enlarging a part of the central figure of FIG. The edge point can be detected by a conventionally known method, or can be detected by a method as in the third embodiment. By connecting adjacent edge points (for example, P1 and P2) to each other, a contour segment K1 forming a part of the contour of the work W is formed. Similarly, the contour segment K2 is formed in the edge detection region F2. This is done by the contour segment forming unit 302.

The contour segment may be formed by the user specifying it using an input device such as a mouse 7. In this case, the user specifies at least the start point and the end point of the contour segment by operating the mouse 7 while looking at the inspection image displayed on the display unit 5. The contour segment forming unit 302, which has been designated by the user as a start point and an end point, uses a straight line connecting the start point and the end point as the contour segment. In this case, the user may additionally specify an intermediate point between the start point and the end point. Further, the contour segment is not limited to a straight line, and can be formed into a curved line by an operation of a mouse 7 or the like.

As shown on the lower side of FIG. 52, the closed region forming portion 303 connects the end portion K1a of the contour segment K1 and the end portion K2a of the contour segment K2 closest to the end portion K1a by a connecting line L1 formed of a straight line. do. Further, the end portion K1b of the contour segment K1 and the end portion K2b of the contour segment K2 closest to the end portion K1b are connected by a connecting line L2 formed of a straight line. As a result, the open ends of the contour segments K1 and K2 are eliminated, and the closed region M is formed. This closed area M becomes an inspection target area, and the inspection unit 4a becomes an area for executing defect inspection.

The separation distance between the end K1a of the contour segment K1 to be connected and the end K2a of the contour segment K2 is set longer than the distance between adjacent edge points (for example, P1 and P2). .. As a result, the end portion K1a of the contour segment K1 and the end portion K2a of the contour segment K2 are less likely to be erroneously recognized as edge points.

Further, when it is desired to set substantially the entire work W as the inspection target area in the work W as shown in FIG. 53, as shown in the upper side of the figure, the edge detection areas F1 to F6 surround the peripheral portion of the work W. To form. Contour segments K1 to K6 are formed in the edge detection regions F1 to F6.

After the contour segments K1 to K6 are formed, as shown in the lower part of the figure, the ends of the contour segments K1 to K6 are connected by connecting lines L1 to L6 to form a closed region M. In this example, since the closed region M can be formed so as to remove the corner portion of the work W, for example, an inspection target region can be formed by removing the chamfered portion and the R processed portion of the corner portion of the work W.

In both the case shown in FIG. 52 and the case shown in FIG. 53, when connecting the contour segments, it may be possible to connect after all the edge detection regions are formed, or the edge detection regions are formed. It may be connected in order. Further, when connecting the contour segments, it is possible to prevent the connection process from being performed on the already connected ends. It is also possible to prevent contour segments that intersect each other from being connected to each other. Further, it is also possible to perform a connection process so that the connection line is the shortest. Furthermore, the connecting ends may be configured to be manually selectable and changeable by the user.

It is also possible to prevent the connection process from being performed when the connection line for connecting the contour segments exceeds a predetermined length. In this case, the contour segment does not form a closed area. If the length of the connecting line is longer than the predetermined length, it means that the two contour segments are far apart from each other. In this case, by not performing the connecting process, an erroneous connecting process is performed. You can avoid it. When the user can manually select the end to be connected, it is preferable that the connecting line can be connected even if the length of the connecting line exceeds a predetermined length.

Further, the connecting line is not limited to a straight line, and may be, for example, a quadratic curve, a cubic curve, or the like, particularly a Pezier curve, a spline curve, or the like. As a result, the ends of the contour segments can be smoothly connected. Further, the connecting line may be a combination of a curved line and a straight line. It is also possible to connect one contour segment and another contour segment by a curve having a common tangent vector with the tangent vector at the end of the one contour segment.

FIG. 54 is a diagram for explaining a case where one work W is imaged in two parts. The imaging unit 3 images a portion above the imaging boundary line R1 of the work W to obtain an image for upper partial inspection (image for first partial inspection), and is below the imaging boundary line R2 of the work W. The portion is imaged to obtain an image for lower partial inspection (image for second partial inspection). Therefore, the predetermined range between the imaging boundary line R1 and the imaging boundary line R2 in the work W is an overlapping portion included in the imaging range of both the upper partial inspection image and the lower partial inspection image. The setting of the imaging range of the work W is not limited to the above example, and may be an imaging mode in which the work W is divided in the left-right direction, or an imaging mode in which the work W is divided into three or more.

In this case, as shown on the left side of FIG. 54, edge detection regions F1 to F6 are formed in the upper partial inspection image. At this time, in the upper partial inspection image, the work W exists until it touches the imaging boundary line R1, but the operation of extending the lower ends of the edge detection regions F1, F3, F4, and F6 to the limit where they touch the imaging boundary line R1. Since it is difficult to perform the above on the upper partial inspection image, when the user performs an operation of extending the lower end of the edge detection areas F1, F3, F4, and F6, the operation is to stop before touching the imaging boundary line R1. Can be considered.

On the other hand, in the present embodiment, the lower end portions of the edge detection regions F1, F3, F4, and F6 reach the overlapping portion of the upper partial inspection image with the lower partial inspection image, and the contour segment K1 , K3, K4, K6 are located in the overlapping portion, the lower end of the contour segment K1 and the lower end of the contour segment K4 are connected by the connecting line L1 and the lower end of the contour segment K3. And the lower end of the contour segment K6 are connected by a connecting line L2. Thereby, the closed region M can be formed. As shown on the right side of FIG. 54, the same processing can be performed on the lower partial inspection image.

That is, a plurality of contour segments (K1, K3, K4, K6) are formed in an overlapping portion of one partial inspection image (upper partial inspection image) with another partial inspection image (lower partial inspection image). When the end portion is located, even if there is no edge detection region at the end portion of the contour segment, the process of connecting the end portions of the contour segment with a connecting line can be automatically performed.

After that, the portion other than the closed region M is blackened to be the non-inspection target region, while the closed region M (diagonal line region) is whitened to be the inspection target region. Similarly, the non-inspection target area and the inspection target area can be set in the lower partial inspection image. As a result, even when the work W is separately imaged, the entire range to be inspected can be inspected with a simple operation.

(Action and effect of the embodiment)
According to the fourth embodiment, when a user sets a plurality of edge detection areas, a plurality of edge points can be detected in each edge detection area and the edge points can be connected to form a plurality of contour segments. It is possible to connect the ends of these contour segments to form a closed region, and to perform defect inspection of the work W using this closed region as an inspection target region.

(Other embodiments)
The present invention is not limited to the above embodiments 1 to 4. For example, as in the modification 1 shown in FIG. 55, the image generation unit 42 can be incorporated in the image pickup unit 3 as the configuration of the image inspection device 1. As a result, the imaging unit 3 performs up to the creation of the inspection image by the deflation metric processing, and the control unit 4 performs other processing. That is, it is possible to make a smart camera in which the image pickup unit 3 has an image processing function.

Further, as in the modification 2 shown in FIG. 56, the pattern light illumination unit 2 and the image pickup unit 3 can be integrated, and the control unit 4 can be separated. In this case, the pattern light illumination unit 2 may have the image pickup unit 3 built-in, or the image pickup unit 3 may have the pattern light illumination unit 2 built-in.

Further, as in the modification 3 shown in FIG. 57, the image pickup unit 3 and the control unit 4 may be integrated into the image pickup unit 3 and the control unit 4 may be incorporated into the image pickup unit 3.

Further, the pattern light illumination unit 2, the image pickup unit 3, and the control unit 4 can all be integrated.

The above embodiments are merely exemplary in all respects and should not be construed in a limited way. Furthermore, all modifications and modifications that fall within the equivalent scope of the claims are within the scope of the present invention.

As described above, the image inspection apparatus according to the present invention can be used when inspecting defects of inspection objects using images obtained by capturing images of various inspection objects.

1 Image inspection device 2 Pattern light Illumination unit 2b Lighting condition storage unit 3 Imaging unit 3a Light receiving element 4a Inspection unit 5 Display unit 9 Information acquisition unit 10 Storage device 31 Line camera 41 Control unit 42 Image generation unit W Inspection object (work)

Claims (5)

In the image inspecting apparatus for inspecting defects of the inspection object using an image obtained by capturing an image of a test object which moves in the Y-axis direction,
The pattern light having a periodic intensity distribution and pattern light illumination unit for irradiating the inspection object,
An imaging unit having a plurality of light receiving elements arranged in a line and having a line camera that can be installed so that the arrangement direction of the light receiving elements is the X-axis direction orthogonal to the Y-axis.
The phase of the illuminance distribution and the phase of the illuminance distribution is shifted to the Y-axis direction together with the sequentially irradiating the inspection object with a plurality generate X-direction pattern light having a different phase obtained by shifting to the X-axis direction A plurality of Y-direction pattern lights having different phases are generated and sequentially irradiated to the inspection object, and a plurality of X-direction pattern lights having different phases and a plurality of Y-direction pattern lights having different phases are irradiated, respectively. The inspection object is imaged at the timing to generate a plurality of X-direction luminance images corresponding to the plurality of X-direction pattern lights having different phases, and a plurality of Y-direction pattern lights having different phases are generated. an imaging control unit for controlling the pattern light illumination unit and the imaging unit so that to generate a corresponding plurality of Y-direction luminance image,
Based on the principle of deflection metrics, phase data on the surface of the inspection object is generated from each of the plurality of X-direction luminance images and the plurality of Y-direction luminance images imaged by the imaging unit, and the generated phase data respectively. an image generator for generating an inspection image indicating the shape of the inspection object on the basis of,
And a checking unit for performing a defect inspection of the inspection object on the basis of the inspection image generated by the image generating unit,
The imaging control unit, while being configured to accept one of the encoder pulse signal from the outside, when receiving the encoder pulse signal of the one, a plurality of X-direction pattern light the pattern light illumination unit has the above different phases A plurality of illumination trigger signals are sequentially transmitted to the pattern light illumination unit so as to sequentially irradiate a plurality of Y-direction pattern lights having different phases, and the imaging is performed in synchronization with the plurality of illumination trigger signals. An image inspection device characterized in that it is configured to sequentially transmit a plurality of imaging trigger signals to a unit. In the image inspection apparatus according to claim 1,
The image inspection device further includes a UI generation unit that generates a display screen for displaying the inspection image displayed on the display unit.
The image generation unit generates a normal reflection component image from the normal reflection components included in each pixel of the plurality of X-direction luminance images and the plurality of Y-direction luminance images based on the principle of deflection methylation, and also generates the normal reflection component image. A diffuse reflection component image is generated from the diffuse reflection component contained in each pixel of the plurality of X-direction luminance images and the plurality of Y-direction luminance images.
The display screen is configured so that the specular reflection component image and the diffuse reflection component image can be displayed as the inspection image, and on the display screen, the inspection unit selects an inspection image to be used for defect inspection. An image inspection device characterized by accepting. In the image inspection apparatus according to claim 2,
The image generation unit generates the positive reflection component based on the difference of each pixel of the X-direction luminance image corresponding to each set of the X-direction pattern light having the opposite phases among the plurality of X-direction luminance images. In addition to the calculation, the average value is calculated by adding each pixel of the X-direction luminance image corresponding to each set of the X-direction pattern light having the opposite phases, and the average value calculated from the X-direction luminance image and the positive The diffuse reflection component is calculated based on the reflection component, and the difference between each pixel of the Y-direction luminance image corresponding to each set of the Y-direction pattern light having the opposite phases among the plurality of Y-direction luminance images. Based on this, the normal reflection component is calculated, and the average value is calculated by adding each pixel of the Y-direction luminance image corresponding to each of the Y-direction pattern light sets having opposite phases, and the average value is calculated from the Y-direction luminance image. An image inspection apparatus characterized in that a diffuse reflection component is calculated based on the calculated average value and the normal reflection component. In the image inspection apparatus according to any one of claims 1 to 3.
The image pickup control unit is configured to control the pattern light illumination unit and the image pickup unit so as to obtain a plurality of X-direction luminance images having different brightness and a plurality of Y-direction luminance images having different brightness.
The image generation unit synthesizes a plurality of the X-direction luminance images having different brightness, and also synthesizes a plurality of the Y-luminance images having different brightness to obtain a dynamic range wider than the dynamic range of each of the luminance images. An image inspection apparatus characterized in that it is configured to generate the above-mentioned inspection image. In the image inspection apparatus according to any one of claims 1 to 4.
A lighting condition storage unit that stores the lighting conditions of the pattern light illumination unit is incorporated in the pattern light illumination unit.
The pattern light illumination unit receives the illumination trigger signal, an image inspection apparatus characterized by illuminating the test object in accordance with the illumination condition stored in the illumination condition storage unit.

Images