JP2002098650A - Transparent body detection method and system thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a transparent body without depending on a specific material. SOLUTION: The transparent body detection system comprises a light source 11 for polarizing light in a specific direction by a projection-side polarization plate 112 and for projecting the light to a specific region including the transparent body 10, a polarizer 12 that has a light reception side polarization plate 121 for transmitting only specific polarization out of light that is transmitted through the specific region, a television camera 13 for picking up an image obtained by receiving light that is transmitted through the reception side polarization plate 121 using a two-dimensional image, and an image-processing device 14 for detecting the transparent body 10 by utilizing a shading distribution in an image obtained from the television camera 13. The light is transmitted through the reception side polarization plate 121 for projecting light that is polarized in a specific direction to a specific region including the transparent body 10 and transmitting the light that has been transmitted through the specific region only through specific polarization, and is received by a CCD that is not shown within the television camera 13, thus detecting the transparent body 10 by utilizing the shading distribution in an image obtained from the CCD.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a transparent object detecting method for detecting a transparent object and a system therefor.

[0002]

2. Description of the Related Art Conventionally, as a method of detecting an object, a method of irradiating an object with light from an imaging system side and detecting the object by utilizing a difference in brightness caused by a difference in reflectance between the object and a background. Alternatively, a method of irradiating the object with light from the object side opposite to the imaging system and detecting the object by using a difference in brightness caused by a difference in transmittance between the object and the background is generally used.

[0003]

However, in the above-mentioned conventional method, when the object is a transparent body, the transparent body has a low reflectance and a high transmittance, so that contrast with the background cannot be obtained. It was difficult to detect.

Japanese Unexamined Patent Publication No. Hei 6-342083 discloses a structure of a light projecting unit that emits ultraviolet light when detecting a transparent material other than a quartz substance and an object, and a light receiving unit that receives the emitted ultraviolet light. A method of detecting a transparent material and an object by passing a transparent material and a substance between a light emitting part and a light receiving part and blocking ultraviolet light is described. However, there is disclosed a method utilizing a characteristic difference at a specific wavelength of a transparent material, but this method depends on characteristics of a specific material,
It was not a method that could be used for any transparent body.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a transparent object detection method and a system thereof capable of detecting a transparent object without depending on a specific material. .

[0006]

According to a first aspect of the present invention, there is provided a method for detecting a transparent body, comprising: converting a light polarized in a specific direction into a transparent body having a characteristic of changing a polarization direction; The light transmitted through the predetermined area is transmitted through a polarizing means that transmits only a specific polarized light, and then received by a two-dimensional image sensor. The transparent body is detected using a light-dark distribution in an image to be obtained.

According to a second aspect of the present invention, in the method for detecting a transparent object according to the first aspect, the light projected onto the predetermined area is provided in front of an imaging device including the polarization direction of the light and the two-dimensional imaging device. The direction of polarization of the polarizing means installed in the section is made orthogonal to the two-dimensional image sensor to receive light, and the position and size of the transparent body are utilized by utilizing the brightness distribution in an image obtained from the two-dimensional image sensor. Is detected.

According to a third aspect of the present invention, in the method for detecting a transparent object according to the first aspect, the light projected onto the predetermined area is provided before the image pickup apparatus including the polarization direction of the light and the two-dimensional image pickup device. The direction of polarization of the polarizing means provided in the section is made to match, the two-dimensional image sensor receives light, and the position and size of the transparent body are utilized by utilizing the brightness distribution in an image obtained from the two-dimensional image sensor. Is detected.

According to a fourth aspect of the present invention, there is provided a transparent object detection system, comprising: a light projecting device for polarizing light in a specific direction and projecting the light to a predetermined area including a transparent material having a characteristic of changing a polarization direction; A polarizing means for transmitting only a specific polarized light out of the light transmitted through the predetermined region, an imaging device for capturing an image obtained by receiving the light transmitted through the polarizing means as a two-dimensional image, An image processing device for detecting the transparent body by using a light and dark distribution in an obtained image.

According to a fifth aspect of the present invention, in the transparent object detection system according to the fourth aspect, in the imaging device, the polarization direction of the light projected on the predetermined area is orthogonal to the polarization direction of the polarization unit. In this state, an image obtained by receiving the light transmitted through the polarizing unit is captured as a two-dimensional image, and the image processing apparatus uses the light and dark distribution in the image obtained from the imaging apparatus to perform the transparent processing. It is characterized by detecting the position and size of the body.

According to a sixth aspect of the present invention, in the transparent object detection system according to the fourth aspect, in the imaging device, the polarization direction of the light projected on the predetermined area coincides with the polarization direction of the polarization means. In this state, an image obtained by receiving the light transmitted through the polarizing unit is captured as a two-dimensional image, and the image processing apparatus uses the light and dark distribution in the image obtained from the imaging apparatus to perform the transparent processing. It is characterized by detecting the position and size of the body.

According to a seventh aspect of the present invention, in the transparent object detection system according to any one of the fourth to sixth aspects, the imaging device outputs two-dimensional image signals for forming each of the two-dimensional images. The image processing apparatus has a memory for accumulating each pixel value obtained by binarizing each pixel signal from the two-dimensional image sensor, and an image formed by each pixel value of the memory. The transparent body is detected by calculating the coordinates of the position of the bright portion or the dark portion above.

According to an eighth aspect of the present invention, in the transparent object detection system according to the seventh aspect, the image processing device counts the number of pixels for each of the bright or dark portions, and the count value is set in advance. When the number of pixels exceeds the calculated number of pixels, the position of the center of gravity of the block is obtained as the position of the transparent body, and the second moment of the block is obtained as the attitude of the transparent body.

According to a ninth aspect of the present invention, in the transparent object detection system according to the eighth aspect, the image processing device performs an expansion process and an erosion process on an image obtained from each pixel value stored in the memory. Are sequentially performed to erase lump portions smaller than a predetermined number of pixels, and the number of pixels is counted for each lump of bright or dark portions on the processed image.

The invention according to claim 10 is the invention according to claims 4 to 6.
In the transparent object detection system according to any one of the above, the imaging device has a two-dimensional imaging device that outputs each pixel signal that forms the two-dimensional image, the image processing device,
A memory for accumulating each pixel value obtained by converting each pixel signal from the two-dimensional image sensor into a multi-level image, and storing coordinates of a characteristic position of a brightness distribution on an image formed by each pixel value of the memory; It is characterized in that the transparent body is detected by calculation.

According to an eleventh aspect of the present invention, in the transparent object detection system according to the tenth aspect, the image processing apparatus comprises:
Performing a normalized correlation process between a pre-registered transparent body model and the brightness distribution on the image, and detecting the highest position of the correlation value obtained in the normalized correlation process as the position of the transparent body. Features.

According to a twelfth aspect of the present invention, in the transparent object detection system according to the tenth aspect, the image processing apparatus comprises:
A normalization correlation process is performed between a plurality of models of transparent bodies having different rotation directions registered in advance and the brightness distribution on the image, and the highest position of the correlation value obtained by these normalized correlation processes is determined by the transparent body. And the rotation direction of the model at that time is detected as the posture of the transparent body.

According to a thirteenth aspect of the present invention, in the method for detecting a transparent object according to the first aspect, the polarization direction of the polarization means provided in front of an imaging device including the two-dimensional imaging device is variable. The position of the transparent body is detected based on a difference in the amount of light received by the device.

The invention according to claim 14 is the invention according to claims 7 and 10
Or the transparent object detection system according to 13, wherein the image processing device changes the polarization direction of the polarization unit to various directions when the transparent object does not exist as a subject of the imaging device. Captures various images obtained by imaging, measures the brightness in a preset area in each of these images, and projects the position of the polarizing unit when the brightness is minimized to the predetermined area. Determined as the position where the polarization direction of the light and the polarization direction of the polarizing means are orthogonal to each other, and at the determined position of the polarizing means, when the transparent body is present as a subject of the imaging device, It is characterized in that an image obtained by imaging by an imaging device is taken in, the coordinates of the position of a bright part on this image are calculated, and the transparent body is detected.

The invention according to claim 15 is the invention according to claims 7 and 10.
Or the transparent object detection system according to 13, wherein the image processing device changes the polarization direction of the polarization unit to various directions when the transparent object does not exist as a subject of the imaging device. Captures various images obtained by imaging, measures the brightness in a preset area in each of these images, and projects the position of the polarizing unit when the brightness is maximized to the predetermined area. Determined as a position where the polarization direction of the light and the polarization direction of the polarization means coincide, and at the determined position of the polarization means, when the transparent body is present as a subject of the imaging device, An image obtained by imaging by an imaging device is captured, and coordinates of the position of a dark part on the image are calculated to detect the transparent body.

According to a sixteenth aspect of the present invention, in the transparent object detection system according to the seventh or tenth aspect, the robot hand is moved to a position of the detected transparent object, and the detected transparent object is moved. The apparatus further includes a robot controller that controls the robot to adjust the posture of the robot hand to the posture of the body and to pick up the transparent body.

According to a seventeenth aspect of the present invention, there is provided an image processing apparatus comprising:
3. In the transparent object detection system according to any one of the items 2, the image processing apparatus detects a position and an orientation of the transparent object to be inspected, and an image formed by each pixel value of the memory from the detected position and the orientation. Determine the existence region of the transparent object to be inspected above, compare the shape of this existence region with the shape of a non-defective transparent object, and detect a shape defect such as a chip or a burr from the comparison result. It is characterized by the following.

The invention according to claim 18 is the invention according to claims 7 to 1
3. In the transparent object detection system according to any one of the items 2, the image processing apparatus detects a position and an orientation of the transparent object to be inspected, and an image formed by each pixel value of the memory from the detected position and the orientation. Determine the existence region of the transparent object to be inspected above, and then take the second image by imaging the inspection object without using the polarizing means, and obtain a region corresponding to the existence region of the second image. It is characterized in that it is determined whether or not the magnitude and the degree of change of the brightness change portion of the brightness distribution are within a normal range, and to detect appearance defects such as dirt, scratches and cracks.

According to a nineteenth aspect of the present invention, in the transparent body detection system according to the seventh aspect, the transparent body is a transparent foreign substance that can be mixed into an empty bottle or a bottle containing a transparent liquid, and The processing apparatus measures the area value or the shape and size of each lump of the light portion or the dark portion, and according to the measurement result, the transparent foreign matter is mixed into the empty bottle or the bottle containing the transparent liquid. It is characterized in that it has a function of inspecting a transparent foreign substance in a bottle for determining whether or not it is in the bottle.

[0025]

FIG. 1 is a block diagram of a transparent object detection system, and a first embodiment according to the present invention will be described below with reference to FIG.

The transparent object detection system shown in FIG. 1 is roughly divided into a light projecting device 11, a polarizing device 12, a television camera 13, an image processing device 14, a display 15, a robot controller 16, a robot hand 17.

The light projecting device 11 includes an illuminating device 111 for projecting planar light upward, and a transparent component provided on an upper surface side of the illuminating device 111 and conveyed by, for example, a conveyor (not shown). The transparent body 10 is constituted by a light-emitting side polarizing plate 112 mounted on the upper surface, and the planar light of the illumination device 111 is polarized in a specific direction by the light-emitting side polarizing plate 112 to include the transparent body 10. The light is projected to the region of the light emitting side polarizing plate 112. Here, it is assumed that the transparent body 10 in the example of FIG. 1 is a plate-shaped member made by injection molding a polycarbonate resin. Some transparent materials, such as glass, do not change their polarization characteristics, and some resins may change their polarization characteristics depending on how they are made.However, the above-mentioned resins have properties that change their polarization direction by injection molding. This is what we use.

The polarizing device 12 includes, for example, a circular light-receiving side polarizing plate 121 rotatably provided above the upper surface side of the illumination device 111 so as to face the light-emitting side polarizing plate 112, and the light-receiving side polarizing plate 121. Motor 122 such as a step motor for rotating the light receiving side polarizing plate 121 by meshing with the outer peripheral edge of
And a motor driver 123 for driving the motor 122
And an NC controller 124 that drives a motor 122 via a motor driver 123 in accordance with a control signal from the image processing device 14.
1 so that only a specific polarization of the light transmitted through the area of the light-emitting side polarizing plate 112 can be changed.
1, the light is transmitted rearward (upward in the figure).

The television camera 13 includes a CCD (not shown) as a two-dimensional image pickup device and a light receiving optical system (FIG. 1) for receiving light transmitted through the light receiving side polarizing plate 121 and forming an image on a light receiving surface of the CCD. (Not shown), and captures an image obtained by receiving light transmitted through the light-receiving side polarizing plate 121 as a two-dimensional image. Each pixel signal for forming the two-dimensional image is sent from the CCD side to the image processing device 1.
4 is output.

The image processing unit 14 converts the pixel signals from the television camera 13 into binary or multi-level signals by A / D conversion. The A / D converter 141 converts the pixel signals into digital signals. A memory 142 for storing the converted and binarized or multi-valued data, a display control unit 143 for display control of the display 15, and a communication control for performing communication control of control signals to the NC controller 124 and the robot controller 16. It comprises a unit 144 and a CPU 140.

The CPU 140 performs various image processing and various calculations in accordance with a program stored in a ROM (not shown) or the like. For example, the communication control unit 144
To output a control signal for driving the motor 122 to the NC controller 124.

Further, a process for detecting the position, posture, size, and the like of the transparent body 10 is performed using the brightness distribution of the image obtained from the television camera 13, that is, the image of the data stored in the memory 142.

Further, a process of displaying the obtained result such as the position and orientation of the transparent body 10 on the display 15 through the display control unit 143 is performed. Further, the transparent body 10
Control signals obtained from results such as the position and orientation of
The robot controller 16 via the communication control unit 144
Is transmitted.

The display 15 is a display device such as an LCD or a CRT.
Numeral 6 is for controlling the drive of the robot hand 17 in accordance with a control signal from the image processing device 14. The robot hand 17 is mounted on the upper surface of the light-emitter-side polarizing plate 112 in accordance with the drive control of the robot controller 16. The transparent body 10 is picked up and aligned on the tray 18 and packed.

FIGS. 2 to 4 are explanatory views of the principle of detection of a transparent body when the polarization direction of the light-emitter side polarizing plate is orthogonal to the polarization direction of the light-receiving side polarizer. FIGS. FIG. 4 is an explanatory view of a principle of detecting a transparent body when the polarization direction of the plate and the polarization direction of the light-receiving side polarizing plate coincide with each other. Will be described.

When the polarization direction of the light-emitting-side polarizing plate 112 is orthogonal to the polarization direction of the light-receiving-side polarizing plate 121, as shown in FIG. Since the light cannot pass through the light-receiving side polarizing plate 121, it cannot reach the television camera 13. As a result, the image picked up by the television camera 13 becomes one color in the dark part. However, due to the positional relationship, as shown in FIG.
When the transparent body 10 exists between the light-emitting side polarizing plate 112 and the light-receiving side polarizing plate 121, the light whose polarization direction is restricted by the light-emitting side polarizing plate 112 is transmitted through the transparent body 10 to be polarized. Light of a polarization direction that is disturbed in the direction and can transmit through the light-receiving-side polarizing plate 121 is generated in the region where the transparent body 10 exists. As a result, in the image captured by the television camera 13, as shown in FIG. 4, a portion corresponding to the area where the transparent body 10 exists is a bright portion, and the other portions are dark portions.

On the other hand, when the polarization direction of the light-emitter-side polarizing plate 112 matches the polarization direction of the light-receiving-side polarizer 121, as shown in FIG. The transmitted light passes through the light-receiving-side polarizing plate 121 because both polarization directions are the same, and reaches the television camera 13. As a result, the image picked up by the television camera 13 becomes a single bright portion. However, if the transparent body 10 is present between the light-emitter-side polarizing plate 112 and the light-receiving-side polarizer 121 as shown in FIG.
The light whose polarization direction is limited by 2 is transmitted through the transparent body 10 to be disturbed in the polarization direction, and the light in the polarization direction that cannot be transmitted through the light receiving side polarizing plate 121 is generated in the area where the transparent body 10 exists. As a result, as shown in FIG. 7, in the image captured by the television camera 13, a portion corresponding to the area where the transparent body 10 exists is a dark portion, and the other portions are bright portions.

As described above, by providing the polarizing plates on both the light projecting side and the light receiving side, the portion corresponding to the area where the transparent body 10 exists in the image obtained by imaging becomes a bright or dark portion and becomes a background. Because the contrast with
It is possible to detect the position, posture, size, and the like of the transparent body 10.

FIG. 8 shows a flow executed by the image processing apparatus of FIG. 1 when the polarization direction of the light-emitter side polarizing plate is orthogonal to the polarization direction of the light-receiving side polarizer, and FIG. 9 shows the flow shown in FIG. FIGS. 7A and 7B are diagrams illustrating a state of an image change due to the above-described processing, and an operation of the image processing apparatus 14 when both polarization directions are orthogonal to each other will be described with reference to FIGS.

First, when an image is picked up by the camera 13 in step S101, each pixel signal obtained by the image pickup is converted into an A / D signal by the A / D converter 141 in step S102.
It is converted and digitized, and is binarized in step S103. Thereafter, in step S104, the binarized pixel values are stored in the memory 142 as image data.

Here, in the case of an image formed by the binarized pixel values, since the polarization disorder in the transparent body 10 is not uniform, as shown in FIG. In the portion corresponding to the existence region of (1), a dark portion (“a portion where polarization is hardly changed” in the drawing) is likely to occur. Therefore, step S
After 104, the expansion process and the reduction process are sequentially executed by the CPU 140 in steps S105 and S106 in order to fill in the darkened portion. That is, when the dilation processing is executed, as shown in FIG. 9B, the darkened part is crushed to become a bright part, and when the reduction processing is subsequently executed, as shown in FIG. 9C. In the meantime, the bulged outer shape returns to its original state, while the portion painted in the bright portion remains the bright portion, thereby ending the filling process.

Subsequently, in step S107, labeling is performed on the bright portion lump to identify each bright portion, and in step S108, the center of gravity and the second moment are calculated for the bright portion lump having a predetermined area or more. The CPU 140 determines the center of gravity of the obtained bright block as the position of the transparent body 10 and determines the direction of the second moment as the attitude of the transparent body 10.
Is executed by

Thereafter, in step S109, the transparent body 10
The CPU 140 executes a process of transmitting a control signal obtained from the position and orientation data from the communication control unit 144 to the robot controller 16, whereby the drive control of the robot controller 16 controls the light-emitter-side polarizing plate 11.
The transparent body 10 placed on the upper surface of the device 2 and located at the above position is picked up by the robot hand 17 in accordance with the above posture, and aligned and packed in the tray 18. This clogging is repeated until all of the detected transparent bodies 10 are completed (S110).

FIG. 10 is a flowchart executed by the image processing apparatus of FIG. 1 when the polarization direction of the light-emitter side polarizing plate and the light-receiving side polarizer coincide with each other. The operation of the image processing device 14 when the polarization directions match will be described.

First, when an image is captured by the camera 13 in step S201, in step S202, each pixel signal obtained by the imaging is subjected to A / D conversion by the A / D converter 141.
It is converted and digitized, and is binarized in step S203. Thereafter, in step S204, the binarized pixel values are stored in the memory 142 as image data.

In this case, since the portion corresponding to the area where the transparent body 10 exists is a dark portion, steps S205 and S206 are performed.
Then, the reduction process and the expansion process are sequentially executed by the CPU 140, and the filling process is performed.

Subsequently, in step S207, labeling is performed on the dark block to identify each dark portion. In step S208, the center of gravity and the second moment are calculated and obtained for the dark block having a predetermined area or more. The CPU 140 determines the center of gravity of the dark block as the position of the transparent body 10 and determines the direction of the second moment as the attitude of the transparent body 10.
Is executed by

Thereafter, in step S209, the transparent body 10
The CPU 140 executes a process of transmitting a control signal obtained from the position and orientation data of the light-transmitting side from the communication control unit 144 to the robot controller 16.
The transparent body 10 placed on the upper surface of the transparent body 10 and located at the above position is picked up by the robot hand 17 in accordance with the above posture, and aligned and packed in the tray 18. This clogging is repeated until all of the detected transparent bodies 10 are completed (S210).

The processing shown in FIGS. 8 and 10 uses the image formed by the binarized pixel values for each transparent body 10.
The following describes a process for detecting each transparent body 10 using an image formed by multi-valued pixel values.

FIG. 11 shows an image formed by multi-valued pixel values by the image processing apparatus shown in FIG. 1 when the polarization direction of the light-emitting side polarizing plate is orthogonal to the polarization direction of the light-receiving side polarizing plate. FIG. 12 is a diagram showing an example of each transparent body in an image that may occur when an image formed by binarized pixel values is used, and FIG. 13 is a flowchart of FIG. FIG. 4 is a diagram illustrating an example of a template used.

In the process of measuring the center of gravity and the second moment of a light or dark block using an image formed by binarized pixel values, as shown in FIG. When a transparent body 10a having a shape different from that of the body 10 but having the same area is mixed, there is a problem that it cannot be distinguished.

Therefore, if an image formed by multi-valued pixel values is used to determine the position of each transparent body using normalized correlation, the above problem can be solved. In this case, it is assumed that the brightness of each pixel signal is multi-valued to 256 gradations. Then, in order to be able to cope with a rotated transparent body, as shown in FIG. 13, 180 templates rotated every 1 ° are used. The reason for the 180 rotations is that in the case of the rectangular transparent body shown in FIG. 12, since the shape is symmetric, the rotation can be sufficiently distinguished from 0 ° to 180 °. Note that the process of the normalized correlation may be a process generally performed in the related art, and a description thereof will be omitted.

In FIG. 11, when an image is captured by the camera 13 in step S301, in step S302,
Each pixel signal obtained by the imaging is subjected to A / D conversion by the A / D conversion unit 141, digitized, and multi-valued into 256 gradations. Thereafter, in step S303, the multi-valued pixel values are stored in the memory 142 as image data.

Thereafter, in step S304, a normalized correlation value is calculated by scanning the image for each template of each angle shown in FIG. 13, and in step S305, the normalized correlation value having a value equal to or more than a predetermined value is calculated. The CPU 140 executes a process of setting the position of the peak value of the correlation as the position of the transparent body and setting the rotation angle of the template used for the correlation as the posture of the transparent body. Here, when peaks of correlation values due to templates having different rotation angles occur close to each other, a result having a high correlation value is selected.

Thereafter, in step S306, the CPU 140 executes a process of transmitting a control signal obtained from the data on the position and orientation of the transparent body from the communication control unit 144 to the robot controller 16, thereby driving the robot controller 16. Under the control, the transparent body placed on the upper surface of the projection-side polarizing plate 112 and located at the above position is picked up by the robot hand 17 in accordance with the above posture,
The tray 18 is arranged and packed. This clogging is repeated until the operation is completed for all of the detected transparent bodies (S307).

The adjustment of the polarization direction of the light-emitting-side polarizing plate 112 and the polarization direction of the light-receiving-side polarizing plate 121 can be performed visually while viewing the image captured by the television camera 13 on the screen of the display 15. However, the first embodiment is configured to automatically adjust both polarization directions to enable more accurate adjustment. That is, when automatically adjusting both polarization directions to be orthogonal to each other, first, rotate the light-receiving-side polarizing plate 121 in a state where nothing is placed on the upper surface of the fixed light-emitting-side polarizing plate 112, For example, a process of capturing a plurality of images by the television camera 13 at each predetermined rotation pitch is performed. At this time,
A control signal is output from the image processing device 14 to the NC controller 124, and the motor 122 is driven by the drive control of the motor driver 123 to rotate the light-receiving-side polarizing plate 121.

Subsequently, the brightness in a predetermined area in each image is measured, and the rotation angle of the light-receiving-side polarizing plate 121 at which the brightness is minimized is determined as a position where the two polarization directions are orthogonal to each other. , A process of outputting a control signal to the NC controller 124 so that the rotation angle is obtained. As a result, the motor 122 is driven by the drive control of the motor driver 123, the light-receiving-side polarizing plate 121 is rotated and stopped at the rotation angle, and the two polarization directions are orthogonal to each other.

On the other hand, when the automatic adjustment is performed so that the two polarization directions coincide with each other, first, in a state where nothing is placed on the upper surface of the light emitting side polarizing plate 112, the light receiving side polarizing plate 12
1 is rotated, and a process of capturing a plurality of images by the television camera 13 at a predetermined rotation pitch, for example, is performed.

Subsequently, the brightness in a preset area in each image is measured, and the rotation angle of the light-receiving-side polarizing plate 121 at which the brightness becomes maximum is determined as a position where the two polarization directions coincide. , A process of outputting a control signal to the NC controller 124 so that the rotation angle is obtained. As a result, the motor 122 is driven by the drive control of the motor driver 123, the light-receiving-side polarizing plate 121 is rotated and stopped at the rotation angle, and the two polarization directions match.

FIG. 14 is a configuration diagram of a transparent object detection system, and a second embodiment according to the present invention will be described below with reference to FIG.

The transparent object detection system shown in FIG. 14 comprises a light projecting device 11, a television camera 13, and a display 15
Robot controller 16 and robot hand 17
Are provided in the same manner as in the first embodiment. The difference from the first embodiment is that the polarizing device 22 and the image processing device 2
4 and automatically inspects the appearance of the transparent body 10,
According to the inspection result, an appearance inspection sorting system is configured in which the transparent bodies 10 are sorted into non-defective products or defective products, and are arranged and packed in a non-defective tray 181 or a defective tray 182.

The polarizing device 22 is provided, for example, in a circular shape on the upper surface side of the lighting device 111 so as to be slidable so as to face the light projecting side polarizing plate 112.
And a slide plate 220 that holds the light-receiving-side polarizing plate 221 in one surface and has a window (for example, a hole) 222 having substantially the same shape as the light-receiving-side polarizing plate 221 in the other surface. A slide actuator 223 that slides in a horizontal plane, and a slide that drives and controls the slide actuator 223 so that the central axis of the light-receiving side polarizing plate 221 or the window 222 is aligned with the optical axis of the television camera 13 according to a control signal from the image processing device 24. When the center axis of the light receiving side polarizing plate 221 is aligned with the optical axis of the television camera 13 (hereinafter, this state is referred to as “the light receiving side polarizing plate is present”), the light emitting side polarizing plate 112 is formed. In this case, only specific polarized light of the light transmitted through the region is transmitted backward by the light-receiving-side polarizing plate 221. Note that the center axis of the window 222 is
The state in which the optical axis is aligned with the optical axis is hereinafter referred to as “without the light-receiving-side polarizing plate”.

The image processing device 24 includes an A / D converter 141
, A memory 142, a display control unit 143, and a communication control unit 144 in the same manner as the image processing apparatus 14 of the first embodiment.
U240 is provided. However, the robot controller 16 and the slide actuator controller 224 are connected to the communication control unit 144.

The CPU 240 is stored in a ROM (not shown) and performs various image processing and various calculations in accordance with a program different from that of the first embodiment.
The communication control unit 144 performs a process of outputting a control signal for setting the presence or absence of the light receiving side polarizing plate to the slide actuator controller 224.

The position, orientation, and size of the transparent body 10 are determined using the brightness distribution of an image obtained from the television camera 13, that is, an image based on data stored in the memory 142, with the light-receiving-side polarizing plate. A process for detecting the like is performed. In addition, an inspection process of the appearance of the transparent body 10 is performed.

The image and the detection result such as the position and orientation of the transparent body 10 and the inspection result of the appearance are transmitted to the display control unit 14.
The process of displaying the information on the display 15 through the step 3 is performed. Further, a process of transmitting a control signal obtained from a detection result such as a position and a posture of the transparent body 10 and an inspection result of the appearance to the robot controller 16 via the communication control unit 144 is performed.

FIG. 15 to FIG. 19 are explanatory diagrams of the appearance inspection of the transparent body, and the procedure of various inspections of the appearance will be described together with the principle using these figures. Dirt on the transparent body 10,
When inspecting for the presence or absence of a first defect such as a flaw or a crack, first, as shown in FIG. 15, the position of the transparent body 10 and the normalized The posture is determined, and the area where the transparent body 10 exists is specified. Here, when there is no defect in the transparent body 10, the presence area on the image obtained without the light-receiving-side polarizing plate becomes an image with low contrast as shown in FIG.
As shown in (1), if the transparent body 10 has a first defect (dirt in the figure), the portion corresponding to the defect becomes darker than the surroundings, and the inside of the existence region becomes an image having a partially high contrast. For this reason, after specifying the existence region, in the existence region on the image obtained without the light-receiving side polarizing plate,
The number of pixels below a predetermined brightness is measured, and the presence or absence of the first defect is determined according to the measurement result. For example,
If the number of pixels in the measurement result is equal to or greater than a predetermined inspection threshold, it is determined that there is a first defect, and if not, it is determined that there is no first defect.

It should be noted that the brightness distribution in the existing area of the grayscale image obtained without the light-receiving side polarizing plate is differentiated to emphasize the changed portion, and the size of the portion above a predetermined threshold value is also determined. This makes it possible to detect external defects such as finer dirt, scratches and cracks. Here, the image without the light-receiving side polarizing plate may be imaged as the second image using another imaging device arranged in parallel. In this case, the image with the light-receiving side polarizing plate (the second image) may be used. The first image) is associated with the positional relationship of the second image, and the second image is examined for the lightness distribution of an area corresponding to the existence area of the inspection target obtained in the first image, and the lightness change is performed. It may be determined whether or not the size or the degree of change of the portion is within a normal range, and a defect in appearance such as dirt, a scratch, or a crack may be detected.

Next, when inspecting the presence or absence of a second defect such as chipping or burrs, first, as described above, the area where the transparent body 10 exists is specified. Here, if the transparent body 10 has a chipped defect, as shown in FIG. 18, the bright portion in the (binary) image obtained with the light-receiving side polarizing plate is the transparent body 1 in the case of a non-defective product.
0 (specifically, the area determined by the template model shown in FIG. 13).
When the transparent body 10 has a flash defect, as shown in FIG. 19, a bright portion in an image obtained with the light-receiving side polarizing plate is:
In the case of a non-defective product, the region where the transparent body 10 exists is protruded.

For this reason, the number of pixels in the area where the transparent body 10 exists in the case of a non-defective product is known, and the upper and lower limits of the number of pixels are used as criteria for judging the presence or absence of the second defect from the number of pixels. Set the normal range specified in
After specifying the existence area of 0, the number of pixels in the bright part of the transparent body 10 is counted, and if this count value is within the normal range,
If it is determined that there is no second defect and the count value is larger than the upper limit of the normal range, it is determined that there is a flash defect, and if the count value is smaller than the lower limit of the normal range, it is determined that there is a chip defect. Note that, in the case of a non-defective product, the existing region of the transparent body 10 or the normal range including the variation of the non-defective product is known. In the following, if the number of pixels of the protruding part is also equal to or less than the specified value, it is determined that there is no second defect. If the number of pixels of the small part exceeds the specified value, it is determined that there is a chipped defect, and the number of pixels of the protruding part. May exceed the specified value, it may be determined that there is a flash defect.

FIGS. 20 and 21 show the case where the polarization direction of the light-emitter side polarizing plate and the light-receiving side polarizer are orthogonal to each other.
4 is a flow executed by the image processing device of No. 4, and the operation of the image processing device 24 when both polarization directions are orthogonal to each other will be described below. When the polarization directions of the two light beams coincide with each other, the light portion and the dark portion are interchanged, but it goes without saying that the operation is substantially the same.

First, in step S401 in FIG. 20, a control signal indicating the presence of the light-receiving-side polarizing plate is sent from the image processing apparatus to the polarizing device 2.
2 and the slide plate 220 slides in a state where the light-receiving-side polarizing plate is present. Thereafter, when an image is captured by the camera 13 in step S402, the process proceeds to step S403.
Then, each pixel signal obtained by the imaging is converted into an A / D converter
A / D converted and digitized by 1 followed by 25
It is multi-valued into six gradations. Thereafter, in step S404,
The multi-valued pixel values are stored in the memory 142 as image (polarized image) data.

Next, in step S405, the CPU 240 executes a process of scanning the image for each template of each angle shown in FIG. 13 and calculating a normalized correlation value. Subsequently, in step S406, the CPU 240 performs processing of setting the position of the peak value of the correlation as the position of the transparent body and setting the rotation angle of the template used for the correlation as the posture of the transparent body although the correlation value has a value equal to or greater than the predetermined value. Be executed.
At this time, if the peaks of the correlation values due to the templates having different rotation angles occur close to each other, the result with the higher correlation value is selected.

Next, in step S407, a control signal indicating that there is no light-receiving-side polarizing plate is output from the image processing apparatus to the polarization device 22, and the slide plate 220 slides without the light-receiving-side polarizing plate. Thereafter, in step S408, the camera 1
When the imaging is performed in step S3, in step S409, each pixel signal obtained by the imaging is A / D converted by the A / D conversion unit 141, digitized, and multi-valued into 256 gradations. Thereafter, in step S410, the multi-valued pixel values are stored in the memory 142 as image (shade image) data.

Next, in accordance with the position and orientation of the transparent body obtained in step S406, the shape of the model of the corresponding transparent body shown in FIG. The region is set as an inspection area for inspecting the presence or absence of the first defect (S411), and the process proceeds to step S411.
At 412, the CPU 240 executes a process of counting the number of pixels having a predetermined brightness or less in the inspection area.

Then, in step S413 in FIG. 21, it is determined whether or not the counted number of pixels is equal to or greater than the inspection threshold. If the number of pixels is equal to or greater than the inspection threshold (S41).
3 is YES), in step S423, it is determined that there is any defect such as dirt, scratch, or crack.
The process that proceeds to 422 is executed by the CPU 240.

On the other hand, if the number of pixels is smaller than the inspection threshold value (NO in S 413), in step S 414, the polarization image is binarized to generate a polarization binary image, and this is stored in the memory 142. Is executed by the CPU 240. After that, the CPU 240 sequentially executes the expansion processing and the reduction processing in steps S415 and S416 in order to perform the filling processing on the polarization binary image.

Next, in step S417, the shape of the corresponding transparent part model is superimposed on the filled-in polarization binary image in accordance with the position and orientation of the transparent body obtained in step S406. The shape region of the model is defined as a shape abnormality inspection area. Here, the shape abnormality inspection area is an area provided for removing, for example, noise and the like, and is set to be equal to or larger than the size of the transparent object to be inspected including the maximum value of burrs.

In step S 418, the CPU 240 executes a process of counting the number of pixels of a bright portion on the polarization binary image overlapping the shape abnormality inspection area. Then, in step S419, it is determined whether or not the counted number of pixels is larger than a preset upper limit value of the normal range. If the counted pixel number is larger than the upper limit value (YES in S419), step S42 is performed.
In step S422, it is determined that there is a flash defect.
On the other hand, if the counted number of pixels is equal to or smaller than the upper limit value of the normal range (NO in S419), in step S420,
It is determined whether or not the counted number of pixels is smaller than the lower limit of the normal range, and if it is smaller than the lower limit (YE in S420).
S), it is determined that there is a chipped defect in step S425, and thereafter, the process proceeds to step S422. If it is equal to or more than the lower limit (NO in S420), the transparent body is determined to be non-defective in step S421. The process that proceeds to step S422 is executed by CPU 240.

Next, in step S422, the CPU 240 performs a process of transmitting a control signal obtained from the detection result of the position and orientation of the transparent body and the inspection result of the appearance to the robot controller 16 from the image processing apparatus. Thereby, the transparent body is sorted into non-defective or defective products, and the non-defective tray 181 or the defective tray 1 is used.
82 and packed. Such a series of processes is repeatedly executed each time the transparent body is sent from all the processes.

FIG. 22 is a block diagram of the transparent object detection system, and FIG. 23 is a view showing an example of a transparent foreign substance that can remain inside the bottle of FIG. 22. The third embodiment according to the present invention will be described below using these figures. The form will be described.

The transparent object detection system shown in FIG. 22 has a television camera 13 and a display 15 in the same manner as in the first embodiment. A light transmitting side polarizing plate 321 provided at the front of the television camera 13, a rotating device 38, a sensor 39 for detecting the timing of imaging, and an image processing device 34 are provided. It has a transparent foreign matter inspection function in the bottle for inspecting the presence or absence of a transparent foreign matter inside the bottle (a bottle containing a transparent liquid) 20.

However, in FIG. 22, the bottle 20 is
A mechanism for transporting the bottle 20 containing transparent foreign matter therein to a defective buffer is omitted from the drawings.

The light projecting device 21 includes a lighting device 211 for projecting planar light from the side of the bottle 20 placed vertically and a light projecting side polarizing plate 212 provided on the light projecting surface of the lighting device 211. The planar light of the illumination device 211 is polarized in a specific direction by the light-emitting side polarizing plate 212, and
The light is projected in parallel with 2 onto an area including the bottle 20. Here, the light that has exited the illumination device 211 is transmitted to the projection-side polarizing plate 212.
The light passes through the bottle 20, and a part of the light passes through the bottle 20, passes through the light receiving side polarizing plate 321, and forms an image on the light receiving surface of the CCDD of the television camera 13.

The rotary device 38 includes a circular rotary table 381 on which the bottle 20 is placed vertically, a motor 382 such as a step motor for rotating the rotary table 381 by meshing with the outer peripheral edge of the rotary table 381. A motor driver 383 for driving the motor 382 and an NC controller 384 that drives the motor 382 via the motor driver 383 in accordance with a control signal from the image processing device 34. This is for rotating the bottle 20 around an axis for imaging.

The image processing device 34 includes an A / D converter 141
, A memory 142, a display control unit 143, and a communication control unit 144 in the same manner as the image processing apparatus 14 of the first embodiment. Digital I / O3 for importing
45 and a CPU 340. However, an NC controller 384 is connected to the communication control unit 144.

The CPU 340 is stored in a ROM (not shown), and performs various image processing and various calculations in accordance with a program different from that of the first embodiment.
The communication control unit 144 performs a process of outputting a control signal for driving the motor 382 to the NC controller 384.

Further, a process of inspecting the inside of the bottle 20 for the presence of a transparent foreign substance is performed. Here, the principle of the inspection will be described. If the polarization direction of the light-emitter-side polarizing plate 212 and the polarization direction of the light-receiving-side polarizer 321 are set to coincide with each other, and the illumination device 211 illuminates the bottle 20, If a transparent foreign substance exists inside 20, the image obtained by the imaging by the television camera 13 will be a dark lump of the transparent foreign substance, as shown in the example of FIG. Therefore, it is possible to determine the presence or absence of a transparent foreign object according to the presence or absence of a portion that appears as a dark lump in the image. Further, it is possible to detect dirt and other opaque foreign substances in addition to the transparent foreign substances.

On the other hand, when the bottle 20 is illuminated by the illumination device 211 by setting the polarization direction of the light-emitting side polarizing plate 212 and the polarization direction of the light-receiving side polarizing plate 321 to be orthogonal, the transparent foreign matter is bright. Since the image is displayed as a lump, it is possible to determine the presence or absence of a transparent foreign object in accordance with the presence or absence of a portion that appears as a bright lump in the image. However, in this case, in the case of dirt or other opaque foreign substances, they cannot be detected because they do not appear as bright lump.

FIG. 24 is a flowchart executed by the image processing apparatus of FIG. 22 when the polarization direction of the light-emitter side polarizing plate and the light-receiving side polarizer coincide with each other. Next, the operation of the third embodiment will be described.

When an image is captured by the camera 13 in step S501, each pixel signal obtained by the imaging is A / D converted by the A / D converter 141 and digitized in step S502. Valued. Thereafter, in step S504, the binarized pixel values are stored in the memory 142 as image (polarized binary image) data.

Next, in steps S505 and S506,
Reduction processing and expansion processing are sequentially executed by the CPU 340. Here, in the portion corresponding to the region where the transparent foreign matter is present in the polarization binary image, the portion where the polarization is disturbed by the transparent foreign matter is not uniform, so that a brightly missing portion tends to occur. And expansion processing is performed.

Next, in step S507, as shown in FIG. 23, the CPU 340 performs a process of performing labeling on a block in a dark area in a preset inspection area to identify each block and calculating the number of pixels of each block. Is executed by

Subsequently, in step S508, it is determined whether or not the number of pixels of the block of the dark portion is equal to or greater than a preset inspection threshold value. If the number of pixels is equal to or greater than the inspection threshold value (S50)
8 (YES in step S5), while it is determined in step S512 that there is a transparent foreign substance, the number of pixels is not equal to or greater than the inspection threshold (S5).
In step S509, the CPU 340 executes a process of outputting a control signal for rotating the bottle 20 to a predetermined angle to the NC controller 384.

After step S509, step S510
It is determined whether or not the bottle 20 has rotated once, and if it has not rotated once (NO in S510), the process proceeds to step S501.
On the other hand, if it has made one rotation (YES in S510),
The process of determining the bottle 20 as non-defective is executed by the CPU 340.

After step S512, step S513
The CPU 340 executes a process of discharging the bottle 20 determined to have the transparent foreign matter to the defective buffer.

In the above-mentioned flow, the processing procedure is such that the bottle 20 is rotated at a fixed angle and an image is taken.
In the transparent object detection system shown in FIG. 2, while rotating the bottle 20 continuously, while watching the output of the sensor 39, a plurality of images are captured at once by the television camera 13, and the plurality of images are sequentially subjected to the above processing. And another mode for inspecting the inside of the bottle 20 for the presence of a transparent foreign substance.

By the way, in the case of a bottle to be recycled, transparent foreign substances such as cigarette wrapping and Saran wrap that have fallen into the bottle when used may remain in the bottle even after the bottle is washed. In conventional visual inspection, almost no transparent foreign matter left in the bottle could be found. On the other hand, according to the third embodiment, it is possible to automatically find out the transparent foreign matter remaining in such a bottle.

[0099]

As is apparent from the above, according to the first aspect of the present invention, light polarized in a specific direction
The light is projected onto a predetermined area including a transparent body having a characteristic of changing the polarization direction, and the light transmitted through the predetermined area is transmitted through a polarization unit that transmits only specific polarized light.
Since the transparent body is detected by using a two-dimensional image sensor to receive light and utilizing the light and dark distribution in an image obtained from the two-dimensional image sensor, the contrast of a portion where the transparent body exists can be increased. Can be detected.

According to the second aspect of the present invention, in the transparent object detecting method according to the first aspect, the light projected onto the predetermined area is provided with an image pickup apparatus including the polarization direction of the light and the two-dimensional image pickup device. The direction of polarization of the polarizing means provided at the front of the light source is made orthogonal to the two-dimensional imaging device to receive light.
Using the light and dark distribution in the image obtained from the three-dimensional image sensor, to detect the position and size of the transparent body,
Since the contrast of the portion where the transparent body exists can be increased, the transparent body can be detected.

According to a third aspect of the present invention, in the transparent object detection method according to the first aspect, the light projected onto the predetermined area is converted into an image pickup device including the polarization direction of the light and the two-dimensional image pickup device. The two-dimensional imaging device receives light by matching the polarization direction of the polarizing means installed in front of the two-dimensional imaging device.
Using the light and dark distribution in the image obtained from the three-dimensional image sensor, to detect the position and size of the transparent body,
Since the contrast of the portion where the transparent body exists can be increased, the transparent body can be detected.

According to the fourth aspect of the present invention, there is provided a light projecting device for polarizing light in a specific direction and projecting the light to a predetermined area including a transparent body having a characteristic of changing a polarization direction, and Polarizing means for transmitting only specific polarized light among transmitted light, an imaging device for receiving an image obtained by receiving light transmitted through the polarizing means as a two-dimensional image, and an image obtained from the imaging device Since the image processing apparatus is provided with the image processing device for detecting the transparent body by using the light and dark distribution in the above, the contrast of the portion where the transparent body exists can be increased, so that the transparent body can be detected.

According to the fifth aspect of the present invention, in the transparent object detection system according to the fourth aspect, the imaging device may be configured such that a polarization direction of the light projected on the predetermined area and a polarization direction of the polarization unit are different. In an orthogonal state, an image obtained by receiving the light transmitted through the polarizing unit is captured as a two-dimensional image, and the image processing device uses a light and dark distribution in an image obtained from the imaging device, Since the position and the size of the transparent body are detected, the contrast of the portion where the transparent body exists can be increased, so that the transparent body can be detected.

According to a sixth aspect of the present invention, in the transparent object detection system according to the fourth aspect, the imaging device may be configured so that a polarization direction of the light projected on the predetermined area and a polarization direction of the polarization unit are different. In the coincident state, an image obtained by receiving the light transmitted through the polarizing unit is captured as a two-dimensional image, and the image processing device uses a light-dark distribution in an image obtained from the imaging device, Since the position and the size of the transparent body are detected, the contrast of the portion where the transparent body exists can be increased, so that the transparent body can be detected.

According to the seventh aspect of the present invention, in the transparent object detection system according to any one of the fourth to sixth aspects,
The imaging device has a two-dimensional imaging device that outputs each pixel signal that forms the two-dimensional image, and the image processing device is obtained by binarizing each pixel signal from the two-dimensional imaging device. It has a memory for accumulating each pixel value, and calculates the coordinates of the position of the bright part or dark part on the image formed by each pixel value of this memory to detect the transparent body,
The position and orientation of the transparent body can be accurately obtained by simple processing.

According to an eighth aspect of the present invention, in the transparent object detection system according to the seventh aspect, the image processing apparatus counts the number of pixels for each of the light or dark portions, and the count value is When the number of pixels exceeds a preset number, the position of the center of gravity of the block is determined as the position of the transparent body, and the second moment of the block is determined as the posture of the transparent body. The posture can be obtained with high accuracy.

According to the ninth aspect of the present invention, in the transparent object detection system according to the eighth aspect, the image processing device expands and contracts an image obtained from each pixel value stored in the memory. The processing is sequentially performed to erase a lump portion smaller than a predetermined number of pixels, and the number of pixels is counted for each bright or dark lump on the processed image. And the attitude can be obtained with high accuracy.

According to the tenth aspect, the fourth aspect is provided.
7. In the transparent object detection system according to any one of to 6, the imaging device includes a two-dimensional imaging element that outputs each pixel signal that forms the two-dimensional image, and the image processing device includes the two-dimensional imaging device. It has a memory for storing each pixel value obtained by multi-leveling each pixel signal from the image sensor, and calculating coordinates of a characteristic position of a brightness distribution on an image formed by each pixel value of the memory. Since the transparent body is detected, the position and orientation of the transparent body can be obtained with high accuracy.

According to the eleventh aspect, the first aspect is provided.
0, the image processing apparatus performs a normalized correlation process between a previously registered transparent body model and a brightness distribution on the image, and calculates a correlation value obtained by the normalized correlation process. Since the highest position is detected as the position of the transparent body, the position and orientation of the transparent body can be obtained with high accuracy.

According to the twelfth aspect, according to the first aspect,
0, the image processing apparatus performs a normalized correlation process between a pre-registered model of a plurality of transparent bodies having different rotation directions and a lightness distribution on the image, and the normalized correlation The highest position of the correlation value obtained in the processing is detected as the position of the transparent body, and the rotation direction of the model at that time is detected as the posture of the transparent body. it can.

According to the thirteenth aspect, according to the first aspect,
In the transparent object detection method according to the above, the polarization direction of the polarizing unit installed in front of the imaging device including the two-dimensional imaging device is variable, and the position of the transparent body is determined by a difference in the amount of light received by the imaging device. Since the detection is performed, the polarization direction necessary for suitably detecting the transparent body can be automatically set to an optimum state.

According to a fourteenth aspect of the present invention, in the transparent object detection system according to the seventh, tenth, or thirteenth aspect, the image processing apparatus is configured to execute the operation when the transparent object does not exist as a subject of the imaging device. In addition, by changing the polarization direction of the polarization unit to various directions, various images obtained by imaging with the imaging device are captured, and the brightness in a preset region in each of these images is measured, and the brightness is measured. The position of the polarizing means at the time of the minimum, determined as a position where the polarization direction of the light projected to the predetermined area and the polarization direction of the polarizing means are orthogonal, at the determined position of the polarizing means, When the transparent body is present as a subject of the imaging apparatus, an image obtained by imaging by the imaging apparatus is taken in, the coordinates of the position of a bright part on this image are calculated, and the transparent body is detected. Runode can automatically set the polarization direction necessary for suitably detecting a transparent body in optimum condition.

According to a fifteenth aspect of the present invention, in the transparent object detection system according to the seventh, tenth, or thirteenth aspect, the image processing device is configured to execute the operation when the transparent object does not exist as a subject of the imaging device. In addition, by changing the polarization direction of the polarization unit to various directions, various images obtained by imaging with the imaging device are captured, and the brightness in a preset region in each of these images is measured, and the brightness is measured. The position of the polarizing means at the time of maximum is determined as a position where the polarization direction of the light projected on the predetermined area coincides with the polarization direction of the polarizing means, and at the determined position of the polarizing means, When the transparent body is present as a subject of the imaging apparatus, an image obtained by imaging by the imaging apparatus is taken in, the coordinates of the position of a dark part on this image are calculated, and the transparent body is detected. Runode can automatically set the polarization direction necessary for suitably detecting a transparent body in optimum condition.

According to the sixteenth aspect, the seventh aspect is provided.
Or in the transparent object detection system according to 10, the robot hand, the robot hand is moved to the position of the detected transparent body, the posture of the robot hand is adjusted to the detected posture of the transparent body, the Since the apparatus further includes a robot controller that performs control for picking up the transparent body, the position and orientation of the transparent body can be measured, and the transparent body can be handled by a robot hand.

According to the seventeenth aspect, the seventh aspect is provided.
In the transparent object detection system according to any one of to 12, the image processing apparatus detects a position and an orientation of the transparent object to be inspected, and from the detected position and orientation,
Determine the existence area of the transparent object to be inspected on the image formed by each pixel value of the memory, perform a comparison between the shape of the existence area and the shape of the non-defective transparent body, and from the comparison result, Since a shape defect such as a chip or a burr is detected, it is possible to automatically detect a shape defect such as a chip or a burr that may occur in the transparent body.

According to the eighteenth aspect, the seventh aspect is provided.
Or the transparent object detection system according to 10, wherein the image processing apparatus detects a position and an orientation of the transparent object to be inspected, and, based on the detected position and orientation, on an image formed by each pixel value of the memory. The existence region of the transparent object to be inspected is determined, and then the inspection object is imaged without using the polarizing means, and a second image is captured.
It is determined whether the size or the degree of change in the brightness change portion of the brightness distribution of the area corresponding to the existence area of the image is within a normal range, and dirt, scratches, and external defects such as cracks are detected. Therefore, it is possible to automatically detect appearance defects such as dirt, scratches, and cracks that may occur on the transparent body.

According to the nineteenth aspect, the seventh aspect is provided.
In the transparent object detection system according to the above, the transparent object,
A transparent foreign substance that can be mixed into an empty bottle or a bottle containing a transparent liquid, and the image processing apparatus measures an area value or a shape and a dimension of each of the light or dark portions, and obtains a measurement result. Accordingly, the empty bottle or the bottle containing the transparent liquid has a transparent foreign matter inspection function for determining whether or not the transparent foreign matter is mixed in the empty bottle or the bottle containing the transparent liquid. It becomes possible to detect the transparent foreign matter remaining inside.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a transparent object detection system.

FIG. 2 is an explanatory diagram of a principle of detecting a transparent body when a polarization direction of a light-emitting side polarizing plate and a polarization direction of a light-receiving side polarizing plate are orthogonal to each other.

FIG. 3 is an explanatory diagram of a principle of detecting a transparent body when the polarization direction of a light-emitting-side polarizing plate is orthogonal to the polarization direction of a light-receiving-side polarizing plate.

FIG. 4 is an explanatory diagram of a principle of detecting a transparent body when the polarization direction of a light-emitting side polarizing plate is orthogonal to the polarization direction of a light-receiving side polarizing plate.

FIG. 5 is an explanatory diagram of a principle of detecting a transparent body when the polarization direction of the light-emitting side polarizing plate and the polarization direction of the light-receiving side polarizing plate match.

FIG. 6 is an explanatory diagram of the principle of detecting a transparent body when the polarization direction of the light-emitter side polarizing plate and the light-receiving side polarizer coincide with each other.

FIG. 7 is an explanatory diagram of the principle of detecting a transparent body when the polarization direction of the light-emitter side polarizing plate and the light-receiving side polarizer coincide with each other.

FIG. 8 is a flowchart executed by the image processing apparatus of FIG. 1 when the polarization direction of the light-emitting-side polarizing plate is orthogonal to the polarization direction of the light-receiving-side polarizing plate.

FIG. 9 is a diagram illustrating a state of an image change due to processing executed in the flow of FIG. 8;

FIG. 10 is a flowchart executed by the image processing apparatus of FIG. 1 when the polarization direction of the light-emitting-side polarizing plate matches the polarization direction of the light-receiving-side polarizing plate.

11 is a diagram showing an example in which when the polarization direction of the light-emitter side polarizing plate and the light-receiving side polarizer are orthogonal to each other, the image processing apparatus of FIG. FIG.

FIG. 12 is a diagram illustrating an example of each transparent body in an image that may occur when an image formed by binarized pixel values is used.

FIG. 13 is a diagram illustrating an example of a template used in the flow of FIG. 11 and the like;

FIG. 14 is a configuration diagram of a transparent object detection system.

FIG. 15 is an explanatory diagram of an appearance inspection of a transparent body.

FIG. 16 is an explanatory diagram of an appearance inspection of a transparent body.

FIG. 17 is an explanatory diagram of an appearance inspection of a transparent body.

FIG. 18 is an explanatory diagram of an appearance inspection of a transparent body.

FIG. 19 is an explanatory diagram of a visual inspection of a transparent body.

20 is a flowchart executed by the image processing apparatus of FIG. 14 when the polarization direction of the light-emitter-side polarizing plate is orthogonal to the polarization direction of the light-receiving-side polarizer.

21 is a flowchart executed by the image processing apparatus of FIG. 14 when the polarization direction of the light-emitter-side polarizing plate is orthogonal to the polarization direction of the light-receiving-side polarizer.

FIG. 22 is a configuration diagram of a transparent object detection system.

FIG. 23 is a diagram showing an example of a transparent foreign matter that can remain inside the bottle of FIG. 22;

FIG. 24 is a flowchart executed by the image processing apparatus of FIG. 22 when the polarization direction of the light-emitter-side polarizing plate matches the polarization direction of the light-receiving-side polarizer.

[Explanation of symbols]

 REFERENCE SIGNS LIST 10 transparent body 20 bottle 11, 21 light projecting device 112, 212 light emitting side polarizing plate 12, 22 polarizing device 121, 321 light receiving side polarizing plate 13 television camera 14, 24, 34 image processing device 15 display 16 robot controller 17 robot hand

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Shinji Hatazawa 1048 Kazuma Kadoma, Kadoma City, Osaka Prefecture F-term in Matsushita Electric Works, Ltd. (reference) 2G051 AA41 AB03 AB15 BA11 CA04 CB02 DA03 DA13 EA11 EA12 EA16 EB01 EC01 ED09 ED15 5B057 AA04 BA02 DA03 DA08 DC08 DC34 5L096 BA03 CA02 CA17 EA02 EA12 FA34 FA54 FA69 GA28 GA34

Claims (19)

[Claims] 1. A light polarized in a specific direction is projected onto a predetermined region including a transparent body having a characteristic of changing the polarization direction, and the light transmitted through the predetermined region is transmitted through only a specific polarization. And transmitting the polarized light to the two-dimensional imaging device and receiving the light, and detecting the transparent body using a light / dark distribution in an image obtained from the two-dimensional imaging device. Method. 2. A method according to claim 1, wherein the direction of polarization of the light projected onto the predetermined region is orthogonal to the direction of polarization of the polarization unit provided at the front of an imaging device including the two-dimensional imaging device.
2. The transparent object detection according to claim 1, wherein the two-dimensional image sensor receives light, and a position and a size of the transparent object are detected by using a brightness distribution in an image obtained from the two-dimensional image sensor. Method. 3. The light projected on the predetermined area is made to have the polarization direction of the light coincide with the polarization direction of the polarization means provided at the front of an imaging device including the two-dimensional imaging device.
2. The transparent object detection according to claim 1, wherein the two-dimensional image sensor receives light, and a position and a size of the transparent object are detected by using a brightness distribution in an image obtained from the two-dimensional image sensor. Method. 4. A light projecting device for polarizing light in a specific direction and projecting the light to a predetermined region including a transparent body having a characteristic of changing a polarization direction, and a specific polarization of light transmitted through the predetermined region. A polarizing means for transmitting only the light, and an image obtained by receiving the light transmitted through the polarizing means.
A transparent object detection system, comprising: an imaging device that captures an image with a two-dimensional image; and an image processing device that detects the transparent object by using light and dark distribution in an image obtained from the imaging device. 5. An image obtained by receiving light transmitted through said polarizing means in a state where the direction of polarization of the light projected on said predetermined area is orthogonal to the direction of polarization of said polarizing means. The image processing apparatus detects a position and a size of the transparent body using a light-dark distribution in an image obtained from the imaging apparatus. The transparent object detection system according to the above. 6. An image obtained by receiving light transmitted through the polarizing means in a state where the polarization direction of the light projected on the predetermined area and the polarization direction of the polarizing means coincide with each other. The image processing apparatus detects a position and a size of the transparent body using a light-dark distribution in an image obtained from the imaging apparatus. The transparent object detection system according to the above. 7. The image pickup device has a two-dimensional image pickup device that outputs each pixel signal forming the two-dimensional image, and the image processing device converts each pixel signal from the two-dimensional image pickup device into two. A memory for accumulating each pixel value obtained by the binarization, calculating coordinates of a position of a bright portion or a dark portion on an image formed by each pixel value of the memory, and detecting the transparent body. The transparent object detection system according to any one of claims 4 to 6, wherein: 8. The image processing apparatus counts the number of pixels for each block of the light or dark portion, and when the count value exceeds a predetermined number of pixels, the center of gravity of the block is determined by the transparent body. 8. The transparent object detection system according to claim 7, wherein the position of the transparent object is obtained, and the second moment of the mass is obtained as the posture of the transparent object. 9. The image processing apparatus according to claim 1, wherein the image processing device sequentially performs an expansion process and an erosion process on an image obtained from each pixel value stored in the memory, and erases a lump portion smaller than a predetermined pixel. 9. The transparent object detection system according to claim 8, wherein the number of pixels is counted for each bright or dark block on the processed image. 10. The image pickup device has a two-dimensional image pickup device for outputting each pixel signal forming the two-dimensional image, and the image processing device multiplies each pixel signal from the two-dimensional image pickup device. A memory for storing each pixel value obtained by binarizing, calculating coordinates of a characteristic position of a brightness distribution on an image formed by each pixel value of the memory, and detecting the transparent body. The transparent object detection system according to any one of claims 4 to 6, wherein: 11. The image processing apparatus performs a normalized correlation process between a pre-registered transparent body model and a brightness distribution on the image, and a position having the highest correlation value obtained by the normalized correlation process. Is detected as the position of the transparent body. 12. The image processing apparatus performs a normalized correlation process between a plurality of models of transparent bodies registered in advance having different rotation directions and a lightness distribution on the image, and is obtained by the normalized correlation process. 11. The transparent object detection system according to claim 10, wherein a position having the highest correlation value is detected as the position of the transparent object, and a rotation direction of the model at that time is detected as an attitude of the transparent object. 13. A polarization direction of said polarization means installed at a front part of an imaging device including said two-dimensional imaging device is variable,
2. The transparent object detection method according to claim 1, wherein the position of the transparent object is detected based on a difference in the amount of light received by the imaging device. 14. The image processing apparatus according to claim 1, wherein when the transparent object does not exist as a subject of the imaging device, the polarization direction of the polarization unit is changed to various directions to obtain various images obtained by the imaging device. And the brightness of a predetermined area in each of these images is measured, and the position of the polarizing means when the brightness is minimized is determined by the polarization direction of the light projected on the predetermined area. And the polarization direction of the polarizing means are determined as orthogonal. At the determined position of the polarizing means, when the transparent body is present as a subject of the imaging apparatus, the position is obtained by imaging with the imaging apparatus. 14. The transparent object detection system according to claim 7, wherein the transparent object is detected by capturing an image to be obtained, calculating coordinates of a position of a bright portion on the image, and detecting the transparent object. 15. The image processing apparatus according to claim 1, wherein, when the transparent body does not exist as a subject of the imaging device, the polarization direction of the polarization unit is changed to various directions to obtain various images obtained by the imaging device. And the brightness of a predetermined area in each of these images is measured, and the position of the polarizing unit when the brightness is maximized is determined by the polarization direction of the light projected on the predetermined area. Is determined as a position where the polarization direction of the polarization unit coincides with the polarization direction. When the transparent body is present as a subject of the imaging device at the determined position of the polarization unit, the position is obtained by imaging with the imaging device. 14. The transparent object detection system according to claim 7, wherein the transparent object is detected by capturing an image to be obtained, calculating coordinates of a position of a dark portion on the image, and detecting the transparent object. 16. A control for moving the robot hand to the position of the transparent body detected by the robot hand, adjusting the posture of the robot hand to the detected posture of the transparent body, and picking up the transparent body. The transparent object detection system according to claim 7, further comprising: a robot controller that performs the following. 17. The image processing apparatus detects a position and an orientation of a transparent object to be inspected, and detects the transparent object to be inspected on an image formed by each pixel value of the memory based on the detected position and orientation. 8. A method according to claim 7, further comprising: determining an existence region of the object, comparing the shape of the existence region with the shape of a non-defective transparent body, and detecting a shape defect such as a chip or a burr from the comparison result. 13. The transparent object detection system according to any one of items 1 to 12. 18. The image processing apparatus detects a position and a posture of a transparent object to be inspected, and, based on the detected position and posture, a transparent object to be inspected on an image formed by each pixel value of the memory. Next, an image of the inspection object is taken without using the polarizing means, and a second region is determined.
Of the brightness distribution of the brightness distribution of the area corresponding to the presence area of the second image is determined to be within a normal range, and the dirt, scratches, cracks, etc. 13. The transparent object detection system according to claim 7, wherein a defect in appearance is detected. 19. The transparent body is a transparent foreign substance that can be mixed into an empty bottle or a bottle containing a transparent liquid, and the image processing device is configured to determine an area value or a shape for each block of the bright portion or the dark portion. Have a transparent foreign matter inspection function in the bottle that determines whether or not the transparent foreign matter is mixed in the empty bottle or the bottle containing the transparent liquid according to the measurement result. The transparent object detection system according to claim 7, wherein: