JP6555071B2 - Object position recognition method and object transfer method - Google Patents

Description

  The present invention relates to a technique of an object position recognition method. More specifically, the present invention can quickly and accurately find a detection target positioned at the highest position among detection targets that are a plurality of overlapping objects. The present invention relates to a technique of an object position recognition method.

2. Description of the Related Art Conventionally, various techniques are known for finding a detection target positioned at the highest position from among a plurality of detection targets that overlap each other (hereinafter referred to as “detection target group” as appropriate).
For example, “Patent Document 1” discloses a technique for determining the vertical positional relationship of each detection target from among a group of detection target overlapping each other based on the density distribution of the image in the image information. .
In “Patent Document 2”, the center of a predetermined detection figure (detection circle) is run on the outline of the detection object group, and based on the number of intersections between the detection circle and the outline of the detection object group, A technique for determining whether or not the number of overlapping detection objects is one or more is disclosed.

JP-A-8-271223 Japanese Patent Laid-Open No. 11-114860

However, in the technique according to the above-mentioned “Patent Document 1”, for example, when the density difference between overlapping detection objects between the detection objects that overlap each other is very small, it is difficult to determine the vertical positional relationship of these detection objects. There was a case.
Further, in the “Patent Document 2”, it is possible to grasp the number of arbitrary detection objects that overlap a certain detection object in a group of detection objects that overlap each other. It was difficult to judge the positional relationship in the vertical direction.

  The present invention has been made in view of the above-described problems of the present situation, and quickly and accurately finds a detection object positioned at the highest position among detection objects that are a plurality of overlapping objects. An object of the present invention is to provide a method for recognizing the position of an object.

  The problem to be solved by the present invention is as described above. Next, means for solving the problem will be described.

  That is, the object position recognition method according to the present invention is an object position recognition method for recognizing an object positioned at the highest position among a plurality of overlapping objects, and the plurality of objects are illuminated obliquely from above. A first step of irradiating light to obtain image data of the plurality of objects and obtaining contour lines of the plurality of objects; and the image data from which the shadows and the contour lines of the plurality of objects are obtained A second step of performing binarization processing on the third step, a third step of discriminating a region blocked by the shadow and the contour, and a first step of forming a virtual inscribed circle having the maximum diameter in the region A fourth step and a fifth step of recognizing the position of the object located at the highest position with the center position of the other virtual inscribed circle excluding the virtual inscribed circle exceeding the size of the object. It is characterized by being.

As described above, in the present invention, the illumination light L is irradiated obliquely from above on a plurality of objects, and it is easy to form shadows around the plurality of objects (detection targets). As in the technique disclosed in “Patent Document 1”, the difference in density between overlapping objects is rarely small, and the contour lines of the plurality of objects can be recognized more clearly. The search for the object located can be performed more accurately.
Further, according to the object position recognition method of the present invention, the number of arbitrary objects overlapping with a certain object (detection target object) is not grasped as in the technique according to “Patent Document 2” described above. Based on the first to fifth steps, it is possible to recognize the object located at the highest position from among the plurality of overlapping objects.

  In the object position recognition method according to the present invention, it is preferable that the object is made of a disk-shaped member.

  According to the object position recognition method having such a configuration, the shape of the virtual inscribed circle formed in the fourth step can be set to be equal to the shape of the object (detection target). The operation program for executing the first to fifth steps can be simplified, which is economical.

As effects of the present invention, the following effects can be obtained.
That is, according to the object position recognition method according to the present invention, it is possible to quickly and accurately find the detection object positioned at the highest position from among a plurality of overlapping detection objects.

1 is a conceptual diagram illustrating an overall configuration of an object position recognition apparatus according to an embodiment of the present invention. The side view which showed the arrangement | positioning attitude | position of the illumination means. The flowchart which showed the procedure at the time of performing the position recognition of an object with time. It is the figure which showed the content of the image data of a detection target object group, Comprising: The schematic diagram which showed the state which performed the formation process of the outline of the said detection target object group. It is the figure which showed the content of the image data of a detection target object group, Comprising: The schematic diagram which showed the state which performed the binarization process of the said detection target object group. It is the figure which showed the content of the image data of a detection target object group, Comprising: The schematic diagram which showed the state which performed the discrimination | determination process of the area | region of the said detection target object group. It is the figure which showed the content of the image data of a detection target object group, Comprising: The schematic diagram which showed the state which assumed the predetermined virtual inscribed circle for every area | region of the said detection target object group. It is the figure which showed the content of the image data of a detection target group, Comprising: The schematic diagram which showed the state which recognized the detection target located in the highest rank for every area | region of the said detection target group.

  Next, embodiments of the invention will be described.

[Position recognition device 1]
First, an overall configuration of a position recognition apparatus 1 used for implementing an object position recognition method according to the present invention will be described with reference to FIGS. 1 and 2.

  The position recognition device 1 according to the present embodiment searches for and finds the detection target Tg located at the highest position from among a plurality of detection target objects Tg, Tg... (Detection target group T) that overlap each other. As shown in FIG. 1, by irradiating the detection target group T with illumination light L obliquely from above, detecting the contour of the detection target group T, and executing arithmetic processing, This is an apparatus that enables to quickly and accurately find the detection target Tg positioned at the top.

Here, the detection target Tg is made of a plate-like member. In the present embodiment, for example, a disk-shaped object is assumed, but the present invention is not limited to this.
That is, the detection target Tg may be a plate member having a polygonal shape such as a triangular shape or a quadrangular shape.

  However, as will be described later, since the shape of the virtual inscribed circle on the program used for searching for the detection target Tg located at the highest position can be set to be equal to the shape of the detection target Tg. The object Tg is preferably disk-shaped in order to simplify the calculation program.

Further, the plurality of detection objects Tg, Tg,... Overlapped with each other are configured in the same shape.
Specifically, each detection target Tg is configured by a disk member having a diameter of 8 [mm] and a thickness of 0.3 [mm], for example.

  The position recognizing device 1 is mainly composed of an imaging means 10, an illuminating means 20, a picking means 30, a control device 40 that controls the operation of the entire position recognizing device 1, and the like.

The imaging means 10 is for imaging the detection target Tg, and for example, an area camera or a line scan camera provided with a CCD (Charge Coupled Device) or a CMOS (Complementary Metal Oxide Semiconductor) can be used. .
The imaging unit 10 has a function of efficiently acquiring high-resolution image data.

  As shown in FIG. 2, the imaging unit 10 is disposed so as to be able to capture an image from directly above the detection target group T input to the position recognition device 1.

  The image data acquired by the imaging means 10 is sent to the control device 40 and subjected to image processing based on the stored arithmetic program.

Next, the illumination means 20 will be described.
The illumination unit 20 is for irradiating the subject (detection target group T) imaged by the imaging unit 10 with the illumination light L.
The illumination unit 20 is configured by, for example, a direct-type bar illumination that can irradiate the subject including the detection target group T with the irradiation light L at an arbitrary angle.

  When the illumination unit 20 is configured by direct-type ring illumination, illumination light is irradiated on the detection target group T from the obliquely upward direction over the entire circumference, and a shadow is formed around the detection target group T. It is not appropriate because it becomes difficult.

  Further, a plurality of irradiation means 20 and 20 are provided for one position recognition device 1 (for example, two in the present embodiment), and illumination light L is applied to the detection target group T obliquely from above. It arrange | positions so that it may irradiate.

As described above, in the present embodiment, the detection target group T that is a subject is irradiated with the illumination light L obliquely from above, and it is easy to form a shadow around the detection target group T. The contour line of the detection target group T is more prominent.
Therefore, the contour line of the detection target group T can be recognized more clearly, and the search for the detection target Tg positioned at the highest level can be performed more accurately by a calculation program described later.

When the inclination angle θ formed by each illumination unit 20 with respect to the virtual vertical line V is less than 30 ° (θ <30 °), the width of the shadow formed around the detection target group T becomes relatively short. It becomes difficult to recognize.
On the other hand, when the inclination angle θ formed by each illumination unit 20 with respect to the virtual vertical line V exceeds 70 ° (θ> 70 °), the width of the shadow formed around the detection target group T becomes too long, It becomes difficult to recognize.
For these reasons, it is desirable that the inclination angle θ formed by each illumination unit 20 with respect to the virtual vertical line V is 30 ° or more and 70 ° or less (30 ° ≦ θ ≦ 70 °).

  Further, with respect to the radix of the illumination means 20, 20..., It is preferable to provide two or more so that a shadow is formed in all directions around the detection target group T.

Next, the picking means 30 will be described.
As shown in FIG. 1, the picking means 30 is for taking out the detection target Tg located at the highest position from the detection target group T.
The picking means 30 is composed of, for example, a suction pad and is provided with a moving mechanism unit (not shown) that moves in an arbitrary direction.

  Then, the picking means 30 sucks and holds the selected detection target Tg, and then carries the detection target Tg out of the detection target group T via a moving mechanism (not shown). .

Next, the control device 40 will be described.
The control device 40 includes a storage unit and a calculation unit. The storage unit processes the image data captured by the imaging unit 10 together with programs relating to the operation of the imaging unit 10 and the picking unit 30, and the detection target group. A program for finding a detection target positioned at the top in T is stored in advance.

  In addition, the control device 40 is provided with input means such as a keyboard and a barcode reader, and is provided with output means such as a monitor.

  And the control apparatus 40 is comprised so that the driving | operation of the position recognition apparatus 1 whole may be controlled.

[Operation Procedure of Position Recognition Device 1]
Next, the operation procedure of the position recognition apparatus 1 will be described with reference to FIGS.
First, the detection target group T (see FIG. 1) is put into a predetermined place in the position recognition device 1.

In this case, in the detection target group T, a plurality of detection target objects Tg, Tg,...
Further, the detection target group T is irradiated with illumination light L obliquely from above by the illuminating means 20, and the detection target group T has a contour line more clearly raised by a shadow formed around it. It is held in the state.

  The inputted detection target group T is imaged from directly above by the imaging unit 10, and the acquired image data of the captured image is sent to the control device 40.

The control device 40 that has received the image data performs image processing on the image data based on a program stored in advance, and starts a search operation for the detection target Tg located at the highest position in the detection target group T. To do.
Specifically, as shown in FIG. 3, as a first step, an extraction process for extracting and obtaining the contour line of the detection target group T based on the density of the image data is executed (step S01).

  As a result, the image data picked up by the image pickup means 10 is processed into the first image data 100 having the outline 101 (displayed by a solid line) of the detection target group T as shown in FIG.

  When the contour line extraction process is completed, as shown in FIG. 3, as a second step, the binarization process is performed on the first image data 100 from which the contour line 101 of the detection target group T is obtained. (Step S02).

  As a result, as shown in FIG. 5, the first image data 100 is the second image data in which only the outline 201 of the detection target group T and the plurality of shadows 202, 202. 200.

  When the binarization process for the detection target group T is completed, as shown in FIG. 3, as a third step, a region discrimination process for the detection target group T is performed on the second image data 200 ( Step S03).

  As a result, the second image data 200 is divided into regions that are blocked by the contour line 201 and the shadow 202, and, for example, as shown in FIG. 6, the region A1 outside the contour line 301 of the detection target group T, for example. , And the third image data 300 composed of the area A2 and the area A3 inside the outline 301.

  When the region discrimination processing in the detection target group T is completed, as shown in FIG. 3, in the discriminated regions A1, A2, and A3, as shown in FIG. A process of creating the largest virtual inscribed circle that touches is executed (step S04).

  Here, as the “virtual inscribed circle”, a circular virtual figure having a shape equivalent to the detection target Tg is set in the program.

  As a result, as shown in FIG. 7, the third image data 300 is, for example, a fourth image having virtual inscribed circles D1, D2, and D3 that are inscribed in the regions A1, A2, and A3 and have different inner diameter dimensions. The image data 400 is processed.

  When the creation process of the virtual inscribed circles D1, D2, and D3 is finished, as shown in FIG. 3, the virtual inscribed circles D1, D2, and D3 are used as the fifth step in the detection target group T. A selection process of a desired detection target Tg, that is, the detection target Tg positioned at the highest level is executed (step S05).

Specifically, for example, among these virtual inscribed circles D1, D2, and D3, an inner diameter dimension equivalent to the inner diameter dimension of the detection target Tg is selected, and the selected virtual inscribed circle (for example, the book In the embodiment, the position of the center G of the virtual inscribed circles D2 and D3) is recognized as the position of the desired detection target Tg.
That is, the position of the center G of the other virtual inscribed circles D2 and D3 excluding the virtual inscribed circle D1 exceeding the size of the detected object Tg is recognized as the position of the detection object Tg located at the highest position.

  As a result, as shown in FIG. 8, the fourth image data 400 is transferred to the fifth image data 500 in which the detection objects Tg2 and Tg3 positioned at the top are indicated by the virtual inscribed circles D2 and D3. And processed.

  Thus, when the detection target Tg positioned at the highest position is found from the detection target group T input to the position recognition device 1, the control device 40 ends the search operation for the time being.

  Thereafter, the control device 40 transmits an operation signal to the picking means 30, and the picking means 30 that has received the operation signal is detected by the virtual inscribed circles D2 and D3 in the fifth image data 500. The objects Tg2 and Tg3 are supported and carried out to a predetermined place.

  When the carrying out of the detection objects Tg2 and Tg3 by the picking means 30 is completed, the control device 40 repeats the above-described steps S01 to S05 again to search for the detection object Tg newly positioned at the top. Execute.

  Thereafter, the control device 40 transmits an operation signal to the picking unit 30 again, and the picking unit 30 that has received the operation signal supports the newly detected object Tg and moves it to a predetermined place. And carry it out.

  As described above, the search operation of the detection target Tg positioned at the highest position by the control device 40 and the unloading operation of the detection target Tg by the picking means 30 are alternately and repeatedly executed and input to the position recognition device 1. When all of the detected object groups T are carried out by the picking means 30, the operation of the position recognition device 1 is completed.

  In the present embodiment, after a plurality of virtual inscribed circles D1, D2, and D3 are created in step S04, the selection processing of the detection target Tg positioned at the highest position is executed in step S05. However, after creating one virtual inscribed circle in step S04, it is also possible to perform the selection process of the detection target Tg positioned at the top in step S05.

  For example, after creating the virtual inscribed circle D1 (step S04), it is selected whether or not the virtual inscribed circle D1 is the detection target Tg positioned at the top (step S05), and the virtual inscribed circle D1 is selected. Is not the highest detection target Tg, step S04 is executed again to form a virtual inscribed circle D2, and whether or not the virtual inscribed circle D2 is the highest detection target Tg. Is selected (step S05).

When the virtual inscribed circle D2 is selected as the detection object Tg positioned at the top, the position of the center G of the selected virtual inscribed circle D2 is recognized as the position of the desired detection object Tg, and the search operation is performed. It can also be configured to end for the time being.
In other words, the search operation can be completed for the time being when one detection target Tg positioned at the top is selected.
In this case, after that, the detection target Tg2 at the location indicated by the virtual inscribed circle D2 is carried out by the picking means 30, and the steps S01 to S05 are repeated again to search for the detection target Tg positioned at the top. Is executed again.

100 First image data 101 Contour line 201 Contour line 202 Shading A1 area A2 area A3 area D1 Virtual inscribed circle D2 Virtual inscribed circle D3 Virtual inscribed circle L Illumination light S01 Step (first process)
S02 step (second step)
S03 Step (Third Step)
S04 step (fourth step)
S05 Step (5th process)
Tg Object to be detected (object)

Claims (3)

An object position recognition method for recognizing an object located at the highest position among a plurality of overlapping objects each having an equivalent shape and size ,
Illuminating the plurality of objects with illumination light obliquely from above,
A first step of obtaining image data of the plurality of objects to obtain contour lines of the plurality of objects;
A second step of performing binarization processing on the image data for which the shadows and the contour lines of the plurality of objects have been obtained;
A third step of determining an area blocked by the shadow and the contour line;
A fourth step of forming a virtual inscribed circle of maximum diameter in the region;
Among the other virtual inscribed circles excluding the virtual inscribed circle exceeding the size of the object , the inner diameter dimension is located at the top position with the center position of another virtual inscribed circle equivalent to the inner diameter dimension of the object. A fifth step of recognizing the position of the object to be
Composed of,
An object position recognition method characterized by the above. The object is made of a disk-shaped member,
The object position recognition method according to claim 1, wherein: The object positioned at the uppermost position recognized by the object position recognition method according to claim 1 or 2 is sucked and held by a suction pad,
Carrying out of the object group consisting of the plurality of objects;
An object conveying method characterized by the above.

Images