JPH0773693B2 - Method for recognizing an object such as an onion with a peeled thin skin - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to an onion or the like in its shape,
When you sort by size or try to measure dimensions, even if the subject is a thin skin (skin) that is attached to the main body such as an onion with some peeled The present invention relates to a method for recognizing an object such as an onion having a peeled thin skin that prevents the thin skin from erroneously determining the shape, size, and the like of the main body when the image is captured and image-processed using a microcomputer or the like.

[0002]

2. Description of the Related Art Conventionally, for example, when shipping onions to the market, they are once collected at a collection site, and the shape, size, etc. are selected, and the onions are packed in boxes or bags and shipped.

In such a case, the above-mentioned selection of onions has generally been performed manually until now, which requires a lot of labor and time and is extremely inefficient. It has become the target.

Therefore, it has already been proposed to efficiently select onions on the conveying line by image processing using a camera. However, this type of image processing, as shown in FIG.
Place the subject b with onion etc. on it and put the arrow c
As described above, light is emitted from above and the subject b is detected as an image by the camera d arranged on the upper surface side.

However, when the above method is used, not only is it difficult to obtain a clear image of the subject b due to the influence of the reflected light from the conveyor belt a and other objects, but also the subject b does not have an onion or the like. When the thin skin (skin) e is partly peeled off from b ′ and reflected light from the thin skin e is also incident on the camera d, the main body b ′ and the thin skin e can be distinguished from each other. This makes it difficult to correctly sort the onions based on their size and shape.

For this reason, the shape of the body b'of the subject b is assumed by assuming the shape of the body b'with a lift in consideration of the shape of the subject b to some extent. However, there are drawbacks such that the shape obtained by the camera is different from the actual shape, and a lot of time is required for the lift process.

In addition, as described above, a photographed image is taken by using the camera.
Use charge-coupled devices or optoelectronic devices instead of capturing images
And irradiate the transported object with a light beam from a light source,
The projection image of the subject formed by this is obtained, and this projection image
The projected image part obtained by peeling thin skin from the image,
Determined based on preset values in the vertical and horizontal directions
It is possible to consider that the part that was cut is an unnecessary part and cut it off.
It has been proposed (Japanese Patent Laid-Open No. 62-102875). Only
However, the means of such projection type shape selection also determines the above set values.
An error may occur depending on the content, and the projected image
In addition, the complicated image processing will be required.
It

[0008]

SUMMARY OF THE INVENTION The present invention is based on the above-mentioned prior art.
In view of the problems of technology, we should not use projected images.
Place the subject on a translucent carrier and illuminate it.
Focus on the subject under the transmitted light by applying
By taking a picture of the camera,
Detection of analytes without being affected by reflected light
Body that not only allows the light of the subject to pass through
And a thin skin that is translucent,
Small and easy to distinguish,
Accurately detects the actual shape and size of the body excluding the skin
However, trying to improve the reliability of selection and measurement is
Is the purpose.

[0009]

The present invention has the above-mentioned object.
In order to achieve this, a subject with a peeled skin should
On the way to place it on the sending object at a desired interval and transfer it
And illuminate the subject from one side of the carrier.
Then, using the transmitted light by this,
Take a picture with the subject in focus with the placed camera,
From the captured image obtained by this, the
Removed only the shadow image part to detect the subject
Subject identification such as onions with peeled thin skin characterized by
It is an attempt to provide a method of knowledge.

[0010]

[ Function ] Subject whose thin skin is partially peeled and extended
On a translucent carrier, and then
Side of the light source, and then the camera on the other side of the light source.
La, the focal point is the subject, that is, its outline
Get the image taken by the camera by adjusting to the minute
Therefore, the depth of focus of the transmissive carrier is deep.
And the carrier is clearly distinguished from the image of the subject.
be able to.

Further, since the peeled thin skin easily transmits light,
The distinction between the non-translucent body and the thin skin can be distinguished by the camera.
It becomes clear as the light intensity is large and small.
The actual shape and size of the body can be easily identified.
The Rukoto.

[0012]

EXAMPLE An example of the method according to the present invention will be described below with reference to the apparatus illustrated in FIG.
In the figure, 1 is the transport
Conveying members such as a conveyor belt, etc. are 2, 3, 4 respectively.
A drive roller for driving the roller
A tension roller, and a ball on the carrier 1.
Place one subject 5 such as a finger at a predetermined interval,
Sequentially cover the required locations where a box not shown is prepared.
The sample 5 can be transported.

The carrier 1 substantially transmits light.
To prevent this, use a wire mesh made of stainless steel or a transparent material.
And a bottom plate of the carrier 1 in the illustrated example.
A light source 6 such as a lighting fixture is arranged on the side,
Then, illuminate the subject 5 on the carrier 1 from the bottom side.
I am trying to get it.

On the optical axis of the light source 6, in the illustrated example
Has a camera 7 arranged on the upper surface side of the carrier 1.
Of course, the light source 6 is on the upper surface side of the carrier 1, and the light source 6 is on the lower surface side.
The cameras 7 may be installed respectively.

When the above-mentioned apparatus is used, the light source 6 is used.
The subject 5 is irradiated with light from the lower surface side of the carrier body 1,
On the side, the subject 5 is photographed by the camera 7, and the photograph is taken.
You will be capturing images. At this time, the focus of the camera 7 is
The focus of the carrier 1 can be adjusted by aligning with the outer shape of the sample 5.
Since the depth becomes deeper, the carrier 1 takes an image of the subject 5.
It will be clearly distinguishable from the image.

Further , the transmitted light from the light source 6 is used to
Since the image of the subject 5 is taken by the camera 7,
Peeling thin skin 5b partially peeled from the main body 5a of No. 5 and extending
Transmits the above light, while the main body 5a
Images that are taken by the camera 7 as light is not substantially transmitted
The peeled thin skin 5b of the subject 5 is
It can be clearly distinguished based on the amount of received light in 7
The Rukoto.

Contrary to the illustrated example, the upper surface side
More illumination and capture the image taken by the camera 7 on the bottom side.
When carrying out jamming, the carrying body 1 and
The photographed image of the subject 5 and the peeled thin skin 5b and the main body 5a
Of course, a clear distinction is made.

In this embodiment, the onion carrying line is used.
For image processing when selecting the shape and size
As mentioned above, the present invention is not limited to the
Also adopted when measuring the external dimensions of the body with a camera
be able to.

[0019]

The present invention is implemented as described above.
Therefore, the subject is placed on a translucent carrier.
Let's illuminate it from one direction and let the camera from the other direction.
The captured image, and at this time,
By focusing the focus of the camera on the outline of the subject,
So that the depth of focus is different from the
Therefore, it is not affected by the reflected light like the conventional method.
Therefore, it is possible to clearly distinguish between the subject and the carrier.
Image processing can be performed.

Further, a photographed image of the subject is taken with a camera.
When entering, the transmitted light from the illumination was used as described above.
Therefore, it seems that the peeled thin skin that has peeled off from the main body has partially extended
Unlike the main body, the peeled thin skin transmits light even if it is the subject.
Since it is easy to distinguish from the main body,
For image processing of objects with peeled thin skin such as green onions
Even if you do not recognize the wrong shape, size, size, etc.
Reliable results.

[Brief description of drawings]

FIG. 1 is a use state explanatory view showing an example of a device that can be used for carrying out a method of recognizing an object such as an onion having a peeled thin skin according to the present invention.

FIG. 2 is an explanatory diagram showing an implementation state of a conventional object recognition method using a camera.

[Explanation of symbols]

 1 Transport Object 5 Specimen 5b Peeling Thin Skin 7 Camera

Claims (1)

[Claims] 1. An object having a peeled thin skin is placed on a translucent carrier at a desired interval, and the object is transferred to the object from one surface side of the carrier. Illuminate, using the transmitted light by this, the same as the above, the camera arranged on the other surface side of the transporting body is used to focus and image the subject ,
From the captured image obtained by this , the
A method of recognizing an object such as an onion having a peeled thin skin, wherein only the shadow image portion is removed to detect the object.