JP3290021B2 - Bottle color recognition method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for recognizing the color of a bottle based on transmitted light or reflected light.

[0002]

2. Description of the Related Art When collecting and recycling various discarded bottles, it is necessary to sort these various bottles by color. For example, there is an apparatus as disclosed in Japanese Patent Application Laid-Open No. Hei 5-27747. This apparatus sorts and sorts bottles by size, recognizes the colors of the bottles by color recognition means, and sorts and collects the bottles by color.

Further, as disclosed in Japanese Patent Application Laid-Open No. 3-200025, a method of irradiating a bottle with light from an illuminating device and recognizing the bottle color based on the transmitted light transmitted through the bottle is used. There is one that automatically sorts various types of bottles.

[0004]

An apparatus for recognizing the color of a bottle using the above-mentioned transmitted light or the like is incorporated in an automatic bottle sorting apparatus. In such a case, if the irradiation light corresponding to the missing portion is determined as a part of the image data, there is a problem that the bottle color recognition based on the image data becomes inaccurate.

[0005]

[0006]

A bottle color recognition method according to the present invention irradiates a conveyed bottle with light, performs image processing using the transmitted light or reflected light, and recognizes the color of the bottle. In the method, when a transmitted light or a reflected light of a bottle is imaged by a camera at a predetermined position, two specific image signals having different scanning positions with respect to the height direction of the bottle are captured, and the specific image signal in the imaging region is taken out of each specific image signal Recognize the color of the bottle by extracting any one of R, G, and B maximum luminance values, comparing both extracted maximum luminance values, and performing image processing based on the smaller image signal. It is characterized by doing.

[0007]

[0008]

[0009]

According to the present invention, image data in two image extraction lines are fetched, the maximum value having the highest luminance value is extracted from each of these image data, and the smaller one of the maximum values is selected and the maximum value is selected. By performing the bottle color recognition processing based on the smaller image data of the bottle, the middle part of the bottle is missing, and even if the irradiation light corresponding to this missing part is captured as a part of the image data, The image data of the corresponding image extraction line is not used, and the color recognition of the bottle is performed based on the image data of the other image extraction line.

[0010]

[0011]

An embodiment of the present invention will be described below with reference to the drawings.

FIGS. 1 and 2 show a schematic configuration of a bottle color recognition apparatus.

In FIG. 1, reference numeral 1 denotes a transport conveyor, which transports bottles sent in a state of being aligned by an alignment device (not shown) to a sorting device 3 via a color recognition device 2 and has an upstream side. And a second conveyor 12 on the downstream side connected to the first conveyor 11.
It is composed of

At the end of the first conveyor 11 on the loading side,
The above-described alignment device (not shown) is connected, and bottles supplied from a bottle supply device such as a hopper (not shown) are aligned in a predetermined state, and the bottles are sequentially carried into the first conveyor 11 one by one.

The first conveyor 11 and the second conveyor 12
Means that the feed speed of the second conveyor 12 is
Is set to be faster than the feed speed. By setting the transport speed between the first conveyor 11 and the second conveyor 12 in this manner, the bottles transported from the first conveyor 11 to the second conveyor 12 have a higher transport speed and are positioned on the first conveyor 11. The distance between the bottle and the bottle increases. Therefore, even if the front and rear bottles are conveyed in a continuous state on the first conveyor 11, the bottle on the front side (downstream side) is transferred to the second conveyor 12 and conveyed by the second conveyor 12, so that the rear The side (upstream) bottle will be separated and the front and rear bottles will be spaced apart.

The color recognition device 2 is installed on the upstream side of the second conveyor 12, and illuminates a bottle flowing on the second conveyor 12 with light, and a second device facing the illumination device 21. A camera 2 provided on the other side of the conveyor 12 for imaging light emitted from the illumination device 21 via a bottle.
2 and an image processing device 23 that processes an image captured by the camera 22 and recognizes the color of the bottle. That is, the color recognition apparatus in this embodiment is an image processing system using transmitted light of a bottle, but may use a reflected light.

The illumination device 21 and the camera 22 are installed in a cover 24. The cover body 24 forms a dark room that shields the illumination device 21 and the camera 22 from the outside and also shields a part of the second conveyor 12 on which the illumination device 21 and the camera 22 are arranged. This is a station that captures the transmitted light that has passed through. Thus, since the transmitted light of the bottle is read in the dark room, accurate color recognition data can be obtained without being affected by noise light.

A detection sensor (a bottle detection device) 25 for detecting that the bottle flowing on the second conveyor 12 has reached this position is provided on the upstream side of the illumination device 21. The output signal from the detection sensor 25 is output from the control device C
Input to O.

As the detection sensor 25, for example, a phototube, a photodiode, a phototransistor, an optical sensor such as CdS, or the like is used. For example, the bottle is formed by blocking light emitted from the light projector 25a to the light receiver 25b. Detect the presence or absence of

The sorting device 3 is installed on the downstream side of the second conveyor 12 and, based on the result of the bottle color recognition by the image processing device 23, sorts each bottle appropriately by color through the control device CO. 46, each collection unit 41-4
5 housed.

The collection units 41 to 45 include, for example, a first collection unit 41 for collecting colorless bottles from the left side shown in FIG. 1, a second collection unit 42 for collecting pale blue bottles, and a black bottle. The third collecting section 43 for collecting, the fourth for collecting brown bottles
A collection unit 44 and a fifth collection unit 45 for collecting dark green bottles on the rightmost side are configured to be able to collect five types of bottles. In addition, the number of collection units is not limited to this example, and a necessary number of sorting units may be provided.

As the distributing means 46, a pusher is provided in each of the first to fifth collecting units 41 to 45, and based on the recognition result of the image processing device 23, the corresponding pusher is operated via the control device CO. By doing so, the bottles are forcibly pushed into the corresponding first to fifth collection units 41 to 45, and the bottles are sorted. Alternatively, the bottles are turned over the upper surface of the second conveyor 12 so as to protrude. For example, one provided with a guide member for guiding the flowing bottle to each of the first to fifth recovery units 41 to 45 is used.

Accordingly, if the image processing apparatus 23 recognizes that the bottle flowing on the second conveyor 12 is, for example, colorless by a recognition method described later, the bottle determined to be colorless is distributed by the first collection unit 41 by the sorting means 46. Pay out and collect.

A residue storage unit 5 is provided below the downstream discharge end of the second conveyor 12, and the residue storage unit 5 is transported together with the bottle by the first conveyor 11 and the second conveyor 12. The collected garbage and the bottles not to be sorted are stored.

The image processing device 23 receives data captured by the camera 21 as image data (image signal) via an image processing unit 231.

As the image data, for example, data as shown in FIG. 3 is obtained. In FIG. 3, a line indicated by a dashed line is an image extraction line D2 for extracting an image of a bottle.
, The image data extracted from the image extraction line D2 is shown in the lower part. The image data is analyzed into the luminances r, g, and b of the three primary colors of the transmitted light. A detailed recognition method will be described later. The symbol Z in FIG. 3 is a label portion affixed to the bottle, and the brightness level of transmitted light transmitted through this portion is naturally smaller than that of other portions.

Further, the timing for taking in the image data is performed based on the detection of the detection sensor 25. Specifically, as shown in FIG. 4, when the bottle is conveyed on the second conveyor 12 and the front end of the bottle is detected by the detection sensor 25, the output signal of the detection sensor 25 rises from L to H (see T1), When the paper is further conveyed and the rear end of the bottle passes, the output signal falls from H to L (see T2). When the signal falls at this T2, that is, when the signal is H
When the mode is switched from L to L, the image data is fetched from the control device CO to the image processing device 23, and an instruction is issued to perform the bottle color recognition process. The image processing device 23 captures the image data after a predetermined time has elapsed, and performs color recognition.

Here, the details of the recognition method for recognizing the color of the bottle by the image processing device 23 will be described with reference to FIGS.
This will be described with reference to the flowchart shown in FIG.

First, in recognizing the bottle color, the irradiation light emitted from the illumination device 21 is not transmitted through the bottle, that is, the three primary colors [R (red), G (Green), B (blue)].

This measurement is performed immediately before the recognized bottle is conveyed by the second conveyor 12 and reaches the imaging area where the image processing by the color recognition device 2 can be performed. It is executed within the time immediately after it ends and disappears from the screen of the camera 22.

The measured luminance values are not actually constant, and the average of these luminance values is determined as Ra, Ga, and Ba (step S1), and is set as the R, G, and B luminance values.

Then, in a step S2, it is determined whether or not a bottle is detected. The detection of the bottle is performed based on whether or not the output signal of the detection sensor 25 has risen from L to H as described above.

At this time, the image processing device 23 sets any one of the measured values R, G, and B of the irradiating light not transmitted through the bottle as a reference value and sets the reference value and other reference values. The difference between the two luminance values is calculated.

For example, of the measured values Ra, Ga and Ba, B
When the luminance value of a is used as a reference value, differences dR and dG between the luminance value of Ba and the luminance values of other Ra and Ga are obtained (step S3).

[0035]

DR = Ba-Ra dG = Ba-Ga Then, in step S4, based on the signal when the output signal from the detection sensor 25 falls from H to L,
Image data obtained by capturing the transmitted light emitted from the illumination device 21 and transmitted through the bottle by the camera 22 is processed by the image processing device 2.
Take in 3.

At this time, as shown in the upper part of FIG. 7, two types of data from two horizontal image extraction lines D1 and D2 having different height levels are taken in from the front end to the rear end of the bottle. It is.

The lower part of FIG. 7 shows the luminance data of R among the data (R, G, B) thus taken in. In FIG. 7, a broken line indicates image data on the image extraction line D1, and a solid line indicates image data on the image extraction line D2.

Then, the maximum values having the highest luminance values are extracted as r1 and r2 from both the image data (step S5). The range in which the maximum value is extracted is a range in which the transmitted light of the bottle on each of the image extraction lines D1 and D2 can be obtained. That is, in the image line D1, a range from the front end B to the rear end A of the bottle, and in the image line D2, the middle part D to the rear end C (=
A).

Next, in step S6, the maximum value r1 at the point Z1 on the image extraction line D1 is compared with the maximum value r2 at the point Z2 on the image extraction line D2, and the smaller one of the maximum values is selected. (Step S7, S
8). That is, if the maximum value r1 is smaller, i = 1, and if the maximum value r2 is smaller, i = 2, and the bottle color recognition process is performed based on the smaller image data.

For example, as shown in FIG. 7, if the maximum value r2 is smaller, the maximum value between C and D of the image data from the image extraction line D2, that is, the three primary colors of the transmitted light at the point Z2. Each luminance value is measured, and the measured values are defined as ri, gi, bi (i = 2) (step S9, r, g, b at Z2 point in the image data shown in FIG. 3).

The ri and g thus extracted and measured
i and bi are luminance values R of the irradiation light of the illumination device 21;
Based on G and B, the brightness values of r and g are corrected using b as a reference value.

The values dr and dg are represented by the following equations (step S10).

[0043]

## EQU00002 ## dr = dR.times.ri / Ra dg = dG.times.gi / Ga Accordingly, the values dr and dg are respectively added to the luminance values r and g of the transmitted light to obtain a corrected luminance value.

[0044]

## EQU3 ## r _new = ri + dr g _new = gi + dg can be obtained (step S11). Since b is a reference value, b _new = b.

Next, r _new and g for the reference b
_new luminance ratio

[0046]

## EQU4 ## r _new / big _new / bi is obtained (step S12), and the bottle color is recognized by applying to the luminance ratio coordinates shown in FIG. 8 (step S13).

The brightness ratio coordinates are created in advance based on a plurality of brightness ratio data, and are input to the image processing device 23 as a data table in advance.

When the color of the bottle is recognized as described above, the sorting unit 46 corresponding to the color is operated via the control device CO, and the bottle is removed from the first to fifth collection units 41 to 45. Is collected in the corresponding collection unit (step S14).

The above operation is repeated as needed until the bottle sorting operation is completed (step S15).

In the present embodiment, the description has been made with reference to B. However, similar results can be obtained by performing the above-described correction with reference to R, or G.

As described above, the image data in the image extraction lines D1 and D2 are fetched, and the maximum values having the highest luminance values are extracted as r1 and r2 from these image data, and the smaller one of the maximum values is selected. In order to perform the color recognition processing of the bottle based on the image data having the smaller maximum value, the middle part of the bottle is missing as shown in FIG. 7, and this missing part is over the image extraction line D1. Even in this case, the color of the bottle is not determined by the light that has not passed through the bottle, and accurate bottle color recognition can be performed based on the image data from the image extraction line D2.

FIG. 9 shows an image processing apparatus 2 for another bottle.
3 shows the captured image data of the luminance value of R.
In this case, the bottom of the bottle is missing instead of the middle of the bottle as shown in FIG. 7, so that color recognition may be performed based on image data from either image extraction line D1 or D2. As described above, the processing is executed according to the flow charts of FIGS. That is, since the maximum value r1 at the Z1 point on the image extraction line D1 is smaller than the maximum value r2 at the Z2 point on the image extraction line D2,
Assuming that i = 1, the maximum value between A and B of the image data from the image line D1, that is, the luminance values of the three primary colors of the transmitted light at the point Z1, are measured, and the color recognition processing of the bottle is performed.

FIG. 10 shows another embodiment of the timing for fetching image data.

In the image data input from the camera 22 to the image processing device 23, a point Q is designated in the image area 221 in advance, and if a fall in luminance occurs at this point Q, the control device CO It is determined that the image has entered the area, and the image data is taken into the image processing device 23 after a certain period of time has elapsed from this time. The setting of the elapsed time is performed based on the transport speed of the second conveyor 12.

[0055]

As described above, according to the bottle color recognition method of the present invention, the image data in the two image extraction lines is fetched, and the maximum value having the highest luminance is extracted from both the image data. In order to select the smaller one of the maximum values and perform the color recognition processing of the bottle based on the image data of the smaller maximum value, the middle part of the bottle is missing and this missing part is extracted as one image. Even if the line passes, the image data is not adopted, and the color recognition of the bottle is performed based on the image data of the other image extraction line, and the bottle color can be accurately recognized based on this image data. it can.

[0056]

[Brief description of the drawings]

FIG. 1 is a plan view showing a schematic configuration of a bottle color recognition device according to the present invention.

FIG. 2 is a side view showing a schematic configuration of a bottle color recognition device according to the present invention.

FIG. 3 is a diagram illustrating an example of image data of transmitted light transmitted through a bottle.

FIG. 4 is a diagram illustrating the timing of capturing image data by a detection sensor.

FIG. 5 is a flowchart illustrating an operation for recognizing a bottle color.

FIG. 6 is a flowchart illustrating an operation for recognizing a bottle color.

FIG. 7 is a diagram showing luminance data of R captured by the method of the present invention.

FIG. 8 is a luminance ratio coordinate diagram for bottle color recognition.

FIG. 9 shows the R of other bottles incorporated by the method of the present invention.
FIG. 5 is a diagram showing luminance data of the first embodiment.

FIG. 10 is a diagram showing another embodiment of the timing for taking in image data.

[Explanation of symbols]

 Reference Signs List 21 lighting device 22 camera 221 image area 23 image processing device 25 detection sensor (bottle detection device) CO control device D1 image extraction line D2 image extraction line

──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Mitsuaki Hirata 30 Environmental Engineering Division, Shinmeiwa Kogyo Co., Ltd. 30, Ono City, Ono City, Hyogo Prefecture (56) References JP-A-5-126761 (JP, A) JP-A-3-111705 (JP, A) JP-A-5-60514 (JP, A) JP-A-6-31251 (JP, A) JP-A-56-140243 (JP, A) JP-A-5-81697 (JP, A) , U) (58) Fields surveyed (Int. Cl. ⁷ , DB name) G01J 3/46-3/51 G01N 21/25-21/27 G01N 21/84-21/958 B07C 5/432 G06T 1 / 00-1/40 G06T 3/00-5/50 G06T 9/00-9/40

Claims (1)

(57) [Claims] 1. A method for irradiating a conveyed bottle with light, performing image processing using the transmitted light or reflected light, and recognizing the color of the bottle, comprising the steps of: When taking an image with a camera, two specific image signals having different scanning positions in the height direction of the bottle are taken in, and one of R, G, and B of the specific image signal in the imaging region is set to the maximum luminance value. , And comparing the extracted maximum luminance values, and performing image processing based on the smaller image signal to recognize the color of the bottle.