JP2000153239A - Device for discriminating glass bottles - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device which an automatically sort smoke bottles from light transmissible bottles. SOLUTION: This sorting device is provided with a light source 3 for projecting light to bottles, a light receiving device 5 for receiving transmission light from the light source 3 through the bottles, and treating device 6 for discriminating the bottles by a signal from the light receiving device 5. The treating device 6 has a means for calculating light receiving strength and light receiving distribution of the light receiving device 5 and a means for judging the difference between the light transmissible bottles and the smoke bottles scattering light such as frosted glass bottles from the light receiving strength and light receiving distribution.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to
In particular, the present invention relates to an apparatus for automatically identifying waste glass bottles to be recycled.

[0002]

2. Description of the Related Art At present, collection of waste bottles has been actively carried out as an object of recycling in the treatment of municipal solid waste. These are mainly sorted by color,
Crushed into cullet, used as a raw material to produce new bottles of the same color. Some municipalities will collect them as returnable bottles.

As a cullet, a transparent bottle is particularly valuable, and it is an important technical problem to separate and remove a smoked bottle in order to increase the sorting purity of the transparent bottle. We had to rely on the eyes of the sorting workers.

[0004] As a prior art of glass bottle sorting, as disclosed in Japanese Patent Application Laid-Open No. 8-108145 ("empty bottle sorting apparatus"), the bottom of a bottle is imaged by a camera, and the light-colored bottle color is the image. In the periphery of the bottom of the bottle (the portion where the selective absorption of transmitted scattered light is large), the dark bottle color indicates the number of pixels within the specific hue range in the central portion of the bottle (the portion where the selective absorption of transmitted scattered light is small). There is a method of determining the color by using

Further, as another method for sorting glass bottles, as disclosed in Japanese Patent Application Laid-Open No. Hei 7-96254 ("Method for Color Separation of Glass Bottles"), a light diffusing plate for making a light source have a uniform light distribution. , And image the bottle side with a camera,
The acquired image is divided into many small areas by the image processing apparatus, and the luminance of each area and the R (red), G (green), and B (blue) of the light diffusion plate are determined.
In some cases, a small area to be subjected to color determination is selected based on the difference from the component, and the color is determined from the R, G, and B components of the same area.

However, the method described in the above-mentioned patent publication is a technique limited to bottle color identification, and has the following problems.

That is, the bottle colors are mainly transparent, brown, blue, green, and black, and each has a smoked bottle. In the automatic color sorting method, (1) when the bottle color is measured, Judgment methods such as counting the number of pixels within a specific hue range of the color measurement area and judging the bottle color, or dividing the image into small areas, determining the color measurement area and measuring the color of that area are performed. The removal of smoked bottles was not considered. (2) A method in which color determination is performed in an arbitrary small area is adopted, and this method also does not consider removal of smoke bottles.

It is an object of the present invention to provide an apparatus which can automatically sort smoke bottles from transmission bottles.

[0008]

SUMMARY OF THE INVENTION The present invention provides a light source for projecting light to a bottle, a light receiving device for receiving light transmitted through the bottle from the light source, and a bottle identification processing device for identifying the bottle based on a signal from the light receiving device. A bottle identification processing device comprising: a calculating means for calculating the light reception intensity and the light reception distribution of the light receiving device; and a transmission bottle that transmits light from the light reception intensity and the light reception distribution, such as a ground glass. And a judging means for judging a smoke bottle that scatters light to the glass bottle.

In the above-mentioned glass bottle identifying apparatus, the calculating means has a function of processing the difference of the received light intensity, a function of detecting a difference peak value from the difference value, and calculating an absolute value of the peak value;
And a function of calculating the distance between the difference peak values, wherein the determining means compares the absolute value of the difference peak value and the distance between the difference peak values with a preset threshold value, and It may be provided with a function for identifying bottles.

The present invention also provides a surface light source having a uniform light distribution for projecting light onto a bottle, an image input device for capturing a transmitted image from the light source via the bottle, and a bottle based on a video signal from the image input device. A glass bottle identification device comprising: an image processing device for identifying a surface light source, wherein the image processing device includes an identification bar for a surface light source, a function of extracting the identification bar of the surface light source, and calculating an area of the identification bar portion. , A function for calculating a histogram indicating the relationship between brightness and the number of pixels, a function for calculating a bottle image area, and discriminating a transmission bottle and a smoke bottle based on the size of the bottle image area in the identification bar extraction image. It is a glass bottle identification device having a function.

In the above-described glass bottle identification apparatus of the present invention, the image processing apparatus has a region-of-interest setting means for selecting a bottle image range, and has a function of extracting an identification bar portion of a surface light source in the region of interest. It may have a function of calculating the area of the identification bar portion in the region, a function of calculating the histogram in the region of interest, and a function of calculating the area of the bottle image in the region of interest.

[0012]

Embodiments of the present invention will be described with reference to the drawings. Hereinafter, embodiments of the present invention will be specifically described with reference to the drawings. FIG. 1 shows the overall configuration of a bottle sorting apparatus according to an embodiment of the present invention. In FIG. 1, a light source 3 having a good linearity for projecting light to a bottle is provided in a dark box 1.
And a plurality of light-receiving sensors 5 arranged in series with the transport path 2 for receiving the transmitted light from the light source 3 via the bottle, and the transport path 2 for transporting the bottles one by one is a light source in the dark box 1. 3
And the light receiving sensor 5.

A bottle detecting sensor 4 for detecting that a bottle being transported on the transport path 2 has arrived at a predetermined position, and a processing device 6 for identifying the bottle based on a light receiving signal from the light receiving sensor 5, Sorting damper 7 that sorts waste bottles into transparent bottles / smoke bottles, collection box 8 for each bottle
a and 8b are provided adjacent to the transport path 2 respectively.

FIG. 2 is a flowchart for sorting bottles. In this flowchart, it is detected whether or not the bottle has arrived at a predetermined position in the dark box 1 which can be processed. When the arrival of the bottle is detected, the light transmitted through the bottle is received by the light receiving sensor 5 and the processing shown in FIG. As a process of the device 6, a calculation process of calculating a received light intensity distribution based on a received light intensity signal of each light receiving sensor 5 and position data of each light receiving sensor 5 known in advance is performed, and a transmission bottle and a smoke bottle are obtained based on the result of the calculation process. And a bottle sorting process for opening and closing a sorting damper (not shown) based on the result of the above-described determination process.

In this embodiment, there is an alignment device (not shown) in the transport path 2 on the upstream side of the bottle identification device, which aligns the bottles in the longitudinal direction and supplies the bottles one by one. When a bottle is conveyed on the conveyance path 2 and passes through a bottle detection sensor 4 disposed on the conveyance path 2 upstream of the light receiving sensor 5 to detect the bottle, the light reception sensor is set at a predetermined position. 5 outputs a bottle arrival signal to the processing device 6 for receiving the transmitted light of the bottle.

The processing device 6 receiving the bottle arrival signal receives light transmitted through the bottle by the light receiving sensor 5. Regarding the light receiving signal, in the case of a transmission bottle, the light source is light (laser) with good straightness and little reflection, so that the transmitted light passes with almost the same intensity as shown in FIG. The light is received by the sensor 5. The processing device 6 obtains the intensity distribution of the transmitted light as shown in FIG.

In the case of a smoke bottle, the light source is reflected on the surface of the bottle and the amount of transmitted light is reduced, and is received by the light receiving sensor 5 as shown in FIG. FIG. 3 (a), FIG.
As shown in (b), in the case of a transmission bottle, only the position of the light source projecting portion has a large light receiving intensity in the case of a transmission bottle, and in the case of a smoke bottle, the reflection of light is large. The light at the position also becomes darker and becomes blurred as a whole. Therefore, the peak value of the received light intensity distribution becomes smaller and the distribution becomes larger.

Next, a method for determining the type of bottle will be described with reference to FIG. In the light reception intensity distribution of the transmission bottle, the light reception peak sharply rises, and the peak value itself is at a high level. On the other hand, in the smoke bottle, the light reception peak becomes gentle and the peak value also becomes a low level. Table 1 summarizes the above characteristics.

[0019]

[Table 1] ================================= Bottle Peak Value Light Receiving Distribution −−−−−−− −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− −−−−−−− ------------------------ ====================

The following processing is performed to facilitate identification by the processing apparatus of the present invention. (1) Perform difference processing of received light intensity. (2) Check the absolute value of the difference peak value. (3) Look at the distance between the difference peak values.

From the result of the above process (2), the state of the peak value high / low and the rise of the peak value can be grasped. From the result of the above process (3), the state of the light reception distribution can be grasped.

First, the processing of the above-mentioned processing (1) and (2) will be described. If the absolute value of the difference between the light receiving signals of the adjacent light receiving sensors 5 is taken, two peak values of the difference can be obtained.

That is, when a plurality of outputs of the light receiving sensors are arranged (correction is performed so that there is no output deviation at this time), one peak value is obtained in the received light intensity distribution. Next, the difference ΔP (equivalent to the differential value of the signal) of the received light intensity output of the adjacent light receiving sensor 5, for example, ΔP _i = 5 _i −5 _i−1 is taken (see FIGS. 4C and 4G). 4) When the absolute value (| ΔP |) is taken, the peak value is as shown in FIG.
There are two as shown in (h).

Here, a transmission bottle whose peak value of the received light intensity is Pf will be considered. The normalization is performed by dividing the absolute value | △ P | of the peak value of the differential received light intensity of the transmission bottle by the peak value Pf of the received light intensity. d △ Pf = i △ Pf | / Pf

The same normalization as that for the transmission bottle is performed for the smoke bin having the peak received light intensity Ps. d △ Ps = i △ Ps | / Ps After the above calculation, as shown in FIG. 5, for example, the transmitted light of the transmission bottle is received by the sensor 5. At that time, the adjacent sensor 5
Assuming that four light beams are received, | △ Pf | = １ ／ · Pf d △ Pf = (１ ／ · Pf) /Pf=１／=0.5

When the transmitted light of the smoked bottle is received by the sensor 5, the transmitted light is scattered from the transmitted bottle. Therefore, assuming that the light is received by six adjacent sensors 5, | ５Ps | = 1/3 · Ps d △ Ps = (1/3 · Ps) /Ps=1/3=0.33
...

D △ Pf ≧ 0.50 (ie, when transmitted light is received by up to four adjacent sensors 5) d △ Ps <0.34 (ie, transmitted light is received by six or more adjacent sensors 5) Here, the coefficient α (0 <α <1) is set to, for example, 0.4 (d △ P
When s <α <d △ Pf) and d △ Pf> α, the bottle may be transparent, and d △ Ps <α
At that time, the bottle may be a smoked bottle. Therefore, in this case, the light receiving sensor 5 that has passed through the bottle
When the peak value of P is P, the coefficient α = 0.
4 and whether or not the normalized d △ P is larger or smaller than α can be determined as a transmission bottle or a smoke bottle.
ΔPf = 0.5 indicates a transmission bottle, and d △ Ps = 0.3 indicates a smoke bottle.

Next, the process (3) will be described. The left side of the peak value of the differential received light intensity is the first peak value, and the right side is the second peak value. If the distance d between the two peaks is long, the light receiving distribution can be regarded as wide, and in that case, it is a smoke bottle. If d is short, the light reception distribution can be regarded as narrow, and in that case, it can be considered that the light is transmitted.

Here, a distance threshold value dt is set based on the size of the bottle and the resolution of the light receiving sensor 5. Here, the threshold value dt is a threshold value for determining whether there is a transmission bottle or a smoke bottle based on the distance between two peak values of the absolute value | △ P | , And the threshold value dt is a value that is a boundary between the distance between the two peak values of the absolute value of the transmission bottle and the distance d between the two peak values of the absolute value of the smoke bottle.

As described above, according to the result of the processing (2),
The state of the peak value high / low and the rise of the peak value can be grasped, and from the result of the processing (3), the state of the light reception distribution can be determined, and the coefficient α (0 <α <1) and the peak value of the difference light reception intensity When the distance threshold value dt is determined, the transmission bottle and the smoke bottle can be distinguished as shown in Table 2.

[0031]

[Table 2] ================================ │ΔP│ ≧ α ・ P │ΔP│ <α ・ P −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− d ≦ dt Transmission bottle − * ¹⁾ d> dt − * ²⁾ Smoke bottle =============================== * 1) Depending on the resolution of the light receiving sensor (number within a certain range), transmission Can be distinguished from bottles * 2) Can be distinguished from smoked bottles depending on the resolution of the light-receiving sensor. Bottles whose type is determined in this way are sorted by the sorting damper 7 and collected by the collection box 8.

In the above-described embodiment, bottle identification is performed by a combination of a laser, which is light having good linearity, as a light source, and a light-receiving sensor that receives the transmitted light of the bottle. FIG. 6 shows a method of using a color camera (hereinafter referred to as an image pickup unit) that includes a plurality of black identification bars in such a surface light source and that receives light transmitted through a bottle from the surface light source as the light receiving unit 5. It will be described using FIG.

FIG. 6 shows the overall configuration of this embodiment.
6, in a dark box 1, a surface light source 13 for projecting light to a bottle, and a plurality of light receiving cameras 1 arranged in series on a transport path 12 for receiving transmitted light from the light source 13 via the bottle.
5, and a transport path 12 for transporting the bottles one by one is disposed so as to pass through between the light source 13 and the light receiving sensor 15 in the dark box 11. A bottle detection sensor 14 for detecting that a bottle being conveyed on the conveyance path 12 has reached a predetermined position, a processing device 16 for identifying the bottle based on a light reception signal from the light reception sensor 14, and a bottle identified by the processing device 16 And a collection box 18a, 18b for each bottle are provided adjacent to the transport path 2, respectively.

In the case of a transmission bottle as shown in FIG. 7A, since the light from the light source is not affected by scattering / interference, it can be identified as a bar shadow in the imaging section. As shown in FIG. 7B, in the case of a smoke bottle, the light from the light source has a large scattering / interference, so that the amount of light transmitted through the bottle is small and the light becomes blurred. become unable.

The bottle sorting method of the present embodiment utilizing the above effects will be described with reference to the drawings. First, a description will be given of preprocessing for determining whether a transmission bottle is a smoke bottle. FIG. 8A shows, as a mask image, a binary image in which the identification bar portion of the surface light source image having the identification bar is “1” (bright portion) and the light emitting portion is “0” (dark portion). It was created. Then, the input image of each of the transmission bottle and the smoke bottle is subjected to AND operation of the images (FIG. 8).
(B) and FIG. 8 (c)).

In the case of a transparent bottle, FIG.
As shown in FIG. 8B, a dark image (an image in which only the identification bar is extracted) is obtained. In the case of a smoke bottle, theoretically, FIG.
As shown in (c), the portion of the smoke bottle on the identification bar is extracted.

However, in order to determine a transparent bottle and a smoke bottle, it is necessary to pay attention only to the bottle area on the identification bar, and it is necessary to delete unnecessary image information. Therefore, the processed image (identification bar extracted image: FIG. 8B, FIG.
For (c)), a region of interest (rectangular frame) is set.
The procedure for setting the region of interest is shown below. Here, the region of interest refers to an image range in which only necessary parts are selected, not all input images when performing image processing.

An input image as shown in FIG. 9A is obtained from the imaging unit. Next, a histogram indicating the relationship between the brightness and the number of pixels as shown in FIG. 9B is calculated from the input image. Mountains with low brightness on the horizontal axis of the histogram (groups close to 0) can be determined to be shadows of identification bars and bottles. Also, the peak with the brightest brightness (group near 255) can be determined as light directly received from the surface light source and is the background of the bottle. Then, it can be determined that the group in the center of the horizontal axis is bottled. From this determination, the bottle portion in the histogram can be extracted.

Here, binarization is performed between a portion determined to be a bottle and a portion not determined to be a bottle. The image is then redrawn based on this, and the horizontal (X-axis) end points (X ₀ , X ₁ ) and the vertical (Y-axis) end points (Y
₀ , Y ₁ ) to determine a rectangular frame connecting them. In FIG. 8C, a rectangular frame surrounding the bottle is set as a region of interest for performing image processing.

The above-mentioned region of interest is set in the identification bar extracted images (FIGS. 8 (b) and 8 (c)) processed earlier. As a result, unnecessary image information can be deleted, and the processing time can be reduced.

Next, the image in the region of interest will be described. Here, the method for determining the transmission bottle and the smoke bottle can be determined based on the amount (area) of image information having brightness that can be regarded as a bottle in the region of interest.

That is, as shown in the histogram of FIG. 9B, the range between lightness la and lightness lb is bottle image information, so how many images having lightness la to lb exist in the identification bar extracted image. This makes it possible to determine whether the transmission bottle is a smoke bottle or a smoke bottle.

As a specific procedure, (1) the area AR of the region of interest is calculated. (2) The mask area AM (range of brightness “0”) of the mask image in the region of interest is deleted because it is invalid image information. This is based on the area AR of the region of interest and the area AM of the mask part.
You can just pull The area in the region of interest remaining after this calculation is defined as the effective image area AR ′. The effective image area AR ′ is the area of the bar portion in the region of interest. When this relationship is expressed by an equation, AR-AM = AR ′. (3) Lightness la to lb which is bottle information on the histogram
If the area between them (the number of pixels) is AB, the effective image area A
The transmission bottle and the smoke bottle can be distinguished by the ratio R of how much the pixel (area) of the inner bottle of R 'occupies.

When this relationship is expressed by an equation, R = AB / AR '. At this time, the transmission bottle area is set to Abf, and in principle, Abf ≒ 0. However, since the actual bottle has a curved portion, the transmitted light is refracted and the background portion (illumination portion) is placed on the identification bar. As shown in FIG. 10A, a slight white portion where the transmission bottle area Abf ≠ 0 is not generated even outside the region of interest as shown in FIG.

Assuming that the area of the smoke bottle is Abs shown in FIG. 10B, the relationship of Abf <Abs holds. Therefore, the identification condition is R ≧ C. Here, C is a smoke bottle identification threshold: Abf
/ AR '<C <Abs / AR').

Thus, according to the embodiment of the present invention,
With a simple configuration, it is possible to accurately identify transmission bottles / smoke bottles. Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and various modifications can be made based on the technical idea of the present invention.

[0047]

According to the present invention, it is possible to identify transmission bottles / smoke bottles accurately with a simple configuration, and (1) the burden on the operator is greatly reduced. (2) Long-term operation and ensuring sorting accuracy. (3) Realization of high-precision sorting of transparent bottles by series connection with bottle color identification device. There is such an effect.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram of a bottle identification method using straight-ahead light (laser) according to an embodiment of the present invention.

FIG. 2 is a flowchart of bottle identification of the embodiment shown in FIG.

FIG. 3 is a diagram showing the principle of bottle identification using the straight traveling light (laser) of the embodiment shown in FIG.

FIG. 4 is a view for explaining the principle of judging a bottle based on a light receiving signal of the embodiment shown in FIG. 1;

FIG. 5 is a diagram for explaining a method of determining a reference value for determining the type of the bottle from the degree of spread of transmitted light of the bottle according to the embodiment shown in FIG. 1;

FIG. 6 is an overall configuration diagram of a bottle identification method using a surface light source according to an embodiment of the present invention.

FIG. 7 is a diagram showing the principle of bottle identification using the surface light source according to the embodiment of the present invention.

FIG. 8 is a diagram showing the principle of bottle identification using the surface light source according to the embodiment of the present invention.

FIG. 9 is a diagram showing the principle of bottle identification using the surface light source according to the embodiment of the present invention.

FIG. 10 is a diagram showing the principle of bottle identification using the surface light source according to the embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1, 11 Dark box 2, 12 Conveying path 3, 13, Light source with good straightness 4, 14, Detection sensor 5, 15, Light receiving sensor 6, 16, Processing device 7, 17, Sorting damper 8, 18, Collection box

Continuation of the front page (72) Inventor Katsumi Konaka 1-2-10 Isogo, Isogo-ku, Yokohama-shi, Kanagawa Prefecture F-term in Yokohama Engineering Center Babcock Hitachi, Ltd. (Reference) 2G051 AA14 AB20 BA10 CA03 CA04 CB02 DA06 DA11 EA16 EB01 ED01 ED07 GC04 GC17 GD02 2G059 AA05 BB08 DD12 EE01 GG01 HH02 KK04 MM02 MM05 PP01 3F079 AD12 CA31 CB25 CB29 CB32 CB34 CC06 DA11

Claims (4)

[Claims] 1. A glass bottle identification comprising: a light source for projecting light onto a bottle; a light receiving device for receiving light transmitted through the bottle from the light source; and a bottle identification processing device for identifying the bottle based on a signal from the light receiving device. An apparatus, wherein the bottle identification processing device is configured to calculate a light receiving intensity and a light receiving distribution of the light receiving device, and determine a transmission bottle transmitting light and a smoke bottle scattering light from the light receiving intensity and the light receiving distribution. And a means for identifying a glass bottle. 2. The calculating means has a function of processing a difference in received light intensity.
A function of detecting a difference peak value from the difference value and calculating an absolute value of the peak value; and a function of calculating a distance between the difference peak values. 2. The glass bottle identification device according to claim 1, further comprising a function of comparing the distance between the peak values with a preset threshold value, and a function of identifying the bottle based on the comparison result. 3. A surface light source having a uniform light distribution projected on a bottle, an image input device for capturing a transmitted image from the light source through the bottle, and an image for identifying the bottle based on a video signal from the image input device. A glass bottle identification device including a processing device, wherein the image processing device includes an identification bar in the surface light source, a function of extracting the identification bar of the surface light source, and a function of calculating the area of the identification bar portion, It has a function of calculating a histogram indicating the relationship between brightness and the number of pixels, a function of calculating a bottle image area, and a function of distinguishing a transmitted bottle and a smoke bottle by the size of the bottle image area in the identification bar extraction image. A glass bottle identification device characterized by the above-mentioned. 4. The image processing apparatus has a region of interest setting means for selecting a bottle image range, a function of extracting an identification bar portion of a surface light source in the region of interest, and an area of the identification bar portion in the region of interest. 4. The glass bottle identification apparatus according to claim 3, further comprising a function of calculating a histogram in the region of interest and a function of calculating a bottle image area in the region of interest.