JP3318809B2 - Bottle sorting 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 bottle sorting method for sorting, by color, bottles made of glass or translucent resin or the like for accommodating beverages and the like.

[0002]

2. Description of the Related Art In general, from the viewpoint of effective use of resources and the like, separation and recovery of glass bottles and reuse thereof are actively performed.
The collection form of glass bottles from each household differs slightly depending on each local government, such as collection by dedicated color boxes and collective bag collection, but in many cases, collection bottles are classified by color at the local government's resource center. Then, it is handed over to a collection company as cullet and recycled as raw material for glass. In many cases, the color separation of collection bottles is performed manually. However, in recent years, the need for automation has been increasing due to the need to improve working conditions and to improve the efficiency of separation and color identification.
From such a tendency, a glass bottle sorting device using a CCD camera or the like has been developed.

FIG. 15 shows a first conventional example of a glass bottle sorting apparatus (see JP-A-6-142618). In the first conventional example, a roller conveyor 11 for continuously transporting a glass bottle G, a television camera 12 for photographing the glass bottle on the roller conveyor 11, and a television camera 12 below the roller conveyor 11 are provided to face the television camera 12. A lighting 13, a color selection device 14 for determining the color of the glass bottle G based on an image signal from the television camera 12,
A monitor 15 for displaying the glass bottle G whose color has been determined constitutes a glass bottle sorting device.

[0004] The color selection device 14 detects the outer shape of the glass bottle G, and sets an area setting unit for setting a plurality of detection areas to be a color discrimination point in the glass bottle G, and detects and determines the color of the detection area. It comprises a discriminating unit and a storage unit in which the reference outer shape of the glass bottle G and reference color information including a plurality of colors and colors to be colored on the glass bottle G are stored.

In operation, when the glass bottle G transported by the roller conveyor 11 is photographed by the television camera 12, an image signal is output from the television camera 12 to the color selection device 14. Then, based on the image signal from the color selection device 14, the area setting unit of the color selection device 14 detects the external shape of the glass bottle G, extracts the reference external shape of the glass bottle G from the storage unit, and detects the external shape of the detected glass bottle G, The size ratio is determined by comparing the reference outer shape drawn from the storage unit, and a plurality of detection areas are set in the glass bottle G in proportion based on the determined ratio. Further, the determining unit detects the color of the detection area set by the area setting unit, and if the color of at least one of the detection areas is detected, the detected color is stored in the storage unit. The reference color information is used to determine which color the color belongs to.

FIG. 16 is a schematic view showing a bottle sorting method according to a second conventional example (see Japanese Patent Application Laid-Open No. 6-142617). In the second conventional example, the longitudinal direction of the glass bottles A is aligned and aligned in the width direction of the conveyor belt of the conveyor by the aligning device 21, and the glass bottles A flattened by the aligning device 21 by the control device 23 are accurately aligned. Sorting device 22 one by one
Adjusted to be delivered to the glass bottle A sent from here
Were separated one by one by the shape and color by the separation device 22.

{Third Conventional Example} FIG. 17 is a schematic diagram showing a bottle sorting method according to a third conventional example (see Japanese Patent Application Laid-Open No. 6-31251). In the third conventional example, the glass bottle 31 put into the hopper 32 is sent by the vibration filter 33 to the vibration aligning machine 36 in an amount corresponding to the processing capacity. Vials fed to the vibrating aligner 36 are aligned in area 37, spaced in area 38, and passed through a pass sensor 42.
Go through. The image processing device 41 includes a passage sensor 42
The color of the glass bottle 31 is determined in synchronism with the signal of (a), and the bottle selection controller 43 controls the bottle selection device 44 based on the determination result. In addition, 40 is a camera, and 39 is a lighting device.

[0008]

In the first, second and third prior arts described above, it is easy to judge a glass bottle having a relatively low degree of transparency, but the color is almost transparent. Judgment of a bottle is difficult because the difference in data to be compared is small. In order to solve such a problem, the following fourth conventional bottle sorting method is disclosed.

<< Fourth Conventional Example >> FIG. 18 shows a bottle sorting method according to a fourth conventional example. In the fourth conventional example, the transmitted light of the glass bottle 3 is photographed by the CCD cameras 4a and 4b by the image processing high-frequency illumination 2 through the diffusion plate 1, and the R (red) is transmitted by the camera controller 5 and the image processing device 6. , G
(Green) and B (blue) components are used for determination. In this case, the determination of a colored bottle that is nearly transparent becomes more difficult as the brightness of the illumination increases. Therefore, not only one CCD camera (4a) but also another CCD camera (4b) with a reduced amount of illumination is provided to perform bottle color sorting.

However, in the case of the fourth conventional example, it is necessary to process images of two cameras, and there is a disadvantage that the cost of parts is high.

The present invention has been made in view of the above circumstances, and has as its object to provide an efficient bottle sorting method capable of sorting light-colored bottles that are nearly transparent, at a relatively low cost.

[0012]

Means for Solving the Problems According to a first aspect of the present invention, there is provided a method for sorting bottles made of glass or light-transmissive resin by color, comprising: The two-side end coordinate detecting step of detecting the coordinates, the center line coordinate calculating step of calculating the coordinates of the middle line of the coordinates of the two side end lines as the center line coordinates of the bottle, and the center line coordinate calculating step An imaging step of imaging the bottle at the center line coordinates, and a color determination step of processing the image captured in the imaging step to perform a color determination , the two-side end coordinate detection step,
The bottle is placed sideways by the placing means, and the bottle is placed behind the bottle.
While irradiating light from the illumination means arranged on the side,
The brightness detector placed on the front side is oriented in the direction of the center axis of the bottle.
Scans intermittently at predetermined intervals in the perpendicular vertical scanning direction.
The brightness detected by the brightness detector is equal to or greater than a certain brightness difference.
The coordinates in the vertical scanning direction when the
An upper end coordinate detecting step of detecting as coordinates, and the placing means described above.
The coordinates of the point of contact between the
A bottom coordinate detection step of detecting the bottom coordinate of the storage bottle.
No.

[0013]

According to a second aspect of the present invention, there is provided:
What is claimed is: 1. A method for sorting bottles made of glass or translucent resin by color, comprising: a saturation detecting step of detecting the saturation of the bottle by a saturation detecting means; and a brightness of the bottle by a brightness detecting means. A brightness / lightness ratio calculating step for calculating a ratio between the saturation detected in the saturation detecting step and the lightness detected in the brightness detecting step; and a chroma / brightness ratio calculating step. A certifying step of certifying that the bottle is a dark-colored bottle when the value of the ratio calculated in the above step is larger than a predetermined reference value, and certifying that the bottle is a light-colored bottle when the value of the ratio is smaller than the predetermined reference value. A dark-colored bottle sorting step of performing color sorting for a dark-colored bottle when it is determined to be a colored bottle, and a light-colored bottle sorting step of performing color sorting for a light-colored bottle when it is determined to be a light-colored bottle in the above-described certification step. And

[0015] The problem solving means according to claim 3 of the present invention is as follows.
The dark bottle selection step includes a first color component detection step of detecting a predetermined color component of the bottle, and a detection value of the color component detected in the color component detection step in a coordinate space for the dark bottle. A first map step of arranging, and a color identification of the detected value of the color component detected in the color component detection step in the coordinate space for the dark color bottle arranged in the first map step. The light-colored bottle sorting step, wherein the light-colored bottle sorting process comprises: converting the dark-colored bottle coordinate space into a light-colored bottle coordinate space so that the color component of the illumination means for irradiating the bottle coincides with the coordinate origin. A coordinate conversion step for converting, a second color component detection step for detecting a predetermined color component of the bottle, and a detection value of the color component detected in the color component detection step arranged in a coordinate space for the light-colored bottle. A second mapping step, and a coordinate space for the light-colored bottles arranged in the second mapping step. In, and a second color identification step of performing the color identification of the detected value of the color component detected by the color component detecting step.

[0016] The means for solving the problem according to claim 4 of the present invention is:
Sorting to sort bottles made of glass or translucent resin by color
A method for detecting coordinates of both end lines of the bottle, the method comprising:
End coordinate detection step, and coordinates of an intermediate line between the coordinates of the both end lines
Is calculated as a center line coordinate of the bottle.
And the center line calculated in the center line coordinate calculating step.
An imaging step of imaging the bottle at coordinates, and
A color determination step of performing color determination by processing the captured image.
The imaging step includes imaging a predetermined area including a plurality of detection points around the center line coordinates of the bottle, and the color determination step includes a step of detecting a predetermined area at the plurality of detection points in the predetermined area. A color component detection step of detecting each color component,
A calculating step of calculating a sum or an average value of the plurality of detection points for each of the predetermined color components detected in the color component detecting step.

According to a fifth aspect of the present invention, there is provided:
The predetermined area is set symmetrically about the center line of the bottle.

[0018] The problem solving means according to claim 6 of the present invention is as follows.
The calculating step is a first average value calculating step of calculating an average value of the predetermined color components of all the detection points in the predetermined area;
Among all the detection points in the predetermined area, a different value recognition step of recognizing a plurality of different value detection points indicating detection values relatively different from the average value calculated in the first average value calculation step, A second average value calculating step of calculating an average value of the predetermined color components of remaining detection points excluding the different value detection points from all the detection points in the predetermined area.

[0019] The problem solving means according to claim 7 of the present invention is:
The different value recognition step includes a step of selecting a predetermined number of detection points in order from the detection point having the largest value difference as compared with the average value and setting the selected number as the different value detection point.

The problem solving means according to claim 8 of the present invention is:
7. The bottle sorting method according to claim 6 , further comprising an area specifying step of specifying the predetermined area so as to avoid the adhering matter on the surface of the bottle.

According to a ninth aspect of the present invention, there is provided a means for solving the problem,
A sorting method for sorting bottles made of glass or translucent resin according to color, wherein positioning is performed so that a predetermined region in a captured image of an imaging device arranged on the front side of the bottle matches a bottom side surface of the bottle. A positioning step, an imaging step of irradiating light from an illuminating unit arranged behind the bottle, and imaging the bottom side surface of the bottle from the front side of the bottle with the imaging device, and the imaging step And a color determination step of performing a color determination by processing the image captured in step (a).

According to a tenth aspect of the present invention, the color determining step includes a color component detecting step of respectively detecting predetermined color components at a plurality of detection points in the predetermined area; A selection step of selecting, from among the predetermined color components detected in the step, a value having the largest difference from preset transparent data, or a fixed number of value groups in order from the value having the largest difference; A selection value determination step of performing color determination only on the value or value group selected in the selection step.

According to an eleventh aspect of the present invention, in the color determination step, the coordinates are determined such that an origin of a coordinate space used for color determination coincides with a color component of illumination means for irradiating the bottle. A coordinate correction step of correcting a space, a color component detection step of respectively detecting predetermined color components at a plurality of detection points in the predetermined region, and a detection value of the color component detected in the color component detection step, A mapping step of arranging in the coordinate space corrected in the coordinate correction step, an origin distance calculating step of calculating a distance from the origin of the detected value of each color component in the corrected coordinate space, and an origin distance calculation A selection step of selecting the largest value among the detected values of the color components calculated in the step, or a fixed number of value groups in order from the largest value, and a value or a value group selected in the selection step Color judgment only for And a selection value determining step for.

According to a twelfth aspect of the present invention, there is provided an illumination color detecting step for detecting a color component of the illuminating means, wherein the illuminating color detecting step detects a plurality of color components of the illuminating means. Detecting at each point, and calculating a sum or an average value of the color components for the plurality of detection points.

According to a thirteenth aspect of the present invention, there is provided a sorting method for sorting bottles made of glass or light-transmitting resin by color, and determining whether the bottles have a predetermined transparency. the transparency determiner step of running, when it is determined to have a predetermined transparency in the transparency determination step, 請
Step of positioning Motomeko 9, sequentially executes imaging process and the color determination process.

According to a fourteenth aspect of the present invention, in the transparency determining step, the coordinates are determined so that an origin of a coordinate space used for color determination coincides with a color component of the illuminating means for irradiating the bottle. A coordinate correction step of correcting a space, a color component detection step of detecting a predetermined color component of the bottle, and a coordinate space in which the detected value of the color component detected in the color component detection step is corrected in the coordinate correction step And a color identification step of performing color identification of the detection value of the color component detected in the color component detection step on the corrected coordinate space.

[0027]

In the bottle sorting method according to the first and fifth aspects of the present invention, since the color judgment is performed at the center line of the bottle, it is not affected by the radius of curvature of the bottle, and data can be acquired with little variation. .

[0028] In the bottle sorting method according to claim 1 of the present invention, after making the coordinate detection of the upper end of the bottle, it is not necessary to obtain in the same manner as the upper end coordinate of the lower end, thus and reliably very short The center line can be determined.

In general, glass bottles for storing beverages and the like include a dark bottle set to a deep color for the purpose of protecting the contents from external light and the like, and a light-colored bottle having a higher transparency to make the contents easier to see from the outside. It is roughly divided into two types. In view of this, in the bottle sorting method according to claim 2 of the present invention, it is determined whether the bottle is a dark bottle or a light bottle based on the ratio between the saturation and the lightness of the bottle, and the color is sorted by different sorting processes. Perform sorting. Then, especially when the transparency is high and color selection is difficult, accurate selection can be performed by increasing the selection accuracy. In particular, in the bottle sorting method according to the third aspect of the present invention, in the coordinate conversion step of the light-colored bottle sorting step, the coordinate origin is matched with the color component of the illuminating means. Make accurate decisions.

In the bottle sorting method according to the fourth aspect of the present invention, a predetermined color component is detected at a plurality of detection points within a predetermined area, and the bottle selection is performed based on a sum or an average value of the detected color components. Sex can be dramatically improved.

In the bottle sorting method according to the fifth aspect of the present invention, by setting the predetermined area vertically symmetrical with respect to the center line of the bottle, the predetermined area is relatively unaffected by the radius of curvature of the bottle, and data having little variation is provided. Acquisition can be performed.

In the bottle sorting method according to the sixth aspect of the present invention, after calculating the average value of the color components of all the detection points in the predetermined area, a detection value that is significantly different from the average value is recognized, and the remaining values excluding this are excluded. The average value of only the detection points is calculated and bottle sorting is performed. In this way, when a label or the like is affixed to the surface of the bottle, when there is a dirty portion, or when halation occurs partially, the bottle can be selected by removing such a portion. The reliability of the acquired data can be greatly improved. In particular, in the bottle sorting method according to the seventh aspect of the present invention, a predetermined number of detection points are selected in order from the detection point having the largest value difference as compared with the average value and are set as different value detection points, so that efficient data Can be reduced.

In the bottle sorting method according to the eighth aspect of the present invention, when there is an attached matter such as a label or dirt on the surface of the bottle, a predetermined region can be excluded from such a portion in advance. Thereby, it is possible to prevent the average value in the first average value calculation step from being far from the data of only the surface of the bottle, and it is possible to accurately recognize the different value in the different value recognition step. The calculation of the average value in the average value calculation step becomes accurate.

In the bottle sorting method according to the ninth aspect of the present invention, in particular, in the case of sorting bottles of light-colored bottles, the color judgment is performed on the light passing through the bottom side surface of the bottle, so that the direction of the transmitted light from the illuminating means can be minimized. The color can be determined for the thick part, and the bottle can be accurately selected as a colored bottle with high saturation.

In the bottle sorting method according to claim 10 of the present invention,
By detecting a predetermined color component for a plurality of detection points in a predetermined area and performing bin sorting based on the sum or average value thereof, the reliability of the acquired data can be significantly improved.

In the bottle sorting method according to claim 11 of the present invention,
In the coordinate correction step of the light-colored bottle selection step, since the coordinate origin is matched with the color component of the illuminating means, accurate judgment can be made even when the transparency is extremely high and it is difficult to judge the coloring. Also, the distance from the origin of the detected value of each color component in the corrected coordinate space is calculated, and a color group is selected by selecting the largest value or a fixed number of values in order from the largest value. Since it is performed, the difference with respect to the transparent color of the bottle can be accurately determined with a very simple calculation. Therefore, bottle sorting can be performed without requiring a great deal of processing time.

In the bottle sorting method according to claim 12 of the present invention,
The color components of the illumination means are detected at a plurality of detection points in the space outside the bottle, and the reliability of the color components of the illumination means is enhanced by the sum or average value, so that the color components are converted or corrected for a light-colored bottle. As a result, the origin of the coordinate space can be accurately matched with the color component of the illumination means, so that the color determination can be performed accurately.

In the bottle sorting method according to the thirteenth aspect of the present invention,
Only if the bottle has a predetermined transparency step of positioning according to claim 9, by performing the color selection by sequentially executing the imaging process and the color determination process, in particular accurately perform color sorting transparent bottle. In the case of other than a transparent bottle, color selection may be performed by a normal method that does not require positioning, and processing can be performed more quickly than in the case of a transparent bottle.

In the bottle sorting method according to claim 14 of the present invention,
When the color is determined in the transparency determining step according to claim 13,
Since the origin of the coordinate space is matched with the color component of the illumination means, the transparency can be accurately determined.

[0040]

【Example】

<Structure> FIG. 1 is a plan view showing the overall structure of a bottle sorting apparatus according to one embodiment of the present invention. In FIG. 1, reference numeral 51 denotes a glass bottle;
Is a conveyor (roller conveyor or belt conveyor: mounting means) for conveying the glass bottle 51; 53 is a color recognition unit for performing color recognition of the glass bottle 51; 54 is an illumination device (illumination) for illuminating the glass bottle 51 in the color recognition unit 53. Means, 55, a camera (imaging device) for imaging the glass bottle 51 illuminated by the illumination device 54, 56, an image processing unit for processing an image imaged by the camera 55, 57, a color recognized by the color recognition unit 53 Separating section for separately sorting glass bottles 51, 58 is sorting section 5
7, a black bottle collector for collecting black bottles, 59 a brown bottle collector for collecting brown bottles, 61 a dark green bottle collector for collecting dark green bottles, 62 a colorless bottle for collecting colorless bottles Collection device, 63 is a different color bottle collection device to collect bottles of other colors,
Reference numeral 64 denotes a control unit that controls the selection operation of the selection unit 57 based on the data processed by the image processing unit 56, and 65 denotes a bottle detection sensor that detects the presence or absence of the glass bottle 51.

Here, the glass bottle 51 handled in this embodiment is
Are generally used for accommodating beverages and the like, and have various characteristics such as size and color. Generally, the glass bottle 5
1 is broadly classified into two types: a dark-colored bottle set to a deep color for the purpose of protecting the contents from external light and the like, and a light-colored bottle whose transparency is enhanced to make the contents easily viewable from the outside. Here, in the case of a dark-colored bottle, since coloring such as green or brown with high saturation is applied and lightness is set low, if the saturation is C and the lightness is L, the former has a very high C / L. (“0.5” or more). It should be noted that C / L shows a very large value even for bottles that appear almost black at first glance to the naked eye because they are actually colored dark green. On the other hand, in the case of a light-colored bottle, the C / L is very small (ie, less than “0.1”). And the intermediate value, that is, C / L of “0.1” or more and less than “0.5” is a very small number. As will be described later, the bottle sorting apparatus according to the present embodiment is configured to perform accurate color sorting for both the dark bottle and the light bottle. The glass bottle 51 is placed on the transport conveyor 52 in a sideways state, and transported to the color recognition unit 53 and the sorting unit 57. Regarding the orientation of the glass bottle 51 on the conveyor 52, either the mouth side or the bottom side of the bottle may be positioned in the traveling direction.

As shown in FIG. 2, a plurality of (for example, four) illuminating devices 54 are erected and supported by a holder (not shown).
And a diffusion plate 72 made of glass or translucent resin for diffusing light from the white fluorescent lamp 71 into the glass bottle 51.

The camera 55 is a general CCD (ch).
The image capturing area of the camera 55 is set so large that most of the general glass bottles 51 shipped to the market can be included in the image.

The image processing unit 56 includes the black bottle collection unit 58, the brown bottle collection unit 59, the dark green bottle collection unit 6
1. The colorless bottle collecting device 62 and the different color bottle collecting device 6
3 to determine which of the glass bottles 51 to collect the glass bottle 51. As shown in FIG. 3, m predetermined areas in the imaging area imaged by the camera 55 are set as a window 74, and further as shown in FIG. , N ₁ × n ₂ in each window 74
The number of detection points P is set, image processing described below is performed on each of the detection points P, and the color of the glass bottle 51 is determined. The window 74 is, for example, 20 mm × 20 mm,
The detection point P is set to, for example, about 100 points. When the center line of the glass bottle 51 is obtained as a data acquisition line, the camera 55 and the image processing unit 56 function as a brightness detector that detects the brightness of light from the lighting device 54 passing through the glass bottle 51.

On the basis of a command signal from the control section 64, the sorting section 57 sorts the glass bottle 51 conveyed by the conveyor 52 by a handling device (not shown) into the black bottle collecting device 58 and the brown bottle collecting section. Container 59, the dark green bottle collection device 61, the colorless bottle collection device 62, and the different color bottle collection device 63.

The control unit 64 controls the black bottle collection unit 58 and the brown bottle collection unit 5 selected by the image processing unit 56.
9. A device for driving and controlling a handling device of the sorting unit 57 so as to collect the glass bottle 51 to any one of the dark green bottle collection device 61, the colorless bottle collection device 62, and the different color bottle collection device 63. It is.

<Operation> 1.) Overall Operation Flow The operation of the above-configured bottle sorting apparatus will be described. As shown in FIG.
First, in a state where the glass bottle 51 does not enter the imaging region of the camera 55, the illumination device 54 illustrated in FIG. 2 is imaged, and color component data of the illumination device 54 (hereinafter, referred to as illumination data).
( _RL , _GL , _BL ) are fetched (step S1). At this time, illumination data is repeatedly detected for a plurality of detection points (steps S2 and S3). Then, the image processing unit 5
At 6, the average (or sum) of the illumination data for the plurality of detection points P is calculated and stored.

Here, in the case of a small-sized glass bottle 51, the glass bottle 51 can be accommodated in the imaging region, but in the case of a large-sized glass bottle 51, it cannot be completely accommodated in the imaging region. is there. Therefore,
It is determined whether the state in which the glass bottle 51 is conveyed is that the mouth side is the top and the bottom side is the top (step S5), and in particular, the bottom position of the glass bottle 51 in the case of a transparent bottle described later is confirmed. Specifically, at a point in time when the glass bottle 51 that has been conveyed slightly enters the imaging region, luminance data for a plurality of points in the imaging region is input as temporary data (sensor data) (step S4). Then, from the sensor data, the coordinates of the point of the glass bottle 51 closest to the conveyor 52 in the vertical scanning direction (y direction) are detected, and the separation distance of the glass bottle 51 from the conveyor 52 is calculated. If the separation distance is larger than a preset threshold value, it is determined that the portion of the glass bottle 51 projected in the imaging region is on the mouth side (judgment of the bottle direction: step S5). Then, while being conveyed by the conveyer 52, the image is repeatedly taken in until the bottom of the glass bottle 51 enters the imaging area (field of view), and the luminance of all points in the vertical scanning direction (y direction) is changed to the illumination data. When it becomes equal to, the time required for transport from the mouth side to the bottom side is calculated (determination of required delay time: step S5). Thereafter, the user inputs bottle data as actual data for actually selecting the bottle (step S6). On the other hand, in step S5, if the distance of the glass bottle 51 from the transport conveyor 52 is smaller than a preset threshold value, it is determined that the portion of the glass bottle 51 projected on the imaging area is on the bottom side. Then, after the paper is conveyed by a predetermined size slightly, a bottle data as actual data for actually selecting a bottle is inputted (step S6). Then, after determining the range of valid data (step S7), the glass bottle 51 becomes black, brown,
It is recognized whether the color is dark green, colorless, or another color (step S8), and the result is output to the control unit 64 (step S9). Thereafter, the same processing is performed for the plurality of glass bottles 51 (step S10).

2.) Bottle Data Input Flow [Determining Data Acquisition Line] Next, the operation of step S4 shown in FIG. 5 will be described in detail. First, a data acquisition line for inputting bottle data is obtained (step S11 in FIG. 6). As a data acquisition line, there is a method of obtaining data at a fixed height from the mounting surface of the transport conveyor 52 as a base surface. However, from the viewpoint of image stability, the data line is most preferably on the center line of the glass bottle 51. Here, FIG. 7 shows an operation for obtaining the center line MM of the glass bottle 51. As shown in FIG. 1, the glass bottle 51 is placed on the transport conveyor 52 in a state of being laid down, and while being irradiated with light from an illuminating device 54 disposed behind the glass bottle 51,
When the image is scanned in a vertical scanning direction (y direction) orthogonal to the center line MM of the glass bottle 51 and in a descending direction as shown in FIG. The luminance detected by the processing unit 56 is Lg in FIG.
As shown in (1), when the scanning line hits the glass bottle 51, the luminance level is always greatly reduced once regardless of the type of the glass bottle 51. Utilizing this phenomenon, the y coordinate U when the luminance level is reduced to a certain luminance difference (dR in FIG. 7) or more.
Is detected as the upper end coordinates of the glass bottle 51 (upper end coordinate detecting step).

More specifically, an image is taken at a minute interval Δy from y = 0, and the magnitude of one of the color components (R, G, B) or their average value is examined, (For example, R _i , R for the red component R)
_{i + 1} ) When | R _{i + 1} −R _i |> dR, it is assumed that R _{i + 1} is over the image of the glass bottle 51. Here, i is the number of intermittent step movements when the detection point is intermittently scanned in the imaging area of the camera 55 at a minute interval Δy from y = 0. by this,
y coordinate y when the scanning line in the y direction hits the glass bottle 51
The minimum value y _jmin of _j is obtained as the value U.
In addition, y of the contact point between the conveyor 52 and the glass bottle 51
The coordinates y _B = B are detected as the lower end coordinates of the glass bottle 51 (lower end coordinate detection step). Then, in the image processing unit 56, the coordinates ｛(y _jmin + y
_B ) / 2} = {(U + B) / 2} = D is calculated as the center line coordinates of the glass bottle 51 (center line coordinate calculation step). Such a center line coordinate is, as shown in FIG.
By determining the points of the plurality of x direction for each accurately determine the position of the center line M-M as the data acquisition line D _L. Thus, by acquiring data at the center line of the glass bottle 51, data acquisition with less variation can be performed.

The image data input in this process is
After the glass bottle 51 is imaged only once, all the saturation and lightness data are stored in a storage unit having a predetermined storage capacity, and data based on a chromaticity method described later is stored for later sorting. It is desirable to use it for processing.

[Determination of Shading of Bottle] As described above, the glass bottle 51 is roughly classified into two types, a dark-colored bottle and a light-colored bottle. And
In the case of a dark bottle, it is easy to determine the color, but in the case of a light bottle, it is erroneously recognized that an error occurs in the acquired data. Therefore, first, the density of the bottle is determined (step S in FIG. 6).
12) Color discrimination is performed by a different processing method between a dark bottle and a light bottle.

When the light from the illuminating device 54 transmitted through the glass bottle 51 is picked up by the camera 55 and image-processed by the image processing unit 56, if the saturation of a predetermined point in the image is C and the lightness is L, In the case of a color bottle, the C / L becomes very large (“0.5” or more), and in the case of a light-colored bottle, the C / L becomes very small (ie, less than “0.1”). Therefore, if (C / L) ² ≦ 0.008, the bottle is recognized as a light-colored bottle, and if (C / L) ² > 0.008, the bottle is recognized as a dark-colored bottle.

Here, assuming that the red component is R, the blue component is B, and the green component is G, the yellow component Y = (0.3R + 0.5
9G + 0.11B), and from these values, the above C and L can be calculated by the following equations (1) and (2).

[0055]

(Equation 1)

[0056]

(Equation 2)

It should be noted that the data obtained here is also used in the color discrimination processing for dark color bottles and the color discrimination processing for light color bottles described later. Dark bottle color components (R, G, B)
Has a tendency to be clearly separated in any direction from the origin of the color coordinate space, so a highly reliable judgment can be made with a small number of data. However, in the case of a light-colored bottle, the color components (R,
(G, B) Since the judgment must be made while paying attention to a slight difference between the data, it is difficult to make a highly reliable judgment with a small number of data. Therefore, as shown in FIG. 3, the window 74 of rectangular or square formed into m position on the data acquisition line (center line M-M), further as shown in FIG. 4, the detection point of the n ₁ × n ₂ of which The sum or average value of the data for P is processed as data for each of the m points. The data at m points may be averaged and determined as a single value, or the determination may be performed for each of the m points and finally determined by majority decision. By doing so, the variation in data is reduced, and a more reliable determination can be made.

The setting position of the window 74 is set so that the center line of the window 74 coincides with the center line MM of the glass bottle 51 as the above-mentioned data acquisition line.
That is, the window 74 is set vertically symmetrically about the center line MM.

Also, as shown in FIG. 8, a label 76 is attached to the surface of the window 74, there is a stained portion (not shown), or halation is partially generated (not shown). In such a case, if the sum of all the data in the window 74 is simply calculated and determined, this may cause an erroneous determination. In order to prevent such a situation, it is desirable to take a sum or an average of data having similar sizes in the window 74. In this embodiment, the window 74
In the case of averaging the data for n ₁ × n ₂ detection points P, data having a large distance from the average is excluded in order to remove noise components.

More specifically, first, the average value of each color component (R, G, B) of all the detection points P in the window 74 is calculated (first average value calculation step). Among the points P, a plurality of different value detection points indicating detection values relatively different from the average value calculated in the first average value calculation step are recognized (different value recognition step). In the different value recognition step, a predetermined number (n) is sequentially set in order from the detection point having the largest value difference as compared with the average value.
_s ) Select the detection point and use it as the outlier detection point. Thereafter, the average value of the color components (R, G, B) of the remaining detection points excluding the different value detection points from all the detection points P in the window 74 is calculated (second average value calculation step).

As shown in FIG. 8, a label 76 may be attached to the surface of the glass bottle 51. Usually, since the position of the label 76 in the glass bottle 51 is known in advance by the product, the setting position is specified so that the setting of the window 74 is basically excluded from the position of the label 76. Specifically, the window 74 may not be set in a section between the x coordinates x _L1 and x _L2 at both ends of the label 76.

Thus, even if the data at the position where the label 76 is applied is mixed in the actually measured data, the mixed data is exceptional data based on the ratio of the number of mixed data to the number of normal data. We can recognize that there is.

[Determination of Color of Dark Color Bottle] In general, the determination result of the chromaticity method for dark color bottles often matches the color perception by the naked eye. Therefore, if it is determined in step S12 that the bottle is a dark-colored bottle, color discrimination is performed by a general chromaticity method, which is a method that relatively matches human color vision (step S13 in FIG. 6). The chromaticity method is a well-known color conversion method generally known as a method for recognizing coloring of a substance,
As shown in FIG. 9, the value (B−Y) obtained by subtracting the yellow component (Y) from the blue component (B) is taken along the horizontal axis (x), and the vertical axis (y) is taken.
Assuming color component coordinates obtained by subtracting the yellow component (Y) from the red component (R) in the direction (RY), "brown", "black", "green", " The area is divided into "blue" and "other". Specifically, Y, RY, and BY are calculated as in the following Expression 3.

[0064]

(Equation 3)

In FIG. 9, a boundary line of each area serving as a reference for color discrimination is set as a straight line having an angle H with the x-axis of the color component coordinates being 0 °. Here, the angle H can be expressed as in the following Expression 4.

[0066]

(Equation 4)

That is, the angle H = 14 ° to 90 ° is “other”, the angle H = 90 ° to 160 ° is “brown”, the angle H = 160 ° to 180 ° is “black”,
Angle H = 180 ° to 270 ° is “green”, angle H =
"Other" from 270 ° to 300 °, angle H = 30
From 0 ° to 360 ° (= 0 °) and from 0 ° to 14 ° are recognized as “blue”.

In obtaining the data of each component, the data obtained in the process of determining the density of the bottle may be used as it is. In this case, since there are a plurality of acquired data, an average value is obtained.

[Color Discrimination of Light-Colored Bottle] On the other hand, regarding the color discrimination of the light-colored bottle, since the illumination used has a large effect on the coloring of light rays, the discrimination result by the chromaticity method described above does not match the color perception by the naked eye. There are many things.

A black circle L _S shown in the color component coordinates of the chromaticity method shown in FIG. 9 is obtained by mapping chromaticity data (hereinafter, referred to as illumination data) of direct light from illumination by the chromaticity method. It is. Originally, it is desirable that the illumination data coincides with the origin O. However, since the actual illumination is not an achromatic color but has a certain degree of saturation, it is actually a theoretical achromatic state. Is shifted in the upper right direction in FIG. In such a state, for example, when the chromaticity of a light-colored bottle having extremely high transparency is recognized, the data is shifted to the upper right due to the effect of the coloring of the illumination, and a situation occurs in which the chromaticity recognition is incorrect or unrecognizable. Therefore, in the present embodiment, when it is determined in step S12 that the bottle is a light-colored bottle, the origin is corrected and the magnitude of the value is normalized by assuming a coordinate space for the light-colored bottle different from that for the dark-colored bottle. (LND method).

First, let the bottle data be (R, G, B), the actual measured values of the illumination data taken in step S1 in FIG. 5 be ( _RL , _GL , _BL ), and n be the data acquisition number. ,

[0072]

(Equation 5)

Assuming a variable L _n , the red component (r _n ), the blue component (b _n ) and the green component (g _n ) are calculated using the L _n as shown in equations (6) to (8). I do.

[0074]

(Equation 6)

[0075]

(Equation 7)

[0076]

(Equation 8)

Then, the x _n calculated by the equation 9 is plotted on the horizontal (x) axis.
The vertical (y) taking y _n calculated in several 10 to the shaft, to map the data.

[0078]

(Equation 9)

[0079]

(Equation 10)

FIG. 10 is a coordinate space for a light-colored bottle for mapping the data calculated as described above. The coordinates are such that the measured values (R _L , G _L , B _L ) of the illumination data coincide with the origin O. Has been converted. Then, it is determined which region is included in each of the colors shown in FIG. 10, and the color of the glass bottle 51 is determined (step S14 in FIG. 6).

As described above, even if illumination having a slight saturation is used, erroneous recognition due to the illumination can be prevented. In addition, since the illumination data at each coordinate position measured immediately before the actual bottle sorting operation is used, and the coordinate conversion (coordinate correction) is performed using the measured values, the installation location of the bottle sorting device may be moved or the room may be moved. Even when the lighting environment is changed, the coordinate correction can be performed under the condition after the change of the environment, and the influence of the change of the environment on the data can be reduced.

The color components (R, G,
B) has a tendency to be clearly separated from the origin in any direction, so that highly reliable judgment is possible with a small number of data. However, in the case of a light-colored bottle, only a small amount of the color component (R, G, B) data is used. It is necessary to make a determination by paying attention to the difference between the two, and it is difficult to make a highly reliable determination with a small number of data. Therefore, as shown in FIG. 3, the window 74 of rectangular or square formed into m position on the data acquisition line (center line M-M), as further in FIG. 4, n ₁ of which
The sum or average value of xn ₂ data is processed as data of each of the m points. The data of m points may be averaged to determine a single value, which is very good, or the determination may be performed for each of the m points, and the final decision may be made by majority decision. By doing so, the variation in data is reduced, and a more reliable determination can be made.

In obtaining the data of each component, the data obtained in the process of determining the density of the bottle may be used as it is. In this case, since there are a plurality of acquired data, an average value is obtained.

[Processing for a Transparent Bottle] The glass bottle 51 determined to be a transparent bottle in the above-described color determination of a light-colored bottle (step S14) is again determined, and a colored bottle that has been erroneously recognized as transparent is accurately identified. They are going to be sorted out.
Specifically, in step S15 in FIG. 6, it is recognized that the container is a transparent bottle.
1 is viewed from the side, the thickest portion in the direction of the transmitted light from the illumination device 54 is viewed to determine the color (BT).
Law).

As shown in FIG. 11, the window BTW (for example, 10 mm × 40 mm and about 100 detection points P) is moved back and forth and up and down near the bottom BT of the glass bottle 51, and each detection point P Calculations up to 10 are performed, and (x _n , y _n ) is calculated in the same coordinate space (FIG. 10) as in the above-described light-colored bottle color determination processing. And FIG.
Illumination data ( _RL , _GL ,
One point having the largest distance from the point corresponding to B _L ) is selected, and data obtained at this point is used as bottle data. Here, the illumination data captured in step S1 in FIG. _{5 (R L, G L,} B L) so the point converted in the coordinate space in FIG 10 is always the origin O _(0,0), (y ( _rg) ² + x
_(bg) It can be seen that the largest point in ² ) is the optimal point for data acquisition.

The horizontal axis in FIG. 12 represents the number of pixels, and the vertical axis represents x (b−g) and y (r−r) in the coordinate space of FIG.
g). X (b) measured in this way
FIG. 13 shows an example of a state where −g) and y (r−g) are mapped in the coordinate space for the light green bottle. In FIG. 13, a plurality of different bottles are targeted for the transparent bottle, and the same bottle is tried 10 times each. In addition, a straight line represented by y = −0.6 × −10 is used as a boundary line for discriminating between transparent and light green. Thereafter, the result of the color determination is transmitted to the control unit 64 (step S17 in FIG. 6).

A method of viewing the bottom from the side of the same bottle as the light green bottle and the transparent bottle targeted in FIG. 13 (BT
FIG. 1 is a diagram of mapping in the light-colored bottle coordinate space (see the light-colored bottle color discrimination process) by the LND method in FIG.
It is shown in FIG. In FIG. 14, it is extremely difficult to distinguish between a light green bottle and a transparent bottle. However, in FIG. 13, it can be seen that the two can be clearly separated, and the color can be accurately distinguished even for a very light bottle.

Further, once converted to the coordinate space for the light-colored bottle, the distance calculation for finding the optimum point for data acquisition is performed as follows:
Since the calculation can be performed with a simple quadratic function, the calculation can be performed at an extremely high calculation speed.

<< Modifications >> (1) In the above embodiment, | R _{i + 1} −R _i for the data of any adjacent detection point (for example, R _i , R _{i + 1} for the red component R) When |> dR, it is assumed that R _{i + 1} is over the image of the glass bottle 51. However, the illumination data B _L (x, y) previously taken in is used, and R _{i + 1} and the imaging data B _b are used. _Is compared with a predetermined threshold value B _MAX, and when | B _L (x, y) −B _b (x, y) |> B _MAX , B _b (x, y) is determined as bin data. May be.
By taking the difference between the data of the same coordinates,
Even when the brightness of the illumination varies greatly depending on the location, the boundary of the glass bottle 51 can be stably detected.

(2) In the above embodiment, only the largest value among the calculated color component detection values is selected in the origin distance calculation step in the process for a transparent bottle. Alternatively, a fixed number of value groups may be selected.

(3) The present invention is not limited to a glass bottle,
The present invention is also applicable to other bottles made of a transparent resin.

(4) In the above embodiment, it is automatically determined whether the glass bottle 51 is conveyed at the head side or at the bottom side at the time of conveyance. If you decide to transport
The step of determining the transport state can be omitted.

[0093]

According to the first and fifth aspects of the present invention,
Since the color judgment is performed at the center line of the bottle, the effect of the influence of the radius of curvature of the bottle is eliminated, and there is an effect that data can be acquired with little variation.

[0094] Then, according to the present invention according to claim 1, after the coordinate detection of the upper end of the bottle, since it is not necessary to obtain in the same manner as the upper end coordinate of the lower end, an extremely short period of time and reliably center line The effect is that it can be obtained.

According to the second aspect of the present invention, it is determined whether a bottle is a dark bottle or a light bottle based on the ratio between the saturation and the lightness of the bottle, and color selection is performed by different selection steps. When color separation is difficult due to high transparency, accurate selection can be performed by increasing the selection accuracy. In particular, the invention
According to the third aspect , in the coordinate conversion step of the light-colored bottle selection step, the coordinate origin is matched with the color component of the illumination means, so that accurate determination can be performed even when the transparency is extremely high and it is difficult to determine the coloring. There is.

According to the fourth aspect of the present invention, a predetermined color component is detected at a plurality of detection points within a predetermined area, and bin selection is performed based on the sum or average of the detected color components. This has the effect of being able to be increased in depth.

According to the fifth aspect of the present invention, since the predetermined area is set up and down symmetrically with respect to the center line of the bottle, it is relatively unaffected by the radius of curvature of the bottle, and data acquisition with little variation is performed. There is an effect that can be.

According to the sixth aspect of the present invention, after calculating the average value of the color components of all the detection points in the predetermined area, a detection value greatly different from the average value is recognized, and only the remaining detection points excluding this are excluded. Since the bottle is sorted by calculating the average value, if a label or the like is affixed to the surface of the bottle, if there is a dirty part, or if there is partial halation, remove such a part. Bottle selection can be performed, and the reliability of the acquired data can be dramatically improved. In particular, claim 7 of the present invention
According to the method described above, a predetermined number of detection points are selected in order from the detection point having the largest value difference as compared with the average value and are used as the different value detection points, so that there is an effect that data variations can be efficiently reduced.

According to the eighth aspect of the present invention, when there is an attached matter such as a label or dirt on the surface of the bottle, a predetermined area can be excluded from such a portion in advance, so that the first average value calculating step can be performed. The average value can be prevented from deviating from the data of only the surface of the bottle, and the different value can be accurately recognized in the different value recognition step, and the average value can be calculated in the second average value calculation step. Has the effect of being accurate.

According to the ninth aspect of the present invention, in particular, in the case of sorting bottles of light-colored bottles, color judgment is performed on light passing through the bottom side surface of the bottle. There is an effect that the bottle can be determined and the bottle can be accurately selected as a colored bottle having high saturation.

According to the tenth aspect of the present invention, a predetermined color component is detected at a plurality of detection points in a predetermined area, and the bin selection is performed based on the sum or the average value thereof. This has the effect of being able to be increased in depth.

According to the eleventh aspect of the present invention, in the coordinate correcting step of the light-colored bottle selecting step, the coordinate origin is made to coincide with the color component of the illumination means. Can be performed. Also, the distance from the origin of the detected value of each color component in the corrected coordinate space is calculated, and a color group is selected by selecting the largest value or a fixed number of values in order from the largest value. Since it is performed, the difference with respect to the transparent color of the bottle can be accurately determined with a very simple calculation. Therefore, there is an effect that bottle sorting can be performed without requiring a large amount of calculation processing time.

According to the twelfth aspect of the present invention, the color components of the illumination means are detected at a plurality of detection points in the space outside the bottle, and the reliability of the color components of the illumination means is enhanced by the sum or the average value. Therefore, there is an effect that the origin of the coordinate space can be accurately matched with the color component of the illuminating means by conversion or correction for a light-colored bottle, so that color determination can be performed accurately.

According to the thirteenth aspect of the present invention, the positioning step according to the ninth aspect , only when the bottle has a predetermined transparency,
Since the color selection is performed by sequentially executing the imaging step and the color determination step, it is particularly advantageous that the color selection of the transparent bottle can be performed accurately.

According to the fourteenth aspect of the present invention, since the origin of the coordinate space is made coincident with the color component of the illuminating means when performing the color determination in the transparency determining step of the thirteenth aspect , the transparency can be accurately determined. There is an effect that can be.

[Brief description of the drawings]

FIG. 1 is a plan view showing the overall configuration of a bottle sorting apparatus according to one embodiment of the present invention.

FIG. 2 is a diagram showing a lighting device according to an embodiment of the present invention.

FIG. 3 is a diagram illustrating an image processing operation in an image processing unit according to an embodiment of the present invention.

FIG. 4 is a diagram illustrating a window in an image processing operation in the image processing unit according to one embodiment of the present invention.

FIG. 5 is a flowchart illustrating a bottle sorting method according to an embodiment of the present invention.

FIG. 6 is a flowchart showing a bottle sorting method according to one embodiment of the present invention.

FIG. 7 is a principle diagram showing a method for obtaining a center line of a bottle in one embodiment of the present invention.

FIG. 8 is a principle diagram showing an operation of acquiring data on the center line of the bottle in one embodiment of the present invention.

FIG. 9 is a diagram showing coordinates for a dark-colored bottle according to one embodiment of the present invention.

FIG. 10 is a view showing coordinates for a light-colored bottle according to an embodiment of the present invention.

FIG. 11 is a diagram showing an operation of performing color discrimination at the bottom of the bottle in one embodiment of the present invention.

FIG. 12 is a diagram showing the relationship between the position at the bottom of the bottle and the coordinates for the light-colored bottle according to one embodiment of the present invention.

FIG. 13 is a diagram showing a state in which data is mapped to a light-colored bottle coordinate space in one embodiment of the present invention.

FIG. 14 is a diagram showing a state where data is mapped to a coordinate space for a dark-colored bottle in one embodiment of the present invention.

FIG. 15 is a diagram showing a bottle sorting method of the first conventional example.

FIG. 16 is a view showing a bottle sorting method according to a second conventional example.

FIG. 17 is a diagram showing a third conventional bottle sorting method.

FIG. 18 is a diagram illustrating a bottle sorting method according to a fourth conventional example.

[Explanation of symbols]

 51 Glass Bottle 52 Conveyor 53 Color Recognition Unit 54 Illumination Device 55 Camera 56 Image Processing Unit 57 Sorting Unit 58 Black Bottle Collection Unit 59 Brown Bottle Collection Unit 61 Dark Green Bottle Collection Unit 62 Colorless Bottle Collection Unit 63 Different Color Bottle Collection Unit 64 Control Part 65 Bottle detection sensor 71 White fluorescent lamp 72 Diffusion plate 74, BTW window 76 Label BT bottom

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-6-142618 (JP, A) JP-A-57-131040 (JP, A) JP-A-6-63514 (JP, A) JP-A-6-63514 218335 (JP, A) Japanese Utility Model 63-94588 (JP, U) (58) Fields investigated (Int. Cl. ⁷ , DB name) B07C 5/342

Claims (14)

(57) [Claims] 1. A sorting method for sorting bottles made of glass or translucent resin by color, comprising: a step of detecting coordinates of both end lines of the bottle; and a coordinate of the both end lines. A center line coordinate calculating step of calculating the coordinates of the intermediate line as the center line coordinates of the bottle; an imaging step of imaging the bottle with the center line coordinates calculated in the center line coordinate calculating step; A color determination step of performing a color determination by processing the image captured in the step , wherein the two-sided end coordinate detection step includes placing the bottle in a lying state by placing means , and placing the bottle behind the bottle.
While irradiating light from the illumination means arranged on the side,
The brightness detector placed on the front side is oriented in the direction of the center axis of the bottle.
Scans intermittently at predetermined intervals in the perpendicular vertical scanning direction.
The brightness detected by the brightness detector is equal to or greater than a certain brightness difference.
The coordinates in the vertical scanning direction when the
An upper end coordinate detection step of detecting as coordinates, and a contact point between the placing means and the bottle in the vertical scanning direction.
Bottom coordinate detection for detecting the coordinates of the bottle as the bottom coordinates of the bottle.
Bottle sorting method and a step out. 2. Bottles made of glass or translucent resin are colored.
A saturation detecting step of detecting the saturation of the bottle by a saturation detecting means.
When the brightness detection step of detecting the brightness of the bottle at lightness detection means
And the saturation detected in the saturation detection step and the brightness detection step.
Saturation / brightness ratio calculation process for calculating the ratio with the detected brightness
And the value of the ratio calculated in the saturation / brightness ratio calculation step is constant
When it is larger than the standard value, it is recognized as a dark bottle,
Recognized as a light-colored bottle when it is smaller than a certain standard value
In the certification process, when the bottle is certified as a dark bottle in the certification process,
A dark- colored bottle sorting step of color sorting, and a light-colored bottle when it is identified as a light-colored bottle in the certification step.
Bottle sorting method and a light-colored bottle sorting step of performing the color selection. 3. The dark-colored bottle sorting step includes a first color component detecting step of detecting a predetermined color component of the bottle.
If, concentrated the detected value of the color component detected by the color component detecting step Irobin
A first map step of arranging in the coordinate space for the dark bottle, and a coordinate space for the dark-colored bottle arranged in the first map step .
In the process, the color component detected in the color component detection step is detected.
A first color identification step of performing color identification of the output value , wherein the light-colored bottle selection step is such that the color component of the illuminating means for irradiating the bottle matches the coordinate origin.
As described above, the coordinate space for the dark bottle is changed to the coordinate space for the light bottle.
And a second color component detecting step of detecting a predetermined color component of the bottle.
When the detected value of the color component detected by the color component detecting step pale
A second map step of arranging in the coordinate space for the color bottle; and a coordinate space for the light-color bottle arranged in the second map step .
In the process, the color component detected in the color component detection step is detected.
And a second color identification step of performing the color identification of the detection value, claims
2. The bottle sorting method according to 2 . 4. Bottles made of glass or translucent resin are classified by color.
A sorting method for sorting, the both side edge coordinate detection step of detecting the coordinates of the two side lines of the bottle
And the center line coordinates of the bottle with the coordinates of the middle line of the coordinates of the both end lines
And the center line coordinate calculation step of calculating as a target, the center line coordinates calculated by the centerline coordinate calculation step
An image capturing step of capturing the bottle by using the image capturing step, and performing a color determination by processing the image captured in the image capturing step.
A color determination step , wherein the imaging step includes a plurality of steps around the center line coordinates of the bottle.
Imaging a predetermined area including a detection point, wherein the color determination step includes a step of: obtaining a predetermined color composition at the plurality of detection points in the predetermined area.
A color component detection step of detecting each of the components, and each of the predetermined color components detected in the color component detection step.
Sum or average of the plurality of detection points for each
Bottle sorting method comprising a calculation step of calculating a. 5. The predetermined area is centered on a center line of the bottle.
The bottle sorting method according to claim 4, wherein the method is set to be vertically symmetrical . 6. The computing step comprises the step of : calculating an average value of the predetermined color components of all detection points in the predetermined area.
Calculating a first average value, and calculating the first average value among all the detection points in the predetermined area.
Detected value relatively different from the average value calculated in the output step
A different value recognition step of recognizing a plurality of different value detection points, and excluding the different value detection points from all the detection points in the predetermined area.
Calculating the average value of the predetermined color components of the remaining detection points
5. The bottle sorting method according to claim 4, comprising a second average value calculating step . 7. The method according to claim 1, wherein the different value recognition step compares the average value with the average value.
The specified number of detection points in order from the detection point with the largest value difference.
7. The method according to claim 6, further comprising the step of selecting a point to be the outlier detection point.
Bottle sorting method as described . 8. The bottle sorting method according to claim 6, wherein
In order to avoid the predetermined area from being attached to the surface of the bottle,
A bottle sorting method further comprising a region specifying step for defining a bottle. 9. Bottles made of glass or translucent resin are classified by color.
A method for selecting a position in a captured image of an imaging device disposed in front of the bottle.
Position where the fixed area is positioned to fit the bottom side of the bottle
Deciding step, and irradiating light from illumination means arranged behind the bottle.
In the imaging device, the bottom side surface of the bottle is positioned in front of the bottle.
An image pickup step of picking up an image from the side, and performing color judgment by processing the image picked up in the image pickup step
Bottle sorting method and a color determination step. 10. The color judging step includes the steps of: determining a predetermined color component at a plurality of detection points in the predetermined area;
A color component detecting step of respectively detecting the predetermined color component detected by the color component detecting step
That is, the value having the largest difference from the preset transparent data,
Alternatively, a fixed number of value groups are sequentially calculated from the value having the largest difference.
Selection process and color only for selected values or values set by the selection step of selecting
The bottle sorting method according to claim 9, further comprising a selection value determining step of performing a determination . 11. The color judging step comprises the steps of: originating a coordinate space used for color judging;
Correct the coordinate space so that the color component of the light means matches
And a predetermined color component at a plurality of detection points in the predetermined area.
The color component detection step to detect each, the detection value of the color component detected in the color component detection step,
A mask to be placed in the coordinate space corrected in the coordinate correction step.
And detecting the detected values of the respective color components in the corrected coordinate space.
An origin distance calculating step of calculating a distance from the origin, and the detected value of the color component calculated in the origin distance calculating step
Of the largest values, or in order from the largest value
A selection process to select a certain number of value groups, and colors only for the values or value groups selected in the selection process
The bottle sorting method according to claim 9, further comprising a selection value determining step of performing a determination . 12. An illumination for detecting a color component of said illumination means.
The illumination color detection step further comprises detecting a color component of the illumination means at a plurality of detection points.
And summing or averaging the color components for the plurality of detection points
And a step of calculating a value, according to claim 3 or claim 11
Bottle sorting method as described . 13. A bottle made of glass or translucent resin is colored.
A screening method for separately screening, wherein a transparency is determined to determine whether the bottle has a predetermined transparency.
Executing a judging step, judging to have the predetermined transparency in the transparency judging step
Then, the positioning step, the imaging step and the imaging step according to claim 9 are performed.
Bottle sorting method characterized by sequentially executing and color determination process. 14. The transparency judging step includes the steps of: originating a coordinate space used for color judgment;
Correct the coordinate space so that the color component of the light means matches
A coordinate correction step, the color component detecting step of detecting a predetermined color component of the bottle, the detection value of the color component detected by the color component detecting step before
The map to be placed in the coordinate space corrected in the coordinate correction process
In the color component detecting step on the corrected coordinate space.
A color identification step of performing color identification of the detected value of the color component
14. The bottle sorting method according to claim 13, comprising: