JPH05126761A - Empty-bottle separating apparatus - Google Patents

Abstract

PURPOSE:To identify the color of a bottle quickly and automatically with high reliability by identifying the light transmitted through the bottle under inspection and the light, which is not transmitted through the bottle, and thereafter measuring the wavelength component of the light transmitted through the bottle. CONSTITUTION:The light from a light source 7 is scattered with a scattering plate 6 and cast on a glass bottle under inspection 1. The image is picked up with a CCD camera 3. In an image processing device 4, the image plane, wherein the amount of the transmitted light that is directly cast into the camera 3 without passing the bottle 1 is larger than a specified threshold value, is removed based on the image signal obtained with the camera 3, and the image region of the battle is set. Then, the part, wherein the amount of the light is less because of a label and contamination in the image region of the bottle, is avoided, and the region having the highest luminance is made to be the region for identifying the color of the bottle 1. The wavelength components (green, red and blue) of the light in this region are measured, and the color of the bottle 1 is identified based on the magnitudes. Thus, the separating work for the recovered bottles can be automated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an empty bottle sorting apparatus for automatically sorting the colors of glass for the purpose of reusing (recycling) various used glass bottles.

[0002]

2. Description of the Related Art Collected glass bottles are usually conveyed by conveyors at a recycling facility operated by local governments, visually classified by workers into several types of colors, and crushed into cullet bottles. Used as a raw material in the factory.
At present, the color separation of glass bottles depends on the judgment of human eyes, and there is no example of automation.

[0003]

The reason why the sorting of the bottles is not automated and depends on the human eyes is because the size and size of the collected bottles are large.
This is because in addition to the wide variety of shapes and colors, the shapes are also varied, and it is difficult to identify them by machines. That is, even for bottles of the same size, the state of the transmitted light of the bottle changes depending on the label attachment form, the presence or absence of the lid, the degree of stain or the state of damage, etc., and the amount of transmitted light and the shape of its color tone also vary depending on the size of the bottle. Because of the difference, humans have not established a technique for accurately and quickly identifying the parts of the bottle that are selected for color classification in an instant by mechanical means such as image processing.

In view of such a situation, the present invention has an object to realize an identification algorithm similar to a human identification mechanism by using an image signal from an illumination and a CCD camera, and identifies a bottle color. It aims to automate in a fast and reliable way.

[0005]

In the present invention, in order to solve the above-mentioned problems, a scattered light source is arranged below the bottle to be inspected, and a CCD camera is provided above the bottle so that the transmitted light of the bottle can be determined from the brightness of the CCD camera image signal. The function of identifying the area and the area directly incident on the camera from the light source, the function of selecting the partial area having the largest amount of transmitted light in the area of the bottle transmitted light, and the color wavelength component of this partial area, that is, green (G) , Red (R),
An image processing device having a function of discriminating the color of glass from the relative intensity of blue (B) light is provided.

Then, the scattered light source and CC for the bottle to be inspected
There are the following three types of arrangement with the D camera. First, the light source and the CCD camera are arranged in parallel above and below the body of the bottle to be inspected, and a partial area having the largest transmitted light amount is selected from the camera image signal of the bottle transmitted light. Is characterized by.

Secondly, the light source is arranged in parallel with the mouth of the bottle to be inspected and the CCD camera is arranged in parallel with the bottom of the bottle to be inspected, and the diameter of the bottle is determined from the image signal of the camera. Is measured, and a region for color discrimination is selected according to the size of the diameter. Thirdly, the light source is arranged at the mouth and the bottom of the bottle to be inspected, and the CCD camera is arranged so that the transmitted light from the light source arrives in synchronism with the mirror through the anti-rotation strobe. It is characterized in that a partial region having the largest amount of transmitted light is selected from the camera image signal.

[0008]

By the light source through the scattering plate, it is possible to obtain a uniform bottle transmission image without extremely uneven brightness due to the lens effect of the bottle. The bottle transmission image obtained in this way is
Since the amount of transmitted light is smaller than the region where light directly enters from the light source, the region of bottle transmitted light can be identified by dividing the luminance of the image signal in the camera visual field region by a certain threshold value. Furthermore, if you take out the area with the highest brightness in the bottle area, select the area through which light passes perpendicularly to the glass, excluding areas such as labels and stains or the bottom corners where the bottle shape is complicated. Which corresponds to the area that humans select for color discrimination. The color component (G,
The distribution of R, B) has a strong correlation with the color classification of the glass regardless of the light transmittance of the glass, and thus the color classification can be performed by obtaining the ratio of the G, R, B components.

[0009]

FIG. 1 shows a first embodiment of the present invention. Figure 1 below
Therefore, the configuration and the function of the present invention will be described. Light emitted from a light source 7 such as a high-frequency fluorescent lamp is scattered by a scattering plate 6 made of an acrylic plate or the like and passes through the test bottle 1 placed above to form a bottle image. The bottle is placed in parallel with the light source by the hold device 2, and its image is converted into an image signal by the CCD camera (TV camera) 3. The image signal is processed by the image processing device 4, and is mechanically classified on the downstream side by an appropriate bottle sorting device 9 according to the bottle color identification signal.

The image processing apparatus 4 first sets a bottle area for an image obtained by the camera 3, for example, the image shown in FIG. The area outside the bottle in the figure and on the field-of-view screen 11 is an area where the light source directly enters the camera without passing through the bottle, and the amount of transmitted light is larger than that of the bottle image 12. Therefore, it is possible to set the area of the bottle image by excluding the bright portion exceeding a certain threshold value from the pixels of the view screen or by extracting the dark portion. In the bottle area thus obtained, the amount of light is smaller in the label, the stained portion, the thick portion that is not perpendicular to the light source, and the like, as compared with the portion 14 through which the light source passes vertically. Therefore, if the area with the highest brightness is set in the bottle image area, this area becomes an area where the label and the like are avoided, and approximately coincides with the area selected by a person for identifying the color of the bottle. Since the ratio of the light wavelength components (G, R, B) of the region 14 thus determined has a strong correlation with the color separation of the glass, the color of the bottle to be inspected can be identified from the size of these light wavelength components. It will be possible. Even for porcelain and metal containers that are similar in shape to bottles on the inspection line, there is no transmitted light in the bottle image, so the brightness is extremely small, so it is possible to easily identify on the same inspection line. An example of the flow of the image processing procedure is shown in FIG.

A second embodiment of the present invention is shown in FIG. The light emitted from the white light emitting body 7 provided at the mouth of the bottle 1 to be inspected passes through the glass at the mouth of the bottle and its surroundings through the scattering plate 6, and directly or indirectly illuminates the inside of the bottle. A circular image is taken by the CCD camera 3 arranged on the axis at the bottom of the bottle. The image signal thus obtained is processed by the image processing device 4 to obtain a color classification signal, and the bottle classification device 9 is controlled by this signal to perform color classification of the bottles.

From the image shown in FIG. 5 obtained by the CCD camera, the diameter 2R of the bottle is calculated by the image processing device from the distribution of the amount of transmitted light. Since the ratio r / R of the distance r from the center of the color judgment area where the transmitted light is relatively uniform and the bottle radius has a substantially constant value in many bottles, this area can be obtained by knowing the above R and bottle center coordinates. The position on the screen is required. In this way, the detection window (in the measurement area for identifying the color, the transmitted light of this portion is analyzed) can be appropriately determined according to the bottle size. For example, in FIG. 5, a window 16 is defined for the image of the large bottle 12, a window 17 is defined for the medium bottle image 13, a window 18 is defined for the small bottle image 14, and a window 19 is defined for the very small bottle 15. By measuring the magnitude of the color components (R, G, B) of the window thus obtained, the shade to be identified can be calculated from these values.

A third embodiment of the present invention is shown in FIG. Generally, it is relatively easy to align the bottles in the longitudinal direction for the time being, but it is very difficult to align the head or bottom of the bottles due to the various shapes of the bottles. Therefore, in the third embodiment, as shown in FIG.
5 are arranged in both the head direction and the bottom direction of the test bottle 1, and the CCs are arranged through the reflection mirror 17 and the image synthesis mirror 18
With a D camera 3.1, a bottle bottom image 22 and a bottle head image 23 are obtained as shown in FIG.

As for the image pickup timing, the timing signal from the image processing device 4 is output,
The two light emitters 7 are sequentially made to emit light to synchronize. Further, in color identification, the bottle bottom image 22 and the bottle head image 2
Set 2 to multiple color analysis windows 14 in 3,
By selecting the one with strong light intensity, it is possible to deal with either the bottom or the head of the bottle.

This makes it possible to capture an image of the bottom of the bottle with very few labels into the image processing device, regardless of whether the bottle is facing up or down. In the image processing apparatus, the color analysis area is set at the image position of the bottle, and the distribution of the color components (R, G, B) in this portion has a strong correlation with the color classification of the glass. Color classification can be performed by obtaining the ratio of the B component. Also, the lid remains,
The effect of shading due to the label or the like on the neck can be eliminated by comparing the light intensities of the color analysis areas and determining which is the bottom of the bottle.

The method of using the half mirror 15 shown in FIG. 6 is only an example. A method of mechanically synchronizing the half mirror 15 with a rotary mirror, or a method in which two cameras are respectively in the bottle head direction and the bottle bottom are used. There is also a method of placing the light source in a certain direction and capturing images synchronously with the opposing light source.

[0017]

EFFECTS OF THE INVENTION Since the present invention is configured as described above, it is possible to automate the sorting operation of collecting bottles, which was conventionally performed by a worker, and the human being is exposed to contaminants and exposed to the noise of the inspection line. The chances of working can be greatly reduced, and the effect of improving the working environment can be obtained. Also, as mentioned above, the sorting algorithm and hardware are simple, so it is possible to provide an inspection device at a low cost, and it is possible to provide high-speed and unattended processing. Technical and economic effects such as improvement and reliability of inspection can be obtained.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing a first embodiment of the present invention.

FIG. 2 is an explanatory diagram of image processing according to the first embodiment.

FIG. 3 is a color analysis flowchart of the present invention.

FIG. 4 is a schematic view showing a second embodiment of the present invention.

FIG. 5 is an explanatory diagram of a relation between a bottle image and a color discrimination area according to the second embodiment.

FIG. 6 is a schematic view showing a third embodiment of the present invention.

FIG. 7 is a bottle image view from the neck and bottle bottom directions of the third embodiment.

[Explanation of symbols] 1 bottle to be inspected 3 CCD camera 4 image processing device 7 light source

Claims (4)

[Claims] 1. A light source for irradiating the bottle to be inspected with scattered light,
A CCD camera for photographing the bottle to be inspected illuminated by the light source, and the bottle-transmitted light reaching the CCD camera are distinguished from the light not transmitted through the bottle, and the wavelength component of the light is measured to classify the bottle color. An empty bottle sorting device comprising an image processing device having a function. 2. The light source and the CCD camera are arranged in parallel above and below the body of the bottle to be inspected, and a partial area having the largest transmitted light amount is selected from the camera image signal of the bottle transmitted light. The empty bottle sorting device according to claim 1, wherein 3. The light source is arranged in parallel with the mouth of the bottle to be inspected, and the CCD camera is arranged in parallel with the bottom of the bottle to be inspected, and the diameter of the bottle is determined from the image signal of the camera. 2. The empty bottle sorting device according to claim 1, wherein a region for color discrimination is selected by measuring and measuring the diameter. 4. The CCD camera is arranged so that the light source is arranged at the mouth and the bottom of the bottle to be inspected, and the transmitted light from the light source arrives in synchronism with the mirror through an anti-rotation strobe. The empty bottle sorting apparatus according to claim 1, wherein a partial area having the largest amount of transmitted light is selected from the camera image signal.