JPH0493756A - Apparatus for detecting foreign matter in bottle - Google Patents

Abstract

PURPOSE:To make it possible to prevent erroneous detection due to the deviation of images by operating the minimum value of the absolute value of the output difference between the picture element of a reference image and the image at a corresponding position and the picture element at the surrounding part. CONSTITUTION:A bottle 11 which is conveyed with a conveyer is rotated with spin belts 2-1 and 2-2. After the rotation is stopped, the image is picked up with CCD cameras 6-1 and 6-2 when the bottle passes through position sensors 3-1 and 3-2. At this time, a foreign matter 11a continues rotation together with liquid by inertial force after the stop of the rotation of the bottle. Therefore, two images of the foreign matter 11a are picked up at the different positions. An image processing device 7 operates the minimum value of the absolute value of the output difference based on the output signals. The signals caused by flaws, contamination and the like of the bottle are removed from the value, and only the foreign matter in the content liquid is detected. In this way, the slight deviation between the images which are picked up with two CCD cameras can be completely eliminated, and the high-speed inspection with less erroneous detection can be performed.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a detection device for foreign objects floating in the liquid contained in a transparent bottle, and is also applicable to a detection device whose purpose is to detect only moving objects. Applicable.

[Conventional technology]

As a device to detect floating foreign objects in bottle liquid, two cameras capture images of moving bottles from the same direction and at the same distance.
An inspection device (Japanese Patent Application No. 63227571) that calculates the difference in the brightness of these image pixels and extracts only foreign matter, holds the bottle still and takes two images with one camera, A device (Japanese Unexamined Patent Publication No. 62-220844) that extracts floating foreign matter by calculating the difference has been proposed.

[Problem to be solved by the invention]

The above-mentioned conventional technology has the following problems.

(1) A detection device that images a bottle twice with one camera and extracts only floating foreign matter from the difference between these pixel elements (Japanese Patent Laid-Open No. 6
No. 2-220844), it is necessary to rotate the bottle to give motion to the foreign object before imaging, and to keep the positional relationship between the camera and the camera still during imaging. For this reason, it is necessary for the bottles flowing along the line to stop moving temporarily for imaging, and then to be moved again after imaging is completed. This method has disadvantages such as the acceleration and deceleration of the bottle movement, which takes time and prevents high-speed processing, and the acceleration applied to the bottle generates bubbles, which become noise and cause false detection.

(2) In order to solve the problem mentioned in item (1) above, a device (patent application 1983-1988) was developed which images a moving bottle twice from the same direction using two cameras and extracts foreign substances from the difference between these images. 227
No. 571). With this method, there is no acceleration or deceleration of the bottle, so it has the advantage of being able to inspect at high speed without generating bubbles. In this case, the two images may be misaligned, and the border of the image signal fixed to the bottle, such as the edge of the bottle, scratches, dirt, etc., may be erroneously detected as a foreign object signal.

The present invention aims to prevent erroneous detection due to image shift by image processing in the inspection method described in item (2) above.

[Means for solving the problem] Two sets of image reading devices, an image processing device that calculates an output signal of the image reading device, and a separator that selects defective bottles based on a foreign matter determination signal of the image processing device are provided. In the summer, the pixel of one reference image and
The pixel of the other image at the position corresponding to the same pixel element and the outputs of its surrounding pixel elements are compared.

The absolute minimum value of the difference between the outputs is calculated, and signals caused by scratches, letters, dirt, etc. on the bottle are removed, and only foreign objects in the liquid are detected.

[Effect]

A foreign object detection device in a bottle includes two sets of image reading devices, an image processing device that calculates an output signal of the image reading device, and a separator that selects defective bottles based on a foreign object determination signal of the image processing device. By adding a function to calculate the minimum value of the difference (absolute value) between the outputs of pixels on the reference screen and surrounding pixels on the comparison screen, it is possible to completely eliminate slight deviations in images captured by the two cameras.

〔Example〕

An embodiment of the in-bottle foreign matter detection device according to the present invention is shown in FIG. 1, and the flow of this image processing is shown in FIG.

In FIG. 1, bottles 11 are conveyed by a conveyor 1.
is given rotation by the spin belt 2-1.2-2, and after it stops rotating, when it passes the position sensor 3-1.3-2,
Light # (strobe) Lights up 5-1.5-2 and CC [
]] A close-up image is taken by the camera 6-16-2. Foreign object 11a
Since the bottle continues to rotate together with the liquid due to inertial force even after the bottle stops rotating, the two images are taken at different positions.

By taking the difference between the two images, only the difference due to the positional shift of the foreign object can be removed. Depending on the magnitude of this signal, the separator 8
By controlling the operation of the bottle, it is possible to classify the bottles into good and defective ones.

In addition, 4 is a light source controller, 9 is a defective product line, and 10 is a light source controller.
is a good product line.

2-') is expressed as the brightness Sij of one pixel element in the first image and the brightness S' of the corresponding pixel element in the second image.
A means for adding an operation si function to take the minimum value of the absolute values of nine differences between ij and eight surrounding pixels of the same pixel element is adopted. That is, the difference between the positions (i, j) is δ, J-11in (l S
t, J-3' i+I, j'l, l St, i"'
r, = l, l Sr, J-5' i-+, J1.
l 5=1-5' 6 ++-= +ll , :S
*, J-5' r1=”+ l, Is;, J-5'
rJ+l, l 5l-J-5' I+l, J-11
, l s,, JSr2a-+ 1. lS;, J-5
;-+, i-+ l l.

Similarly, the difference δ' with respect to the second screen can also be defined by the same formula as above. This can be done by replacing S and S' in the above equation. Here, the difference δS, J in coordinates ij is expressed as δSi, j -max (δl+j δ',, j+, and δS ij≠O or δS ij ≧ε (constant) (i
.

j) i! The ii element determines that there is a foreign object.

Hereinafter, this calculation function will be referred to as a "difference filter."

[Operation] In the above calculation, even if there is a positional shift of one pixel,
The image difference is calculated by comparing the brightness with the surrounding 900 elements of the corresponding opponent's screen, and the difference is 1i! ! If it is within the elementary range, it is possible to remove image signals such as stains on a bottle.

Two reference screens were taken, the first image and the second image.
It is clear from the calculation algorithm that the same function can be achieved even with only one. However, if the reference image is the first one. By using the second and both screens, the probability of detecting floating foreign objects increases.

For example, if a foreign object is not detected in the first image but only in the second image, the foreign object signal cannot be detected only by calculation based on the first image (see FIG. 3).

In this embodiment, the difference from the surrounding nine pixels is calculated, but if the deviation is large, a comparison with 5 x 5 = 25 pixels may be performed. Note that if the foreign object moves by one pixel or more during the capture of two images within the bottle, the signal will not be removed by this differential filter.

FIG. 2 shows the flow of image processing, and the function of the differential filter, which is the main part of the present invention, will now be explained with reference to FIG.

i-2, i-1, i, i +1 in FIG. i +2
and j-2j-1,j,j ±1. j+2 is the coordinate indicating the position of a part of the pixel on the image, A is a foreign object, and B is a scratch signal. i device (Lm) (1=i-2~i+2
．． The numbers 1 and 0 corresponding to m=j-2 to j+2) represent brightness, with 1 being dark and 0 being bright, and are binarized for simplicity. Assuming that images 1 and 2 are misaligned by one pixel each in the horizontal and vertical directions, the flaw signal B of image 1 is misaligned as in image 2.

On the other hand, it is assumed that the foreign object moves significantly as the liquid rotates and comes to the position of image 2 in the figure, for example. If we calculate the difference filter between the reference screen and its surrounding 900 pixels as explained earlier, we get s-s'
Difference values such as δ for δ and δ′ for δ′−3 are obtained, and both of them remove the flaw signal. δ and δ
As shown in the figure, the wax calculation δS of ' will detect two foreign object signals for one foreign object.

〔Effect of the invention〕

The present invention provides a method for detecting foreign matter in a bottle, which includes two sets of image reading devices, an image processing device that calculates an output signal of the image reading device, and a separator that selects defective bottles based on a foreign material determination signal of the image processing device. In the detection device, the pixel of one reference image is compared with the output of the pixel of the other image at the position corresponding to the same pixel and its surrounding pixel, and the minimum absolute value of the difference is calculated. However, by providing an image processing device having a function of removing signals caused by scratches, letters, stains, etc. on the bottle based on this value and detecting only foreign matter in the liquid contained in the bottle, the following effects are achieved.

By adding a function to calculate the minimum value of the peripheral pixels of the reference screen and comparison screen to the conventional bottle inspection device for which a patent has been applied for in Japanese Patent Application No. 63-227571, etc., it is possible to improve the Slight deviations can be completely eliminated, and it is possible to provide high-speed, highly reliable actual bottle inspection with few false positives.

In addition, we have automated visual inspections that were previously performed by humans.
It can save labor.

[Brief explanation of the drawing]

Fig. 1 is a schematic configuration diagram of a foreign object detection device in a bottle according to an embodiment of the present invention, Fig. 2 is a diagram showing the procedure flow of foreign object detection and image processing contents, and Fig. 3 is a diagram showing the function of a differential filter. It is. 1... Conveyor, 2-1.2-2... Spin belt 3-1.3-2... Bottle position sensor. 5-1.5-2...Light source, 6-1.6-2.
··camera. 7... Image processing device 11... Bottle. 8...Separator. 11a...Foreign object

Claims (1)

[Claims] A foreign object detection device in a bottle includes two sets of image reading devices, an image processing device that calculates output signals of the image reading devices, and a separator that selects defective bottles based on the foreign object determination signal of the image processing devices, Compare the pixel of one reference image with the output of the pixel of the other image at the position corresponding to the same pixel and its surrounding pixel, calculate the minimum absolute value of the difference, and calculate this value. 1. A foreign object detection device in a bottle, comprising an image processing device having a function of detecting only foreign objects in the liquid content by removing signals caused by scratches, letters, dirt, etc. on the bottle.