JP2008002917A - Content inspection method of liquid packed in container, content inspection device and method of manufacturing liquid food packed in container - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a content inspection method of a liquid packed in a container capable of accurately grasping the occurrence of flocs with the passage of time of production. <P>SOLUTION: The liquid in the container is inspected through an imaging process, an image recognizing process, a reference pixel number setting process and a floc occurrence judging process. In the imaging process, the image of the flowing liquid is taken so as to prove a time lag to output the obtained image data. In the image recognizing process, the temporal change of the image data is analyzed to recognize a fine floating substance containing air bubbles and flocs and the amount of the line floating substance is recognized as the number of pixels. In the reference pixel number setting process, the reference number of pixels is set on the basis of the image data obtained by imaging in a state that flocs are not contained at all or not almost contained in the liquid. In the floc occurrence judging process, the number of pixels recognized in the image recognizing process is compared with the reference number of pixels set in the reference pixel number setting process to judge whether flocs exceeding an allowable amount are produced. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a method for inspecting the contents of a liquid in a container, a content inspection apparatus, and a method for producing a liquid food in a container, and particularly relates to determination of occurrence of a flock.

Conventionally, various methods have been proposed for quality inspection of liquid food in containers. In particular, since contamination with liquid food is a major problem, various foreign substance inspection methods for inspecting foreign substances in liquid have been proposed. Examples of the foreign substance inspection method include a visual method, an X-ray method, and a method of analyzing an image captured by a camera. Among these methods, image analysis is often used (see, for example, Patent Document 1). ). Moreover, in the foreign material inspection which performs image analysis, the method of distinguishing the damage | wound and pattern of a container from a foreign material is proposed (for example, refer patent document 2, 3).
JP 2002-122550 A Japanese Patent Publication No. 4-6900 Japanese Patent Laid-Open No. 8-159989

  By the way, in liquid foods such as liquor and vinegar, a flock (a part of the liquid is solidified) may be generated in the filling process of filling the container with the liquid. Although this floc is not a foreign object, when it can be recognized by the human eye beyond the allowable amount, it looks bad and causes a problem in quality, which may lead to a deterioration in the image for the customer. However, a device for inspecting the occurrence of floc has not been proposed. Moreover, even if the conventionally proposed foreign matter inspection method is used, it is impossible to grasp the occurrence of flocs with the lapse of production time.

  The present invention has been made in view of the above-mentioned problems, and a first object thereof is to provide a content inspection method and a content inspection device for a liquid in a container capable of accurately grasping the occurrence of floc associated with the lapse of production time. There is. In addition, a second object is to provide a method for producing a liquid food in a container that can produce a liquid food in a container that does not deteriorate in quality due to the inclusion of flock.

  When a liquid is filled into a container using a cleaned liquid filling apparatus to produce a liquid food in a container, a flock is hardly generated at the start of production. However, flocs may gradually occur as production progresses. The initial floc is very small, but the floc tends to become gradually larger. On the other hand, bubbles are generated from the start of production. However, unless the production conditions are changed, the amount of bubbles basically does not change even if the production proceeds. Bubbles are generated due to the impact of when the liquid is dropped when the container is transported or filled, and changes depending on the production lot. The amount of bubbles may vary depending on the liquid filling speed, the liquid filling temperature, the container transport speed, and the like.

  However, an extremely small floc is difficult to recognize with human eyes (100 μm or less), and is smaller than a level normally recognized as a foreign object (200 μm or more). Since this extremely small floc is not a foreign matter, it may not be a problem in quality, but it is necessary to grasp the occurrence of a floc at the small floc stage in order to prevent the occurrence of a large floc. Also, when recognizing extremely small flocs by image processing, it is necessary to increase the accuracy of a camera or the like that images the liquid, but it is difficult to distinguish small flocs from scratches or bubbles in containers.

  As a result of intensive studies by the present inventors in view of the above problems, attention was paid to the fact that no flocs are generated (or very little) at the beginning of liquid filling. Then, in the state of capturing and storing the micro suspended matter other than the liquid containing bubbles at the beginning of filling, the subsequent tendency of the micro suspended matter is monitored, and the micro suspended matter at the beginning of filling and the subsequent micro suspended matter are It was newly found that the occurrence of floc can be detected if the compared value deviates from the allowable amount. And based on this novel knowledge, the present inventors came to complete the following outstanding invention.

  That is, according to the first aspect of the present invention, there is provided an imaging step of capturing an image of the liquid with a time difference in a state where the liquid in the container is flowing, and outputting the obtained image data, and temporal change of the image data. And recognizing a minute suspended matter other than the liquid in the liquid, and recognizing the amount of the minute suspended matter including bubbles and flocs as the number of pixels, and the floc is completely or almost not in the liquid. A reference pixel number setting step for setting a reference pixel number based on the image data obtained by imaging without being included, the pixel number recognized in the image recognition step, and a setting in the reference pixel number setting step A content inspection of the liquid in the container, comprising a step of determining whether or not a floc exceeding the allowable amount is generated by comparing the reference pixel number Law and the gist of.

  Therefore, according to the first aspect of the present invention, the number of pixels recognized in the image recognition step is compared with the reference pixel number set in the reference pixel number setting step, and the determination in the flock occurrence determination step is performed, thereby It is possible to accurately grasp the occurrence of flock over time.

  There are two types of flocs: filling flocs and secondary precipitation. “Filling flock” means, for example, that part of the liquid (protein, etc.) is agglomerated at the contact portion between the liquid surface of the liquid stored in the tank in the liquid filling device and the wall surface of the tank depending on conditions such as temperature. It occurs and is assumed to be mixed in the container when the liquid is filled using the liquid filling apparatus. On the other hand, “secondary precipitation” means that after a certain number of days have elapsed since the container was filled with the liquid, a part of the liquid aggregates and is generated as a precipitate / starch. Note that the floc of the present invention is a filling floc.

  According to a second aspect of the present invention, in the first aspect, in the reference pixel number setting step, the filling speed of the liquid into the container, the filling temperature of the liquid, and the conveyance speed of the container after filling are set. The gist is to set the reference pixel number in consideration of at least one element.

  That is, in the invention described in claim 2, focusing on the fact that the state and number of bubbles change when the liquid filling speed, the liquid filling temperature, and the transport speed of the container after filling change, A reference pixel number serving as a reference for detecting a flock is set in consideration of at least one. For this reason, even if the factor that causes the state / number of bubbles to change is changed, it is possible to accurately grasp the occurrence of flocs as the production time elapses.

The gist of a third aspect of the present invention is that, in the first or second aspect, in the imaging step, an image is obtained with a region of 30 cm ² or less in the body portion of the container as a visual field.

That is, in the invention according to claim 3, paying attention to the fact that only one or a small number of flocs and bubbles are generated as foreign matter, but not uniformly in the entire liquid, not the entire container containing the liquid. An image is obtained with a region of 30 cm ² or less in the body part as a field of view. As a result, the number of pixels per area of the image can be increased as compared with the case where the entire container is imaged, so that the resolution of the image is increased and the accuracy with which the occurrence of flocks can be grasped is improved. Further, since the image is obtained with the body portion as the field of view instead of the bottom of the container, for example, when the liquid flows in the clockwise direction with respect to the axial direction of the container, the state of movement of the liquid (floc) appears more clearly. Images can be obtained. Therefore, compared with the case where an image is obtained with the bottom of the container as a visual field, occurrence of floc can be grasped with higher accuracy. Furthermore, since it is possible to reduce the size of the image pickup means for picking up an image of the liquid, it is possible to simplify the inspection apparatus for inspecting the floc.

In claim 3, to obtain an image of 30 cm ² or less in the region in the body portion as a viewing, preferably, 1 cm ² or more 20 cm ² or less in the body portion, more preferably, 10 cm ² or more at the body portion 20 cm ² It is preferable to obtain an image with the following areas as the field of view.

  According to a fourth aspect of the present invention, after performing the filling step of filling the container with the liquid, the container is tested by the content inspection method for the liquid in the container according to any one of the first to third aspects. The gist is a method for producing a liquid food in a container, which is characterized by producing a liquid food in a container.

  Therefore, according to the fourth aspect of the present invention, the number of pixels recognized in the image recognition step is compared with the reference pixel number set in the reference pixel number setting step, and the determination in the flock occurrence determination step is performed, thereby It is possible to accurately grasp the occurrence of flock over time. Therefore, it is possible to manufacture a liquid food in a container that does not deteriorate in quality due to the inclusion of flock.

  According to a fifth aspect of the present invention, there is provided an imaging means for outputting image data obtained by taking an image of the liquid with a time difference in a state where the liquid in the container is flowing, and analyzing temporal changes of the image data Image recognizing means for recognizing a minute suspended matter other than the liquid in the liquid and recognizing the amount of the minute suspended matter including bubbles and flocs as the number of pixels, and a floc is contained in the liquid at all or almost. A reference pixel number setting means for setting a reference pixel number based on the image data obtained by imaging in a non-existing state, and a reference pixel number storage means for storing the reference pixel number set by the reference pixel number setting means And whether the number of pixels recognized by the image recognition means is compared with the reference pixel number stored in the reference pixel number storage means, and whether or not a floc exceeding an allowable amount has occurred. The contents inspection device containers liquid; and a floc generation judging means for judging as its gist.

  Therefore, according to the fifth aspect of the present invention, the flock generation determination unit compares the number of pixels recognized by the image recognition unit with the reference pixel number set by the reference pixel number setting unit, and performs the determination. It is possible to accurately grasp the occurrence of flocs over time of production. The image recognizing unit recognizes the minute suspended matter in the liquid by analyzing the temporal change of the image data, and recognizes the amount of the minute suspended matter as the number of pixels. For this reason, it is possible to prevent the image recognizing means from recognizing a scratch on the container that does not move with the passage of time or the amount of the scratch on the container as the number of pixels. Therefore, the occurrence of floc can be grasped more accurately.

  As described in detail above, according to the first to third aspects of the present invention, it is possible to accurately grasp the occurrence of flocs with the lapse of production time. Moreover, according to the invention of Claim 4, the liquid food in a container without the quality fall by mixing of a flock can be manufactured. Furthermore, according to the fifth aspect of the present invention, it is possible to provide a content inspection apparatus capable of accurately grasping the occurrence of flocs with the lapse of production time.

  Hereinafter, an embodiment in which the present invention is embodied in a content inspection system for implementing a content inspection method for liquid in a container will be described with reference to FIGS.

  As shown in FIGS. 1 and 2, the content inspection system 1 includes a transfer line in which a carry-in conveyor 3, a transfer conveyor 2, and a carry-out conveyor 4 are arranged in order. The carry-in conveyor 3 conveys containers 31 filled with liquid (liquid food) to the conveyor 2 side. The transport conveyor 2 transports the containers 31 transported from the transport conveyor 3 to the transport conveyor 4 side, and the transport conveyor 4 transports the containers 31 transported from the transport conveyor 2 from the content inspection system 1. ing. In addition, each conveyor 2,3,4 of this embodiment is a belt conveyor. That is, each of the conveyors 2, 3 and 4 has a rotating shaft 7 provided at the start end and an end, an endless belt 8 wound between the rotating shafts 7 and supporting the container 31 from the bottom surface side, and a belt And a motor 9 for driving the motor 8.

  As shown in FIG. 2, a container detection sensor 10 is provided on the transport line (the transport conveyor 2 in the present embodiment). The container detection sensor 10 of the present embodiment includes a light emitting unit 10a and a light receiving unit 10b facing each other on both sides of the transport conveyor 2. The container detection sensor 10 is configured to detect the container 31 passing through the transport conveyor 2 in a non-contact state. Specifically, the container detection sensor 10 detects the container 31 when the light from the light emitting unit 10a is blocked by the container 31 and the light receiving unit 10b cannot receive light, and the container detection signal is sent to the CPU 44 of the control device 41. Output. The container detection sensor 10 does not detect the container 31 and does not output a container detection signal when the light receiving unit 10b can receive light from the light emitting unit 10a. Note that the container detection sensor 10 may detect the container 31 passing through the transport conveyor 2 in a contact state. For example, a touch sensor that detects the container 31 when the container 31 is in contact may be used, or a mechanical switch (mechanical switch) that detects the container 31 when the detection unit is pressed against the container 31 may be used.

  The content inspection system 1 includes a liquid filling device 61, a container rotating device 11, a content inspection device 21, and the like. As shown in FIG. 1, the liquid filling device 61 is provided on the upstream side of the carry-in conveyor 3 and fills an empty container 31 with liquid. A filling thermometer 62 is provided inside the liquid filling device 61. The filling thermometer 62 measures the temperature of the liquid stored in the tank in the liquid filling device 61 and outputs a temperature measurement signal to the CPU 44 of the control device 41. Moreover, the cap 31 is attached to the container 31 filled with the liquid by a plugging machine (not shown).

  As shown in FIG. 1 and FIG. 2, the container rotating device 11 has a range from the downstream part of the carry-in conveyor 3 to the upstream part of the transport conveyor 2 and from the downstream part of the transport conveyor 2 to the upstream part of the carry-out conveyor 4. Each of them is provided over the range. The container rotating device 11 is composed of a pair of belt conveyors 12, and a container 31 passing on the transport line is sandwiched between the belt conveyors 12 from the side. As shown in FIG. 2, the belts of both belt conveyors 12 move in the same direction as the direction of passage of the containers 31 on the opposing surfaces. The speed of the belt of one belt conveyor 12 is higher than the speed of the belt of the other belt conveyor 12. As a result, the container 31 passing through the container rotating device 11 rotates in the clockwise direction with respect to the axial direction of the container 31, for example. Accordingly, the liquid filled in the container 31 flows in the same direction as the rotation direction of the container 31 (in this embodiment, the clockwise direction with respect to the axial direction of the container 31). Even if the rotation of the container 31 stops after passing through the container rotating device 11, the liquid in the container 31 continues to rotate clockwise with respect to the axial direction of the container 31 due to inertia.

  The container 31 is not particularly limited as long as light (white light, infrared light, etc.) is transmitted without being irregularly reflected, and examples thereof include a glass bottle and a plastic bottle. Moreover, the liquid with which the container 31 is filled is not particularly limited as long as light is transmitted without being irregularly reflected. For example, liquid seasonings (soy sauce, soy sauce, ponzu, mirin, etc.), juice, etc. Can be mentioned.

  The contents inspection device 21 shown in FIGS. 1 and 2 is an apparatus for inspecting the contents (liquid) of the container 31. The content inspection device 21 includes a control device 41 for comprehensively controlling the content inspection system 1 as a whole. The control device 41 includes a CPU 44, a ROM 45, a RAM 46, an A / D conversion unit 42, a binarization conversion circuit 43, an input / output circuit, and the like. The CPU 44 is electrically connected to the motor 9, the liquid filling device 61, the container rotating device 11, the defective product discharging device 5, the alarm device 6 and the like of the conveyors 2, 3 and 4, and various driving signals. Control them by. The CPU 44 is connected to a keyboard 47 as input means and a display 48 as display means. The keyboard 47 is used for inputting information such as the number of containers 31 (in this embodiment, from the start of production until the 1000th) to the CPU 44 as a condition for starting production. The display 48 displays the number of containers 31 present on the transport line, the production amount of the containers 31, which container 31 has a floc exceeding the allowable amount, and the like. The display 48 may display a front view (see FIG. 1) or a plan view (see FIG. 2) in which the operation state of the content inspection system 1 is simplified. Further, the keyboard 47 and the display 48 may be omitted.

  The content inspection apparatus 21 shown in FIGS. 1 and 2 includes a container side illumination device 52 and a container bottom illumination device 53. The container side lighting device 52 is provided on the side of the central portion of the transport conveyor 2 and emits light from the side of the container 31. On the other hand, two container bottom lighting devices 53 are provided between the transport conveyor 2 and the carry-out conveyor 4 so as to irradiate light from below the container 31. The container side illumination device 52 and the container bottom illumination device 53 of the present embodiment are always lit. Here, examples of the light to be irradiated include white light and infrared rays. However, when the liquid is a transparent liquid, it is preferable to use white light, and the liquid is a dark liquid (for example, soy sauce, soup, ponzu Etc.), it is preferable to use infrared rays.

  The installation location of the container side illumination device 52 and the container bottom illumination device 53 is not limited to this embodiment. For example, the container side lighting device 52 may be installed at a position where light is irradiated from an obliquely lower side or an upper side of the container 31 instead of being installed at a position where light is irradiated from the side of the container 31. In this case, the floc is detected when the light reflected by the floc is captured by the monochrome camera 51a for floc detection. When this method is employed, the floc can be detected with high accuracy when the floc is white or light in color.

  This method is effective because flocs generated in relatively transparent products such as grain vinegar and rice vinegar are often white or light in color. On the other hand, since flocs generated in products such as black vinegar have a deep color, the transmitted illumination method as in this embodiment is effective. Thus, it is preferable to select an optimal installation location of the illumination device and the camera according to the nature and color of the floc generated in the liquid in the container 31.

  As shown in FIGS. 1 and 2, the contents inspection apparatus 21 includes two flock detection monochrome cameras 51 a (imaging means), two side-surface foreign matter detection monochrome cameras 51 b (imaging means), and 2 There are provided monochrome cameras 51c (imaging means) for detecting the bottom side foreign matter. Each of the flock detection monochrome cameras 51a and each of the side surface foreign matter detection monochrome cameras 51b are arranged to face the container side surface illumination device 52 with the transport conveyor 2 interposed therebetween. Each of the flock detection monochrome cameras 51 a is arranged at two locations along the transport direction of the transport conveyor 2. Each side-side foreign object detection monochrome camera 51b is arranged at two locations along the transport direction of the transport conveyor 2 on the downstream side of each flock detection monochrome camera 51a. As shown in FIG. 2, the angle formed by the direction in which each monochrome camera 51a, 51b is viewed (camera optical axis direction) and the transport direction of the transport conveyor 2 (horizontal direction in FIG. 2) is 60 °. Degree. The flock detection monochrome cameras 51a arranged at two positions on the upstream side face obliquely backward, and the side surface foreign matter detection monochrome cameras 51b arranged at two positions on the downstream side face obliquely forward. Each of the monochrome cameras 51a and 51b captures the liquid in the container 31 from the side of the container 31. More specifically, each of the monochrome cameras 51a and 51b captures a shadow caused by light transmitted through the container 31. In the present embodiment, one container 31 is imaged using the flock detection monochrome camera 51a located on the upstream side and the side surface foreign matter detection monochrome camera 51b located on the upstream side. Further, one container 31 is imaged using the flock detection monochrome camera 51a located on the downstream side and the side surface foreign matter detection monochrome camera 51b located on the downstream side. In addition, the monochrome cameras 51a and 51b are configured to take a liquid image with a time difference in a state where the liquid in the container 31 is flowing. More specifically, first, the liquid in the container 31 is imaged using the flock detection monochrome camera 51a located on the upstream side. Next, after a predetermined time has elapsed, the liquid in the same container 31 that has moved downstream is imaged again using the flock detection monochrome camera 51a positioned downstream. Similarly, the liquid in the container 31 is imaged using the monochrome camera 51b for detecting a foreign substance on the side surface located on the upstream side. Next, after the predetermined time has elapsed, the liquid in the same container 31 that has moved to the downstream side is imaged again using the side surface foreign matter detection monochrome camera 51b located on the downstream side. Thereby, the image of the liquid in the container 31 is taken with a time difference. Note that an image of the liquid in the container 31 may be taken using only the monochrome cameras 51a and 51b located on the upstream side or the downstream side while the conveyors 2, 3, and 4 are stopped.

  As shown in FIGS. 1 and 2, each bottom surface-side foreign object detection monochrome camera 51 c is disposed below each container bottom surface illumination device 53, and the liquid in the container 31 is discharged from below the container 31. It is supposed to take an image. In addition, the bottom-side foreign object detection monochrome camera 51c captures an image of the liquid with a time difference while the liquid in the container 31 is flowing. Specifically, first, the liquid in the container 31 is imaged using the monochrome camera 51c for detecting a foreign substance on the bottom side located on the upstream side. Next, after the predetermined time has elapsed, the liquid in the same container 31 is imaged again using the bottom surface side foreign object detection monochrome camera 51c located on the downstream side. Thereby, the image of the liquid in the container 31 is imaged with a time difference. In the state where the conveyors 2, 3, and 4 are stopped, an image of the liquid in the container 31 may be captured using only the bottom surface side foreign matter detection monochrome camera 51c located on the upstream side or the downstream side. .

  By the way, when performing the foreign substance inspection by image processing using the side surface foreign matter detection monochrome camera 51b and the bottom side foreign matter detection monochrome camera 51c, it is necessary to image the entire container 31. This is because even if only a part of the container 31 (for example, about half of the body 32 of the container 31) is imaged, it is not possible to find a foreign object at a location not imaged. This problem can be solved to some extent by taking a plurality of images while the liquid in the container 31 is flowing, but it cannot be said that it can be completely solved.

On the other hand, when performing an inspection for grasping the occurrence of the floc 71 (see FIGS. 3A and 3B) due to the image processing using the flock detection monochrome camera 51a, even if only a part of the container 31 is imaged, The generation tendency of the floc 71 can be sufficiently grasped. Therefore, the monochrome camera 51a for floc detection according to the present embodiment obtains an image with an area of 30 cm ² or less (about 15 cm ^{2 in the} present embodiment) in the body 32 of the container 31 as a visual field. In this way, the number of pixels per same area of the image can be increased as compared with the case where the entire container 31 is imaged, so that the accuracy with which occurrence of the flock 71 can be grasped is improved. Therefore, it is not necessary to use a high-sensitivity camera as the flock detection monochrome camera 51a.

  Each of the monochrome cameras 51a to 51c shown in FIGS. 1 and 2 outputs image data obtained by capturing a liquid image with a time difference to the A / D conversion unit 42 as an imaging signal. Yes. Specifically, the monochrome cameras 51a to 51c output image data obtained by imaging the liquid in one container 31 as a first imaging signal. Furthermore, the monochrome cameras 51a to 51c output image data obtained by imaging the liquid in the same container 31 after a predetermined time has passed as a second imaging signal. Each imaging signal is digitally converted by the A / D converter 42 and output to the binarization conversion circuit 43.

  The binarization conversion circuit 43 converts the image data into “0” (light detection white pixel) or “1” (light non-detection black pixel) based on the digitally converted first image pickup signal and second image pickup signal. A binarization process for binarizing is performed. Specifically, pixels that are black without detecting light due to the presence of flocs 71, bubbles 72, foreign matter 73, scratches 74 on the container 31, and the like (see the hatched portions in FIGS. 3A and 3B) Is set to “1”, and pixels that are white when light is detected because only liquid is present (see the non-hatched portions in FIGS. 3A and 3B) are set to “0”. The binarization conversion circuit 43 outputs the binarized image data to the CPU 44 as a binarization processing signal.

  In the present embodiment, the first image (see FIG. 3A) indicated by the binarization processing signal generated from the first imaging signal and the first processing indicated by the binarization processing signal generated from the second imaging signal. Between two images (see FIG. 3B), the flock 71, the bubble 72, and the foreign material 73 are moved by two pixels from the left to the right. In contrast, the scratch 74 has not moved at all. Further, the foreign matter 73 is a lump that is larger than the floc 71 and the bubbles 72. In the present embodiment, a pixel of 3 pixels or more and 34 pixels or less is counted as one minute floating substance (floc 71 or bubble 72), and a pixel of 35 pixels or more is counted as one foreign substance 73.

  The CPU 44 shown in FIG. 1 analyzes temporal changes in the image data indicated by the binarization processing signal input from the binarization conversion circuit 43, and detects fine suspended matters (floc 71 and bubbles other than the liquid in the liquid). 72). For example, the CPU 44 superimposes the first image (see FIG. 3A) and the second image (see FIG. 3B), and detects the difference between them. Specifically, when a specific pixel is “1” in both images, the pixel is “0”. As a result, the scratch 74 on the container 31 is not recognized as the number of pixels. In addition, when a specific pixel is “1” in one image and “0” in the other image, the pixel is “1”. Furthermore, when a specific pixel is “0” in both images, the pixel is set to “0”. Then, the CPU 44 recognizes the area of the minute floating substance existence region as the number of pixels (the number of pixels that become “1”). That is, the CPU 44 has a function as “image recognition means”.

  Further, the CPU 44 shown in FIG. 1 has the above-described monochromic camera 51a for detecting the flock in a state in which the floc 71 is not or hardly contained in the liquid, that is, in a state in which nothing other than the bubbles 72 is generated in the liquid. The reference pixel number is set based on the image data obtained by imaging. That is, the CPU 44 has a function as “reference pixel number setting means”. The image data for setting the reference pixel number is preferably acquired when the production is started and the production is stabilized. For example, when 100,000 containers 31 filled with liquid are produced, image data is not acquired until the 100th container after the start of production, and image data is acquired in 1000 containers 31 thereafter. The data to be acquired is preferably acquired from at least five or more containers 31, but is preferably acquired from ten or more, more preferably 50 or more containers 31. Here, “data to be acquired” is data of the number of pixels recognized from the difference between the first image (see FIG. 3A) and the second image (see FIG. 3B). The reference pixel number is calculated based on the acquired image data, but can be calculated according to required accuracy. For example, the average value of the area (number of pixels) of the micro-floating matter existing region may be used as the reference pixel number, or the maximum value of the pixel number may be used as the reference pixel number. The CPU 44 stores the set reference pixel number in the RAM 46. That is, the RAM 46 has a function as “reference pixel number storage means”.

  By the way, when the filling of the liquid into the container 31 is performed without delay, the amount of the bubbles 72 is stabilized in the same production lot. However, when the filling speed of the liquid into the container 31, the filling temperature of the liquid, the transport speed of the container 31 after filling, etc., the amount of the bubbles 72 is likely to change, and the detection accuracy of the floc 71 also decreases. End up. In particular, when the filling speed is changed, the amount of the bubbles 72 is greatly changed, so that the detection accuracy of the floc 71 is significantly reduced.

  Therefore, in this embodiment, the CPU 44 shown in FIG. 1 corrects the reference pixel number when at least one element of the filling speed, the filling temperature, and the conveyance speed changes. For example, when the filling speed changes, the CPU 44 performs the following processing. First, the amount (number of pixels) of the bubbles 72 at the filling speed at a stable time (a state where the filling speed does not change) is grasped. Since the bubbles 72 decrease as the filling speed decreases, the CPU 44 decreases the number of pixels recognized as the bubbles 72 (reference pixel number). On the other hand, since the bubbles 72 increase as the filling speed increases, the CPU 44 increases the number of reference pixels. For example, when the filling speed at the time of stabilization is 300 lines / minute, and the number of pixels recognized as the bubble 72 (reference number of pixels) is 400 pixels in that case, the filling speed is slowed down to 200 lines / minute. The reference pixel number is corrected to 300 pixels. In this case, the reduction amount of the reference pixel number is obtained by an empirical rule or data analysis based on the properties of the liquid. It is more preferable to set the amount of bubbles 72 (reference pixel number) in consideration of the time difference from filling to imaging. In the present embodiment, the filling speed is calculated based on an output when the CPU 44 drives and controls the liquid filling device 61.

  When the filling temperature changes, the CPU 44 shown in FIG. 1 performs the following processing. Since the bubbles 72 increase as the filling temperature increases, the CPU 44 increases the number of pixels recognized as the bubbles 72 (reference pixel number). On the other hand, since the bubbles 72 are reduced when the filling temperature is lowered, the CPU 44 reduces the number of reference pixels. In the present embodiment, the filling temperature is detected by the CPU 44 based on the temperature measurement signal input from the filling thermometer 62.

  Further, when the transport speed changes, the CPU 44 performs the following processing. When the transport speed is increased, the time from when the liquid is filled to when the image is picked up by the monochrome camera 51a for detecting the flock is shortened, so that the bubbles are hardly removed and vibration during the transport increases, and the bubbles 72 increase. . In this case, the CPU 44 increases the number of pixels recognized as the bubble 72 (reference pixel number). On the other hand, if the transport speed is slow, the time from when the liquid is filled to when the image is picked up by the monochrome camera 51a for detecting the flock is lengthened, so that the bubbles are considerably eliminated and the vibration during the transport is also reduced. . In this case, the CPU 44 decreases the reference pixel number. In the present embodiment, the conveyance speed is calculated based on the output when the CPU 44 drives and controls the motors 9 of the conveyors 2, 3, 4.

  The CPU 44 shown in FIG. 1 recognizes the number of pixels recognized by imaging the liquid in the container 31 currently on the transport line with the flock detection monochrome camera 51a and the number of reference pixels stored in the RAM 46. To compare. The CPU 44 determines whether or not the difference in the number of pixels exceeds the allowable amount, that is, whether or not the floc 71 exceeding the allowable amount is generated. That is, the CPU 44 has a function as “floc occurrence determination means”. In the present embodiment, the number of pixels that are abnormal as the amount of the bubbles 72 is determined in advance, and when the allowable amount is reached, the occurrence of the floc 71 is recognized. In the present embodiment, when the amount of the bubble 72 is 400 pixels, the occurrence of the floc 71 is recognized when the number of pixels is 500 pixels or more and less than 1000 pixels. The occurrence of flock 71 may be recognized when the number of pixels exceeding 450 pixels is 30% or more, or when the number of pixels exceeds 430 pixels continuously, or the number of pixels is 1. If the number of times exceeds 500 pixels, the occurrence of the floc 71 may be recognized. That is, the allowable amount can be freely set according to the accuracy required by the user, the allowable malfunction range, and the like.

  When it is determined that the difference in the number of pixels exceeds the allowable amount, the CPU 44 shown in FIG. 1 outputs a defective product detection signal to the defective product discharge device 5 and outputs an alarm signal to the alarm device 6. To do. Then, the CPU 44 operates the defective product discharge device 5 based on the defective product detection signal, and performs control to discharge the container 31 in which the occurrence of the flock 71 is detected as a defective product. At this time, the CPU 44 operates the defective product discharge device 5 in accordance with the timing at which the container 31 that has become a defective product is conveyed. The CPU 44 controls the alarm device 6 to output an alarm sound and light.

  Next, a method for producing a containerized liquid food will be described. The liquid food in the container is manufactured by performing a content inspection of the liquid in the container after performing a filling process for filling the container 31 with the liquid. Hereinafter, the contents inspection method for the liquid in the container will be described.

  When the container 31 filled with the liquid is transported and reaches the container rotating device 11, the container 31 is rotated clockwise by the container rotating device 11, and the liquid in the container 31 flows clockwise (flow). Process). After the container 31 passes through the container rotating device 11, the liquid in the container 31 flows through the flock detection monochrome camera 51a, the side surface foreign matter detection monochrome camera 51b, and the bottom surface side foreign matter detection monochrome camera 51c. In this state, an image of the liquid is taken with a time difference, and the obtained image data is output to the CPU 44 as an imaging signal (imaging process). Specifically, when the container 31 reaches the side of the container side illumination device 52, the container 31 is imaged by the flock detection monochrome camera 51a and the side surface foreign object detection monochrome camera 51b. The monochrome cameras 51 a and 51 b capture an image by capturing light transmitted from the container side illumination device 52 on the side of the container 31 through the container 31. On the other hand, when the container 31 reaches above the container bottom surface lighting device 53, the container 31 is imaged by the bottom surface side foreign object detection monochrome camera 51c. The bottom-surface-side foreign object detection monochrome camera 51c captures an image by capturing the reflected light reflected from the floc 71 included in the liquid, which is irradiated from the container bottom illumination device 53 located below the container 31. To do.

  Images captured by the monochrome cameras 51 a to 51 c are output to the A / D conversion unit 42 as imaging signals, are digitally converted, and are output to the binarization conversion circuit 43. The binarization conversion circuit 43 sets the imaging signal to “0” (light detection white pixel, refer to the non-hatched portion in FIGS. 3A and 3B), or “1” (light non-detection black pixel, 3 (a) and FIG. 3 (b) (see hatched portions), a binarized signal is generated, and the binarized signal is output to the CPU 44. Note that the A / D conversion unit 42 and the binarization conversion circuit 43 of this embodiment are provided separately from the CPU 44, but the CPU 44 may have both functions.

  Next, a process performed by the CPU 44 of the control device 41 when performing an inspection for grasping the occurrence of the floc 71 due to the image processing using the flock detection monochrome camera 51a will be described.

  In the process of step S10 shown in FIG. 4, the CPU 44 determines whether or not the liquid filling speed by the liquid filling device 61 has changed. If the filling speed has not changed (step S10: N), the CPU 44 proceeds to the process of step S20. In step S <b> 20, the CPU 44 determines whether or not the liquid filling temperature has changed based on the temperature measurement signal output from the filling thermometer 62. If the filling temperature has not changed (step S20: N), the CPU 44 proceeds to the process of step S30. In step S30, the CPU 44 determines whether or not the transport speed of the container 31 after filling has changed. If the filling speed has not changed (step S30: N), the CPU 44 proceeds to the process of step S50.

  If at least one of the filling speed, the filling temperature, and the conveyance speed has changed (steps S10, S20, S30: Y), the CPU 44 proceeds to the process of step S40. In step S40, the CPU 44 corrects the reference pixel number stored in the RAM 46, and proceeds to the process of step S50.

  In step S50, the CPU 44 determines whether or not a binarization processing signal generated based on the imaging signal from the flock detection monochrome camera 51a is input. When the binarization processing signal is not input to the CPU 44 (step S50: N), the CPU 44 ends the processing here without performing the processing of steps S60 to S110. On the other hand, when the binarization processing signal is input to the CPU 44 (step S50: Y), the CPU 44 proceeds to the process of step S60.

  In step S <b> 60, the CPU 44 analyzes the temporal change of the image data indicated by the binarization processing signal input from the binarization conversion circuit 43, and recognizes a fine suspended substance other than the liquid in the liquid. Then, the CPU 44 recognizes the amount of the fine suspended matter including the floc 71 and the bubbles 72 (area of the minute suspended matter existing region) as the number of pixels (image recognition step).

  In the subsequent step S70, the CPU 44 determines whether or not production is started. Specifically, the CPU 44 determines whether or not the number of input of the container detection signal from the container detection sensor 10 has reached a predetermined number (“1000” in the present embodiment). Note that the method for determining whether or not production is started is not limited to the method using the container detection sensor 10 described above. For example, a timer that starts counting when the first container 31 is filled with liquid is provided, and the CPU 44 determines whether or not the time from the start of production until the 1000 containers 31 pass has elapsed. It may be. If it is time to start production (step S70: Y), the CPU 44 proceeds to the process of step S80. In step S80, the CPU 44 stores the number of pixels recognized in step S60 as the reference pixel number in the RAM 46 (reference pixel number setting step), and ends the processing here.

  On the other hand, when it is not the time of production start (step S70: N), CPU44 transfers to the process of step S90. In step S90, the CPU 44 compares the number of pixels recognized in step S60 with the reference number of pixels stored in the RAM 46 in step S80. In step S100, the CPU 44 determines whether or not the difference in the number of pixels exceeds the allowable amount, that is, whether or not the floc 71 exceeding the allowable amount is generated (floc generation determining step). When the difference in the number of pixels does not exceed the allowable amount (step S100: N), the CPU 44 ends the process here without performing the process of step S110. On the other hand, when the difference in the number of pixels exceeds the allowable amount (step S100: Y), the CPU 44 proceeds to the process of step S110. In this embodiment, the foreign matter 73 is detected later by the foreign matter detection monochrome cameras 51b and 51c. However, if only the lower limit value of the number of pixels is set as a reference when determining the occurrence of the floc 71 exceeding the allowable amount, In the floc occurrence determination step, the occurrence of the floc 71 and the occurrence of the foreign material 73 are collectively detected. Therefore, in this embodiment, not only the lower limit value of the number of pixels but also the upper limit value is set as a reference when determining the occurrence of the floc 71 exceeding the allowable amount. Specifically, when the reference pixel number is 400 pixels, it is determined that the floc 71 is generated when the pixel number is 500 pixels or more and less than 1000 pixels.

  In step S <b> 110, the CPU 44 outputs a defective product detection signal indicating a defective product to the defective product discharge device 5 and outputs an alarm signal to the alarm device 6. As a result, the defective product discharge device 5 discharges the container 31 in which the occurrence of the flock 71 is detected as a defective product, and the alarm device 6 outputs an alarm sound or light.

  Next, processing performed by the CPU 44 of the control device 41 when performing foreign matter inspection by image processing using the side surface foreign matter detection monochrome camera 51b and the bottom side foreign matter detection monochrome camera 51c will be described.

  In the process of step S210 shown in FIG. 5, is the CPU 44 input a binarization processing signal generated based on the imaging signals from the side surface foreign matter detection monochrome camera 51b and the bottom side foreign matter detection monochrome camera 51c? Determine whether or not. When the binarization processing signal is not input to the CPU 44 (step S210: N), the CPU 44 ends the processing here without performing the processing of steps S220 to S240. On the other hand, when the binarization processing signal is input to the CPU 44 (step S210: Y), the CPU 44 proceeds to the process of step S220.

  In step S220, the CPU 44 analyzes the temporal change of the image data indicated by the binarization processing signal input from the binarization conversion circuit 43 and recognizes the foreign matter 73 in the liquid. Then, the CPU 44 recognizes the amount of the foreign matter 73 (the area of the foreign matter existing area) as the number of pixels.

In subsequent step S230, CPU 44 determines whether the number of pixels recognized in step S220 is 1000 pixels or more. When the number of pixels is less than 1000 pixels (step S230: N), the CPU 44 determines that the foreign substance 73 does not exist in the liquid, and ends the process here without performing the process of step S240. On the other hand, when the number of pixels is 1000 pixels or more (step S230: Y), the CPU 44 determines that the foreign substance 73 exists in the liquid, and proceeds to the process of step S240. In step S240, the CPU 44 outputs a defective product detection signal indicating a defective product to the defective product discharge device 5 and outputs an alarm signal to the alarm device 6. As a result, the defective product discharge device 5 discharges the container 31 in which the occurrence of the flock 71 is detected as a defective product, and the alarm device 6 outputs an alarm sound or light.
(Example 1)

  A more specific example 1 carried out to confirm the effect of the present invention will be described below.

  First, produce 100,000 bottles of rice vinegar (acidity 4.5%) in a container 31 (glass bottle) 72 mm in diameter and 201 mm in height, and transport them at a filling speed and transport speed of 400 bottles / min. To do. At this time, the liquid in the container 31 is imaged using a plurality of flock detection monochrome cameras 51a (see FIGS. 1 and 2).

Note that the content that is captured and recorded is the number of micro suspended matter that is counted after the image data is binarized and specified as a substance other than a liquid. In the present embodiment, the average state of occurrence of microscopic suspended matters is recorded in the liquid in the 1000 containers 31 at the start of production. Specifically, the CPU 44 calculates an average value of the number of pixels indicating the amount of the fine suspended matter in the 1000 containers 31 and stores it in the RAM 46. Here, a monochrome camera 51a for floc detection with 1.4 million pixels was used as the imaging means, and high-intensity white illumination was used as the container side illumination device 52 that transmits light from the side of the container 31. The imaging range is a 15 cm ² (5 cm × 3 cm) range of the body 32 of the container 31 (that is, about 1.32 million pixels), and one micro floating object having an area of 3 pixels (area 0.0034 mm ² ) or more. I counted it.

  As a result, the number of fine suspended matters stored at 1000 at the start of production (only the bubbles 72 at the start of production) was 62 to 137, an average of 93. For this reason, in the production products after 1000 from the start of production, when 140 or more minute floating objects are counted, it is set to output a defective product detection signal and an alarm signal as the occurrence of the floc 71. . In the first embodiment, in order to facilitate the generation of the floc 71, the level of the liquid level of the tank in the liquid filling device 61 is adjusted in advance by maintaining it for a certain period of time. In the latter case (approximately 48000th in the present embodiment), a container 31 in which 150 or more micro suspended matters were counted was detected. Here, when the liquid in the container 31 on the conveyor 2 was confirmed, it was confirmed that the floc 71 was generated in the liquid. After this, the frequency of occurrence of floc 71 in the liquid gradually increased.

In the first embodiment, the transport speed is set to 400 lines / minute, but the transport speed may be faster or slower. In the first embodiment, the state of the liquid in the 1000 containers 31 at the start of production is recorded. However, the number of containers 31 is not limited to 1000, and may be increased or decreased depending on the type of liquid. May be. Further, in the first embodiment, the imaging is performed by the monochrome cameras 51a to 51c having 1.4 million pixels and the light is transmitted from the side or the lower side of the container 31, and the imaging range is 15 cm ^2. However, the present invention is limited to these. The conditions may be changed depending on the container 31 and the type of liquid.
(Example 2)

  Next, more specific Example 2 carried out to confirm the effect of the present invention will be described below.

  First, 100,000 bottles of rice vinegar (acidity 4.5%) in the same container 31 as in Example 1 are produced at a filling speed and a conveyance speed of 300 bottles / min. At this time, the liquid in the container 31 is imaged using a plurality of flock detection monochrome cameras 51a (see FIGS. 1 and 2). The contents to be imaged and recorded, the imaging means, the container side illumination device 52, the imaging range, and the like are the same as in the first embodiment.

  As a result, the number of fine suspended matters stored at 1000 at the start of production (only the bubbles 72 at the start of production) was 53 to 104, and an average of 72. Then, considering the case where the speed of the conveyor 2 changes in the middle, the equation of Y = X− (t−5) × 14 is set based on the relationship between the values X and Y calculated in advance (ROM 45 Remembered). Here, X indicates the number of micro suspended matters (bubbles 72) at the start of production ("110" in Example 2), and is set from the number of micro suspended matters (53 to 104). Y indicates a threshold value at which a defective product detection signal and an alarm signal are output as the occurrence of the flock 71. Furthermore, t indicates the time (seconds) from filling the liquid to starting the content inspection. However, when the threshold Y is a numerical value of 10 or less, the threshold Y is set to 10.

  In the second embodiment, in order to facilitate the generation of the floc 71, the liquid level of the tank in the liquid filling device 61 is adjusted in advance by, for example, maintaining the liquid level for a certain time. After that (about 36000th in the present embodiment), a container 31 in which micro suspended matters were counted more than a threshold was detected. Here, when the liquid in the container 31 on the conveyor 2 was confirmed, it was confirmed that the floc 71 was generated in the liquid. After this, the frequency of occurrence of floc 71 in the liquid gradually increased.

In the second embodiment, the conveyance speed is set to a maximum of 300 lines / minute, but the conveyance speed may be faster or slower. In the second embodiment, the state of the liquid in the 1000 containers 31 at the start of production is recorded. However, the number of containers 31 is not limited to 1000, and may be increased or decreased depending on the type of liquid. May be. Further, in the second embodiment, the image is picked up by the monochrome camera 51a for floc detection with 1.4 million pixels, the light is transmitted from the side or the lower side of the container 31, and the image pickup range is 15 cm ^2. The conditions may be changed depending on the type of the container 31 and the liquid. In addition, although the calculation formula is used in the second embodiment, this is an optimal calculation formula only in the case of the second embodiment. Even if the calculation formula is appropriately changed according to the condition and an optimal threshold value is set. Good.

  Therefore, according to the present embodiment, the following effects can be obtained.

  (1) In the contents inspection method for liquid in a container according to the present embodiment, the number of pixels recognized in step S60 is compared with the reference number of pixels stored in the RAM 46 in step S80, and the determination in step S100 is performed. It is possible to accurately grasp the occurrence of the floc 71 as the time elapses. Therefore, it is possible to manufacture a liquid food in a container that does not deteriorate in quality due to the flock 71 being mixed.

  (2) By the way, when the foreign matter 73 is generated in the liquid, it is required to set the accuracy strictly because one container 31 filled with the liquid cannot be adopted as a product. However, in the present embodiment, when a floc 71 (see FIGS. 3A and 3B) whose number of pixels is considerably smaller than that of the foreign material 73 is generated in the liquid, the increase state of the floc 71 is confirmed. Thus, the defective product discharge device 5 discharges the defective product and the alarm device 6 outputs an alarm sound and light. Therefore, not only can the floc 71 in the liquid be checked, but also the foreign matter 73 in the liquid can be checked reliably.

  In addition, you may change embodiment of this invention as follows.

  The CPU 44 of the above embodiment determines that the floc 71 is detected when the difference between the number of pixels recognized based on the imaging signal from the flock detection monochrome camera 51a and the reference number of pixels stored in the RAM 46 reaches an allowable amount. Was aware of the occurrence of However, the CPU 44 may recognize the occurrence of a foreign substance 73 larger than the flock 71 and the bubble 72 when the difference in the number of pixels further increases to a specific amount. For example, when the amount of the bubble 72 is 400 pixels, when the number of pixels is 500 pixels or more (the difference in the number of pixels is 100 pixels or more), the occurrence of the flock 71 or the foreign matter 73 is recognized, and the number of pixels is 1000 pixels or more. If the difference is 600 pixels or more, the generation of the foreign material 73 may be recognized. Specifically, a process for determining whether or not the number of pixels is 1000 pixels or more is added after the process of step S50 shown in FIG. If the number of pixels is less than 1000 pixels, the processes of steps S60 to S110 are executed. If the number of pixels is 1000 pixels or more, the processes of steps S220 and S240 shown in FIG. 5 are executed in order. .

  In this way, the occurrence of the foreign matter 73 in addition to the flock 71 can be detected by the flock detection monochrome camera 51a. Therefore, when the content inspection apparatus 21 is configured by providing the foreign substance inspection apparatus with a function of detecting the occurrence of the flock 71, it is only necessary to change the program, and the existing configuration is not increased without increasing the number of cameras. Can be used as is.

  In the above embodiment, when the floc 71 is detected, the CPU 44 performs control to discharge the defective product by the defective product discharge device 5 and performs control to output an alarm sound or light from the alarm device 6. (See step S70 in FIG. 4). However, when the floc 71 is detected, the CPU 44 may perform control to slow down the liquid filling speed by the liquid filling device 61 or completely stop the transport line (the motors of the conveyors 2, 3 and 4). 9) may be controlled.

  In the above embodiment, the container 31 filled with the liquid is rotated by the container rotating device 11 to cause the liquid to flow in the container 31. However, the container 31 may not be rotated by the container rotating device 11. . This is because even if the container 31 is not rotated by the container rotating device 11, it is highly likely that the liquid naturally flows immediately after the liquid is filled or during the conveyance of the container 31.

  -You may change suitably the number, combination, etc. of the monochrome cameras 51a-51c in the said embodiment. For example, two monochromatic cameras for detecting flock 51a and two monochromatic cameras for detecting side-face foreign matter 51b may be arranged along the conveying direction of the conveying conveyor 2. In this case, since the monochrome cameras 51a and 51b respectively image the upper and lower portions of the trunk portion 32 of the container 31 and determine the occurrence of the floc 71 for each, the occurrence of the floc 71 with the passage of production time is more accurately detected. I can grasp. In addition, the camera optical axis direction of each monochrome camera 51a, 51b and the conveyance direction of the conveyor 2 are orthogonal to each other, and the flock detection monochrome camera 51a group and the side-side foreign object detection monochrome camera 51b group are arranged above and below, respectively. Also good.

  In the above embodiment, the monochrome cameras 51a to 51c are used as the imaging unit, but a color camera or the like may be used as the imaging unit. The imaging unit may be a camera using a tube-type imaging tube, or a camera using a solid-state image sensor that converts an image into a first imaging signal and a second imaging signal. Examples of the solid-state image sensor include a MOS type solid-state image sensor and a CCD (Charge Coupled Device).

  In the above embodiment, the binarization process is performed on the image data in the pre-processing in order to improve the image recognition accuracy of the minute suspended matter. However, the image recognition accuracy of the micro suspended matter may be improved by performing processing such as filtering, smoothing, and sharpening on the image data.

  Next, in addition to the technical ideas described in the claims, the technical ideas grasped by the embodiments described above are listed below.

  (1) The content inspection method for liquid in a container according to any one of claims 1 to 3, wherein a flow step of flowing the liquid in the container is performed before the imaging step.

  (2) In any one of Claims 1 to 3, when it is determined in the floc occurrence determination step that a floc exceeding an allowable amount is generated, an alarm execution step is provided for causing an alarm means to issue an alarm. A method for inspecting the contents of a liquid in a container.

  (3) The defective product according to any one of claims 1 to 3, wherein the defective product discharging means discharges a defective product when it is determined in the flock generation determining step that a floc exceeding an allowable amount is generated. A method for inspecting the contents of a liquid in a container, comprising a discharging step.

  (4) In any one of claims 1 to 3, when it is determined in the flock generation determination step that a flock exceeding an allowable amount is generated, the transfer line that stops the transfer line that transfers the container A method for inspecting the contents of a liquid in a container, comprising a stopping step.

  (5) In Claim 5, the container is transported in a state where the liquid is rotated with respect to the axial direction of the container, and the imaging means images the liquid from the side or the lower side of the container. A device for inspecting the contents of a liquid in a container.

  (6) The content inspection apparatus for a liquid in a container according to claim 5, wherein a plurality of the imaging means are provided.

  (7) Imaging means for outputting image data obtained by taking an image of the liquid with a time difference in a state where the liquid in the container is flowing, and analyzing temporal changes in the image data to analyze the liquid in the liquid Image recognizing means for recognizing a minute suspended matter other than liquid and recognizing the amount of the minute suspended matter including bubbles and flocs as the number of pixels, and imaging with no or little flocs contained in the liquid Based on the obtained image data, reference pixel number setting means for setting a reference pixel number, reference pixel number storage means for storing the reference pixel number set by the reference pixel number setting means, and the image recognition means The number of pixels recognized by the reference pixel number is compared with the number of reference pixels stored in the reference pixel number storage means to determine whether or not a floc exceeding an allowable amount has occurred. The reference pixel number setting means includes at least one element of a filling speed of the liquid into the container, a filling temperature of the liquid, and a transport speed of the container after filling. And setting the reference number of pixels.

(8) Imaging means for outputting image data obtained by taking an image of the liquid with a time difference in a state where the liquid in the container is flowing, and analyzing temporal changes in the image data to analyze the liquid in the liquid Image recognizing means for recognizing a minute suspended matter other than liquid and recognizing the amount of the minute suspended matter including bubbles and flocs as the number of pixels, and imaging with no or little flocs contained in the liquid Based on the obtained image data, reference pixel number setting means for setting a reference pixel number, reference pixel number storage means for storing the reference pixel number set by the reference pixel number setting means, and the image recognition means The number of pixels recognized by the reference pixel number is compared with the number of reference pixels stored in the reference pixel number storage means to determine whether or not a floc exceeding an allowable amount has occurred. Occurrence determination and means, the imaging means, the container contents inspection device containers liquid and obtaining an image of 30 cm ² or less in area as the field of view in the body of the.

  (9) Container flow means for flowing the liquid in the container, imaging means for outputting image data obtained by taking an image of the liquid with a time difference while the liquid is flowing, and time of the image data Image recognition means for recognizing a microscopic suspended matter other than the liquid in the liquid by analyzing a change in the liquid, and recognizing the amount of the microscopic suspended matter including bubbles and flocs as the number of pixels, and there is no floc in the liquid. Or a reference pixel number setting means for setting a reference pixel number on the basis of the image data obtained by imaging in a state of being hardly included, and a reference for storing the reference pixel number set by the reference pixel number setting means The number of pixels storing means, the number of pixels recognized by the image recognizing means, and the reference pixel number stored in the reference pixel number storing means are compared, and the floc exceeding the allowable amount Contents inspection system for containers liquid; and a determining floc occurrence determination means for determining whether occurring.

The front view which shows the content inspection system of this embodiment. The top view which shows a content inspection system. (A) is a figure which shows the 1st image which the binarization processing signal produced | generated from the 1st imaging signal shows, (b) shows the 2nd image which the binarization processing signal produced | generated from the 2nd imaging signal shows Figure. The flowchart which shows the process performed by CPU of a control apparatus. The flowchart which shows the process performed with CPU of a control apparatus.

Explanation of symbols

21 ... Content inspection device 31 ... Container 32 ... Body 44 ... CPU as image recognition means, reference pixel number setting means, and floc occurrence determination means
46 ... RAM as reference pixel number storage means
51a ... Monochrome camera for detecting flock as imaging means 51b ... Monochrome camera for detecting side foreign matter as imaging means 51c ... Monochrome camera for detecting foreign matter on bottom side as imaging means 71 ... Flock 72 ... Bubble

Claims (5)

An imaging step of capturing an image of the liquid with a time difference in a state where the liquid in the container is flowing, and outputting the obtained image data;
An image recognition step of recognizing a minute suspended matter other than the liquid in the liquid by analyzing the temporal change of the image data and recognizing the amount of the minute suspended matter including bubbles and flocs as the number of pixels;
A reference pixel number setting step of setting a reference pixel number based on the image data obtained by imaging in a state where no or almost no floc is contained in the liquid;
A flock generation determination step of comparing the number of pixels recognized in the image recognition step with the reference pixel number set in the reference pixel number setting step to determine whether or not a flock exceeding an allowable amount has occurred. A method for inspecting the contents of a liquid in a container, comprising:   In the reference pixel number setting step, the reference pixel number is determined in consideration of at least one element of a filling speed of the liquid into the container, a filling temperature of the liquid, and a transport speed of the container after filling. The content inspection method for liquid in a container according to claim 1, wherein the content is set. 3. The method for inspecting a content of a container-filled liquid according to claim 1 or 2, wherein in the imaging step, an image is obtained with an area of 30 cm < ² > or less in the trunk of the container as a visual field.   After performing the filling process which fills a container with the liquid, the test | inspection by the content inspection method of the liquid in a container of any one of Claim 1 thru | or 3 is performed, and a liquid food in a container is manufactured. A method for producing liquid food in a container. Imaging means for outputting image data obtained by taking an image of the liquid with a time difference in a state where the liquid in the container is flowing; and
An image recognition means for recognizing a minute suspended matter other than the liquid in the liquid by analyzing temporal change of the image data, and recognizing the amount of the minute suspended matter including bubbles and flocks as the number of pixels;
A reference pixel number setting means for setting a reference pixel number based on the image data obtained by imaging in a state in which no or almost no floc is contained in the liquid;
Reference pixel number storage means for storing the reference pixel number set by the reference pixel number setting means;
A block that compares the number of pixels recognized by the image recognition unit with the reference pixel number stored in the reference pixel number storage unit to determine whether or not a floc exceeding an allowable amount has occurred. An apparatus for inspecting the contents of a liquid in a container, characterized by comprising generation determination means.

Images