JP3985946B2 - Foreign matter inspection device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a foreign matter inspection apparatus suitable for reliably inspecting foreign matter that may be mixed in tobacco leaves, board gum, and the like.
[0002]
[Related background]
A foreign substance inspection apparatus that inspects the contamination of foreign substances in food materials and semi-finished foods and uses the foreign substances to eliminate such foreign substances exclusively detects foreign substances by paying attention to the color of the inspection object. Specifically, a color component of a color image obtained by imaging the inspection object is obtained, and the color component is a normal color component (first inspection color information) of the inspection object, or the inspection object The foreign matter is detected by determining whether or not the foreign color component (second inspection color information) possessed by the foreign matter that may be mixed in is contained.
[0003]
By the way, the first inspection color information specifies a color component having a high appearance frequency as a normal color as much as possible by visually inspecting a color image of a normal inspection object that does not include foreign matters in advance. This is done by setting the normal color component area of the object. At this time, it is necessary to exclude a color component that may be included in a foreign object even if it is a normal color component having a high appearance frequency from the setting target. Further, the second inspection color information is to visually inspect a color image of a foreign material in advance, thereby specifying a color component having a high appearance frequency as an abnormal color as much as possible and setting a color component region of the abnormal color that the foreign material has. Made by. At this time, even if the color component is an abnormal color having a high appearance frequency, it is necessary to exclude the normal color of the inspection object from the setting target.
[0004]
[Problems to be solved by the invention]
However, in order to set the first and second inspection color information described above, it is necessary to prepare a color image of only a normal inspection object that does not include foreign matter, or to prepare a color image of foreign matter. It cannot be denied that a great deal of labor and labor is required for the setting work. In addition, after accurately grasping the relationship between the variation range (distribution tendency) of the color component (normal color) of the inspection object and the variation range of the color component (abnormal color) of the foreign object, the setting work is performed. It is necessary to do.
[0005]
By the way, food materials such as cereals, semi-finished products thereof, and tobacco leaves are agricultural products and have various color components depending on the variety, production area, production time, and the like. In addition, in kneaded products such as board gum, additives such as sugar may remain in a solid state, and their color components are not uniform. Some foreign matters that may be mixed into the inspection target have a color component similar to the inspection target, and some have a color component included in the range of normal color components as the inspection target. . Therefore, it is very difficult to set the first and second inspection color information with high accuracy according to the inspection object, and this is a factor that hinders high-accuracy foreign matter inspection.
[0006]
The present invention has been made in consideration of such circumstances, and its purpose is to have a normality of the inspection object even when the color component of the normal inspection object is similar to the color component of the foreign object. Accurate color components (color component regions) and color components (color component regions) possessed by foreign substances that may be mixed into the inspection object are accurately defined as inspection color information, and using this inspection color information, for example, An object of the present invention is to provide a foreign matter inspection apparatus capable of reliably detecting foreign matter mixed in an inspection object such as tobacco leaves and board gum.
[0007]
[Means for Solving the Problems]
In order to achieve the above-described object, the foreign substance inspection apparatus according to the present invention includes an imaging unit that captures a color image of an inspection object made of, for example, tobacco leaves, food raw materials or semi-finished products thereof, and performs image processing on the color image. To reliably detect foreign matter mixed in the inspection object,
Based on the appearance frequency of each color component in the color image of the inspection object in advance, a normal color component area of the inspection object is obtained (normal color area detecting means), and an area outside the normal color color component area is obtained. A color component (abnormal color) is obtained and set as first inspection color information (first inspection color information setting means), and a predetermined appearance frequency based on the appearance frequency of each color component in the color image of the inspection object. The color components included in the range are collected, and the color component area is set as second inspection color information indicating a color component that may be a foreign object (second inspection color information setting unit).
[0008]
Next, the color image of the inspection object obtained by the imaging unit is color-selected and specific images corresponding to the first and second inspection color information respectively registered by the first and second inspection color information setting units. A color component image area is obtained (color discrimination means), and the size of the image area of the specific color component is determined to detect foreign substances contained in the inspection object (foreign substance detection means).
[0009]
In a preferred aspect of the present invention, the normal color area detecting means obtains an appearance frequency distribution of each color component in the color image of the inspection object, and a color component area that forms a single unit having a predetermined frequency threshold or more from the appearance frequency distribution Is obtained as a normal color region of the inspection object (claim 2). Further, the second inspection color information setting means determines that a color component area is within a predetermined frequency threshold range including the normal color area from the appearance frequency distribution of each color component in the color image of the inspection object including the foreign object. This is realized as what is obtained as a color component area including the color component possessed.
[0010]
Incidentally, the second inspection color information is obtained as one or a plurality of color component areas by changing the predetermined frequency threshold range. Further, the color image is obtained by color separation into, for example, three primary color components composed of R (red), G (green), and B (blue), and the appearance frequency distribution of each color component of the color image is determined by each color in the RGB three-dimensional table. It is calculated | required by each counting the registration frequency of a component (Claim 5).
[0011]
Of course, each color component in the color image can be obtained as a wavelength component in the near-infrared region, for example (claim 6).
[0012]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, a foreign substance inspection apparatus according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram illustrating a schematic configuration of a main part of a foreign matter inspection apparatus according to this embodiment. Reference numeral 1 denotes a camera that captures an image of the inspection object S and obtains a color image thereof. It is a light source (fluorescent lamp). In addition, for example, the tobacco leaf as the inspection object S is composed of a plurality of types of tobacco in which tobacco leaf is preliminarily cut as a leaf piece of a predetermined width, which is combined (blended) according to the brand of the cigarette. .
[0013]
Such a foreign matter inspection apparatus that inspects foreign matter mixed in the cut leaves (inspection object) S is incorporated in a conveyance line including the belt conveyor 3 that conveys the cut leaves (inspection object) S. Then, the tobacco leaves placed on the belt conveyor 3 and transported are imaged from above the belt conveyor 3 using the camera 1 and the color image is taken into an image processing device 4 comprising a computer (PC). By performing image processing based on color components to be described later, it is configured to detect foreign matter mixed in the tobacco leaf. Incidentally, the image of the tobacco leaf by the camera 1 is performed by sequentially taking in a belt-like region over the entire width of the belt conveyor 3 as one color image, as shown by the shaded portion 5 in FIG. The color image is given to the image processing device 4 as a color signal separated into three primary colors composed of, for example, R (red), G (green), and B (blue).
[0014]
The image processing apparatus 4 including a computer includes a color selection function 11 that obtains a color component for each of a plurality of pixels (picture elements) constituting a color image, for example, as shown in FIG. In particular, the color components selected by the color selection function 10 are sequentially registered in the three-dimensional table 11 having the respective RGB color components as axes (parameters), and the number of registrations of each color component is applied to all the pixels constituting the color image. And a frequency counting function 13 that obtains the appearance frequency and obtains the appearance frequency and registers it in the frequency distribution table 12 to obtain the appearance frequency distribution.
[0015]
Then, the inspection color setting function 14 obtains an inspection color for detecting the foreign matter mixed in the inspection object S based on the appearance frequency distribution of each color component registered in the frequency distribution table 12 as will be described later, and the inspection object Inspection color information indicating a color component (unusual color component) that S does not have in the first inspection color table 15 and inspection color information indicating a color component of a foreign substance mixed in the inspection object S (approximate color component) Are registered in the second inspection color tables 16a to 16n, respectively.
[0016]
In other words, the inspection color setting function 14 obtains a group of color components having a high appearance frequency of color components obtained from the color image of the inspection object S including foreign matter as a normal color component area of the inspection object S. Normal color area detection means and first inspection color information indicating a color component (abnormal color) that the inspection object S does not have a color component of the area (complementary area) out of the normal color component area A first inspection color information setting unit that obtains A and registers the inspection color information A in the first inspection color table 15 is provided.
[0017]
The inspection color setting means 14 is also a color component of the inspection object S based on the appearance frequency distribution of each color component registered in the frequency distribution table 12 of the color image of the inspection object S including the foreign matter. A second inspection in which an ambiguous color component (approximate color) region that is also a color component of a foreign object is obtained as second inspection color information B and registered in the second inspection color table 16 (16a, 16n). Color information setting means is provided. Note that a plurality of second inspection color information B can be set with different color component areas. In this case, the second inspection color information Ba,... Bn may be registered in a plurality of second inspection color tables 16a,.
[0018]
The setting (registration) of the first and second inspection color information A, Ba, to Bn in the inspection color tables 15, 16a, to 16n is, for example, a preparation process (initial stage when the foreign matter inspection apparatus is started up. This is executed as a setting process. In the case where batch inspection is performed for a plurality of types of inspection objects, the inspection color information A, Ba, to Bn may be automatically set at the start of each batch.
[0019]
At this time, in each of the inspection color tables 15, 16a to 16n, the inspection color information A, Ba, to Bn is a size when the inspection color information A, Ba,. Information on the size when pixels having Ba,..., Bn exist as pixel areas each having a certain area is simultaneously registered as a threshold value for foreign object determination. This size information (threshold value) is optimally set by a visual inspection result with respect to the foreign matter or a subsequent learning process.
[0020]
As described above, if inspection color information A, Ba,... Bn is set in the first and second inspection color tables 15, 16a,. Using the color image of the object S, an inspection is performed as to whether or not the inspection object S contains foreign matter. The execution of this inspection uses the color comparison function 17, and the color components of the color image of the inspection object S obtained by the color selection function 10 are the first and second inspection color tables 15, 16a, as described above. It is made by comparing and collating whether or not the inspection color information A, Ba,... When pixels (picture elements) of color components corresponding to the first and second inspection color information A, Ba,... Bn are detected, the same inspection color information (color component region) is detected using the pixel integration function 18. The pixels adjacent to each other in the group are integrated as a single unit, and the size (area) is determined using the size determination function 19 to detect foreign matters mixed in the inspection object S. ing.
[0021]
In this foreign object detection, the size (area) of the cluster of pixels integrated using the pixel integration function 18 exceeds a predetermined size (threshold value) preset in the inspection color tables 15, 16a to 16n. Is determined by determining whether or not the size is abnormal, that is, whether or not the size of the normal inspection object S is deviated. As a result of such determination, not only the detection object S having a color component other than the normal color area included in the first inspection color information A is detected as a foreign object, but also the normal color area of the inspection object S. It is possible to reliably detect a foreign object having an ambiguous color component included in the.
[0022]
The setting of the first and second inspection color information A, Ba,... Bn and the foreign object detection using these inspection color information A, Ba,. For example, the tobacco leaves as the inspection object S are processed products of agricultural products, and generally have similar color components, but are not uniform, and are distributed over a certain color component range. However, it is extremely difficult to accurately define the boundary that specifies the normal color of the tobacco leaf, that is, the range of the color component. In addition, some of the foreign substances that may be mixed in the tobacco leaves are the same as the tobacco leaves, such as the tobacco leaf pieces that have been leaked and the trash that has adhered to the containers used to transport the tobacco leaves. Some have such color components.
[0023]
However, the appearance frequency of the color component in the inspection object (tobacco leaf) S has a distribution of spreading around a certain color component as schematically shown in FIG. 3, for example. Even if foreign matter is mixed in the inspection object S, the appearance frequency distribution of the color component does not change significantly. Therefore, in the inspection color setting function 14 described above, a threshold value C1 is set for the appearance frequency of each color component, for example, the appearance frequency is [50 pixels] or more, and the color component having the appearance frequency exceeding the threshold value C1. Is set as a normal color region of a normal inspection object (cigarette leaf) S. In other words, based on the appearance frequency of each color component in the color image obtained by imaging the inspection object (cigarette leaf) S, it is defined as a normal color region T that the normal inspection object (cigarette leaf) S can surely take. is doing. Then, the color component area other than the normal color area T is set as an abnormal color area F which may be a foreign substance, and information on the abnormal color area is registered in the table 15 as first inspection color information A. .
[0024]
For example, as shown in FIG. 4A, the normal color region T and abnormal color region F defined (set) in this way are arranged on a two-dimensional table focusing only on the RB component. One region T to be formed and its complementary region F are respectively represented. Although not shown in the drawing, the normal color region T is represented as a one-dimensional one-color component region on the RGB three-dimensional table.
[0025]
However, as described above, some of the color components possessed by the foreign matter are included in the normal color region T, and even if the color component is a normal inspection object (cigarette leaf), the color component is abnormal. Some are included in the color region F. Accordingly, the inspection color setting function 14 pays attention to the above-described color component appearance frequency distribution, sets two different appearance frequency thresholds C2 and C3, and sets one color component within the range of the appearance frequency thresholds C2 and C3. The color components (approximate colors) which may be collected as one group are defined as one group, and this information is registered in the table 16 as the second inspection color information B.
[0026]
Specifically, when the frequency threshold C1 defining the normal color range is set to [50 pixels], the appearance frequency thresholds C2 and C3 are set as [40 pixels] and [60 pixels], respectively, and the appearance frequencies thereof are set. All the color components included in the range are defined as a region G of color components (approximate colors) that may be foreign substances as shown in FIG. Among the color components included in the color component (approximate color) region G set in this way, the hue (distance between the color components) is greatly separated, and it is difficult to say that the colors are similar. There is also. However, the region G is defined as an approximate color component (region) that forms a group (group) of all the color components obtained as described above apart from the general concept of approximation between hues.
[0027]
In addition, such an approximate color region G is obtained by changing the above-described appearance frequency thresholds C2 and C3 according to the color component of the foreign object, for example, a color having an appearance frequency in the range of [500 pixels] to [1000 pixels]. It can also be set as a component region. Therefore, if a plurality of approximate color regions G are defined in accordance with the color components of foreign matter that may be mixed into the inspection object S, the normal color range of the inspection object (cigarette leaf) S in which the color components of the foreign object are normal Even if it is included in T, the color component can be grasped as information that may be a foreign object.
[0028]
Thus, according to the foreign matter inspection apparatus configured as described above, the normal color range (color component region) of the normal inspection object S is easily set, and the color range of the foreign matter (color component region) is set. Can be set easily. In particular, even when these color ranges overlap each other, the abnormal color region F and the approximate color region G are set based on the appearance frequency of each color component, so that they are included in these regions F and G. When a color component is detected, it can be determined (detected) as the possibility of a foreign object.
[0029]
Then, the pixels having the color components contained in the respective regions F and G are integrated between adjacent pixels to form a group of pixels of the same approximate color, and the detected inspection color information A, Ba,... Bn Since the size of the group (image region) is determined every time, it is possible to reliably detect a foreign substance having a certain size mixed in the inspection object S. In addition, even when the color of the foreign object is included in the normal color range of the inspection object S, the color component is not defined as the color of the inspection object S or the color of the foreign object. The presence of the foreign matter can be reliably detected. That is, whether or not a pixel having a color component that is likely to be a foreign object forms a group (area) of an image area having a predetermined size indicates whether or not the color component is a foreign object. Since the determination is made, it is possible to reliably detect the foreign matter mixed in the inspection object S.
[0030]
The present invention is not limited to the embodiment described above. For example, the inspection object S may be an agricultural product such as cereal or a processed product thereof, or may be a refined food such as a plate gum. By the way, in the case of refined foods such as board gum, when food additives such as sugar remain as a solid, this can be detected as a foreign substance. Further, when obtaining the color image of the inspection object S, when the surface of the belt conveyor 3 is captured as the background image, the color component of the background image is obtained and the color component is subjected to image processing for foreign object detection. Should be removed.
[0031]
When color-selecting a color image, for example, each color component of RGB may be decomposed into 256 gradations to obtain the color component. Further, the threshold values C1, C2, and C3 for the appearance frequency of color components may be set according to the total number of pixels of one color image, the content rate of foreign matter, and the like. Further, here, an example in which the color component is discriminated by focusing on the RGB component of the color image has been described. However, for example, focusing on the wavelength components λ1, λ2,... It is of course possible to identify them. In this case, an infrared camera may be used as the camera 1. It is also possible to perform color discrimination by paying attention to the spectrum of each pixel. In addition, the present invention can be variously modified and implemented without departing from the scope of the invention.
[0032]
【The invention's effect】
As described above, according to the present invention, even if the color of the foreign matter mixed in the inspection object is similar to the color of the normal inspection object, each color component in the color image obtained by imaging the inspection object. Based on the appearance frequency distribution, inspection color information for foreign object detection can be simply set to perform foreign object inspection. Therefore, not only can the inspection color setting process for foreign object inspection be greatly simplified, but also has a great practical effect, such as being able to reliably and reliably detect foreign matter mixed in the inspection object. Played.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram of a main part of a foreign matter inspection apparatus according to an embodiment of the present invention.
2 is a schematic functional configuration diagram of an image processing apparatus in the foreign substance inspection apparatus shown in FIG. 1;
FIG. 3 is a diagram showing a relationship between an appearance frequency distribution of each color component in a color image obtained by imaging an inspection object and a color component region determined based on the appearance frequency.
FIG. 4 is a diagram illustrating a relationship between a normal color region, an abnormal color region, and an approximate color region in an RB two-dimensional color space.
[Explanation of symbols]
1 Camera (imaging means)
4 Image processing device (computer)
DESCRIPTION OF SYMBOLS 10 Color selection function 12 Color component appearance frequency table 14 Inspection color setting means 15, 16a to 16n Inspection color table 17 Color comparison function 19 Large determination function

Claims (7)

Imaging means for capturing a color image of the inspection object;
Normal color region detecting means for obtaining a normal color component region of the inspection object based on the appearance frequency of each color component in the color image of the inspection object in advance;
First inspection color information setting means for setting color component areas other than the normal color color component area detected by the normal color area detection means as first inspection color information;
Second inspection color information for grouping color components included in a predetermined appearance frequency range based on the appearance frequency of each color component in the color image of the inspection object and setting the color component area as second inspection color information Setting means;
A specific color corresponding to the first and second inspection color information respectively registered by the first and second inspection color information setting means by color-selecting the color image of the inspection object obtained by the imaging means Color discriminating means for obtaining an image area of the component;
A foreign matter inspection apparatus comprising foreign matter detection means for detecting a foreign matter contained in the inspection object by determining the size of the image area of the specific color component obtained by the color discrimination means. The normal color area detecting means obtains an appearance frequency distribution of each color component in the color image of the inspection object, and determines a color component area that forms a single unit having a predetermined frequency threshold or more from the appearance frequency distribution as a normality of the inspection object. The foreign matter inspection apparatus according to claim 1, which is obtained as a color area. The second inspection color information setting means determines a color component having a color component region having a color component region within a predetermined frequency threshold range including the normal color region from the appearance frequency distribution of each color component in the color image of the inspection object. The foreign matter inspection device according to claim 1, wherein the foreign matter inspection device is set as a color component region to be included. The foreign matter inspection apparatus according to claim 3, wherein the second inspection color information is obtained as one or a plurality of color component areas. The color image is obtained by color separation into RGB three primary color components, and the appearance frequency distribution of each color component is obtained by counting the number of times each color component is registered in the RGB three-dimensional table. Item 1. A foreign matter inspection apparatus according to Item 1. The foreign matter inspection apparatus according to claim 1, wherein the color component of the color image is obtained as its wavelength component. The foreign object inspection apparatus according to claim 1, wherein the inspection object is made of tobacco leaf, food material, or a semi-finished product thereof.

Images