JPH0365768A - Inspecting device for grain - Google Patents

Abstract

PURPOSE:To simplify the configuration of an inspecting device for grains by discriminating whether the grain is unpolished rice or unhulled rice based on the size of an area where the value of difference between one of red information and green information, and blue information is more than a set value out of three primary color components. CONSTITUTION:An image pickup means S1 is constituted in a color type so that one of the red information and green information and the blue information can be outputted out of the three primary color components. In the area where the value of the difference between one of the red information and green information, and the blue information is more than the set value, when the size of the area is larger than the set value, a discriminating means 101 discriminates that the grain is the unhulled rice and when the size of the area is smaller than the set value, the grain is discriminated as the unpolished rice. Thus, the configuration of the device can be simplified.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides an imaging means for imaging grains, and a method for determining whether the grains are paddy or brown rice based on the imaging information of the imaging means. The present invention relates to a grain inspection device including a determination means for determining grains.

[Conventional technology]

For example, if the paddy after threshing contains brown rice, there is a disadvantage that the incidence of shell cracking increases when it is dried in a dryer. A performance test is being conducted on the threshing equipment to check the rate of paddy contamination.

However, it is desired that the contamination rate of brown rice be 0.5% or less. Therefore, in order to check whether or not the threshing process was being carried out properly, we decided to
Take out a sample of about 000 grains to 5000 grains,
It is necessary to determine whether each grain of these samples is paddy or brown rice.

In addition, in the sorting equipment for sorting and collecting grains after hulling into brown rice and paddy, it is desirable that the contamination rate of paddy in the sorted and collected brown rice be 0.3% or less. Similar to the above performance test of the threshing device, it is necessary to be able to distinguish between brown rice and paddy grain by grain.

By the way, since brown rice and paddy differ in their colors, conventionally, for example, a method has been proposed that captures a color image of the grain and distinguishes between brown rice and paddy based on the color, that is, the ratio of the three primary color components. (For example, see Japanese Patent Application No. 139003/1983, which was previously proposed by the present applicant).

Furthermore, since brown rice and rice grains have different reflectances for near-infrared light, rice grains have the characteristic of being imaged brighter than brown rice.

Therefore, for example, if the area of the area corresponding to the grain imaged while illuminating with near-infrared light is greater than a set value, it is determined to be paddy, and if the area is smaller than the set value, it is determined to be brown rice. Means to do this have also been proposed (for example, see Japanese Patent Application No. 182588/1988, previously proposed by the present applicant).

[Problem to be solved by the invention]

However, the former method for distinguishing between brown rice and paddy based on the ratio of the three primary color components has the disadvantage that image processing is complicated because all the three primary color components are used.

In addition, in the latter method that utilizes the difference in reflectance of near-infrared light, it is necessary to keep the brightness constant during imaging because paddy and brown rice are basically distinguished based on the difference in brightness. Furthermore, since near-infrared light is used, there is a disadvantage that the device configuration is complicated.

The present invention has been made in view of the above circumstances, and its purpose is to simplify the device configuration.

[Means to solve the problem]

A grain inspection device according to the present invention includes an imaging means for taking an image of the grain, and a determining means for determining whether the grain is paddy or brown rice based on the imaging information of the imaging means. Its characteristic structure is as follows.

That is, the imaging means is configured in color so as to output at least one of red information and green information of the three primary color components and blue information, and the discrimination means outputs one of the red information and green information and the If the area of the area is larger than the setting value among the areas where the value of the difference with the blue color information is greater than the setting value, the grain is determined to be paddy, and the area is smaller than the setting value. In some cases, the grain is configured to be determined to be brown rice.

[For production]

Focusing on the values of the three primary color components of grains captured in color images, there is a slight difference in the values of red and green information between brown rice and paddy, but the value of blue information is higher in brown rice than in paddy. It turned out that it was also large.

Therefore, based on the size of the area where the difference value between red information and green information of the three primary color components and blue information is greater than a set value, it is determined whether it is brown rice or paddy. By doing this, while using less image information,
Image processing for determining whether a grain is brown rice or paddy can be simplified while normalizing image information with respect to brightness without reducing the determination accuracy.

〔Effect of the invention〕

Therefore, the device configuration can be simplified, and the process of discriminating between brown rice and paddy can be sped up.

〔Example〕

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to a grain inspection device for inspecting whether or not a threshing device is properly threshing will be described below with reference to the drawings.

To explain, if the paddy after threshing contains brown rice or grains with twig stalks, there is a disadvantage that there is a high incidence of shell cracking and damage to the grains when drying in a dryer. Therefore, performance tests are being conducted on threshing equipment to check the contamination rate of brown rice and paddy with twig stalks in the paddy after threshing. By the way, if the threshing process is excessive, the husks will be removed and the rice will become brown rice, and if it is insufficient, the paddy will be left with twigs and stalks.

It is desired that the contamination rate of brown rice be 0.5% or less, and the contamination rate of twigs and stalks be 3% or less. Therefore, in order to check whether or not the threshing process was carried out properly, 4,000 to 5,000 grains were extracted from the grains after the threshing process.
Samples of grain size 0 are taken out and each grain is determined whether the sample is paddy, brown rice, or with twig stalks.

As shown in FIGS. 12 and 13, the apparatus for inspecting grains after threshing includes a sample hopper (1) that stores a plurality of grains (A) to be inspected, and a sample hopper (1) that stores a plurality of grains (A) to be inspected. ) A rotating body (2) that is driven to rotate around a horizontal axis (P) by an electric motor (M) while adsorbing and holding the grains (A) stored in the container (A) one by one at regular intervals on its outer periphery. , is equipped with a two-dimensional image sensor (S1) using a color TV camera or the like as an imaging means for taking an image of each grain of the grain (A) held by suction on the rotating body (2), and capturing information on the image. By image processing, it is configured to automatically determine whether the grain (A) adsorbed and held on the outer periphery of the rotating body (2) is paddy, brown rice, or grain with twig stalks. There is. However, as will be described later, the imaged grains are automatically separated from the rotating body (2).

The structure for adsorbing and holding the grains (A) on the outer circumference of the rotating body (2) will be described.
), a plurality of suction holes (3) penetrating from the center of the rotation axis (P) toward the outer circumference are formed at equal intervals in the circumferential direction.

Each suction hole (
A ventilation hole (4) is formed which communicates with the end of the rotating body (3) on the rotation center side and penetrates through the back side of the rotating body (2). The ventilation hole (4) communicates with the inside of a cylindrical member (5) that supports the rotating body (2) from the back side in a rotatable state around the axis (P).

However, inside the cylindrical member (5), there is a partition plate (6) formed in a V-shape when viewed from the open end side.
A suction space (6) for sucking the suction hole (3) from the outer circumferential direction toward the center of the rotating body (2)
A) and a discharge space (6B) that blows air from the center toward the outer periphery, and the suction space (6A) is connected to a suction pump (7) and maintained at negative pressure, and The discharge space (6B) is connected to a discharge pump (8) and is maintained at a positive pressure.

That is, while the rotation center side end of the suction hole (3) is in the suction space (6A), the plurality of samples stored in the sample hopper (1) are While the grains (A) are being sucked and held at intervals of - grains, and the rotation center side of the suction hole (3) is within the discharge space (6B), the rotating body (2) The suction and discharge are automatically repeated so that the suction and retention are automatically blown away by the pressure blown toward the outer periphery of the grains (A). It is constructed so that each grain can be inspected.

Then, the image sensor (Sl) detects the grains (A) at an upper part of the rotating body (2), which is a section where the grains (A) are adsorbed and held in the suction holes (3). - It is provided so that it can image each grain.

However, although not detailed, the outer peripheral part of the rotating body (2) is
It is painted black to prevent the influence of reflected light, and the imaging range of the image sensor (S1) is illuminated at a set distance.

In FIG. 13, (9) is a bearing inserted between the rotating body (2) and the cylindrical member (5). or,
(S2) is an encoder for detecting the rotation angle of the rotating body (2), and the detection information is transmitted to the suction hole (S2).
3) is used as control information for starting image processing when the grain (A) held by suction is located within the imaging field of view of the image sensor (S1).

Next, a control configuration for determining whether the grain (A) is paddy, brown rice, or grain with twig stalks will be described based on the imaging information of the image sensor (S1).

As shown in FIG. 1, an image processing means (100) that takes in imaging information of the image sensor (S1) based on information of the encoder (S2) and processes the image; Based on the information of paddy, brown rice,
Discrimination means (101) for discriminating whether it is a twig with stem or
) and a microcomputer-based control device (
10), and a monitor device (11) for displaying imaging information and image-processed information of the image sensor (S+).
) is provided.

To explain image processing, the image sensor (
S1) is a brightness signal (Y) corresponding to the brightness of the imaging information;
Red signal (R) and green signal (G) corresponding to three primary color information
, and a blue signal (B), respectively, and the image sensor (S1
) is captured, first, the luminance signal (
Based on the value of Y) - 2 of 32X32 pixels per screen
It is converted into digitized image information, and the region (X) corresponding to the grain (A) is extracted (see FIG. 8).

However, since the twig stalk is much narrower than the grain (A),
As shown in Figure 6, when an image is captured in a focused state, as mentioned above, the image information is quantized relatively coarsely at 32 x 32 pixels, and due to quantization errors in image processing, the extracted information is There is a possibility that the region (X) is divided into a grain region (Xa) corresponding to the grain (A) and a twig stalk region (Xb) corresponding to the twig stalk, leading to erroneous determination.

Therefore, by intentionally causing the image sensor (S1) to take an image in an out-of-focus state, as shown in FIG.
This is done by expanding the image signal components corresponding to grains and twigs so that quantization errors are less likely to occur.

Next, image processing for determining whether the grain (A) is brown rice, paddy, or twig-stem based on the information of the extracted region (X) will be described.

If the threshing process is insufficient, twigs will be attached. As a result, the extracted area (X) becomes larger by 5, which is the area of the twig area (Xb) corresponding to the twig stump, than when no twig stump is attached. On the other hand, if the threshing process is excessive, the grains (A) will be hulled and become brown rice, which will be smaller in size than unhulled paddy.

Therefore, basically, if the area of the area (X) is larger than the size of properly threshed paddy without twig stumps, it is determined that the paddy has twig stumps. It's like letting them do it. Incidentally, since the size of the region (X) is proportional to the number of pixels belonging to the region (X), it is determined that the larger the number of pixels, the larger the area.

However, the size of grains (A) varies due to individual differences even if they are of the same type, so it is difficult to distinguish between paddy and brown rice based only on the size of the area (X). There is a risk of inaccuracy.

Therefore, as will be described in detail later, the image sensor (
The color signal information output from S1) is used to distinguish between paddy and brown rice.

Incidentally, since brown rice is hulled, its size becomes smaller than paddy, so it is determined whether it is brown rice or paddy based on the area of the region (X) extracted based on the value of the luminance signal (Y). Although it is not impossible to distinguish between paddy and brown rice, depending on individual differences, paddy and brown rice may have the same area, so 100%
It is difficult to distinguish (see Figure 3).

Adding an explanation to the process for distinguishing between paddy and brown rice based on the color described above, if we focus on the color of grains (A) when they are paddy and when they are brown rice, brown rice is looks whiter than paddy.

Therefore, when comparing the distribution state of the three primary color information, it is found that the fourth
As shown in the figure, the value of the red light (R) is not much different between paddy and brown rice, but the value of the blue light (B) is significantly greater for brown rice than for paddy. Becomes large.

Although not shown, the value of the green signal (G) is similar to the red signal (R), and there is no significant difference between paddy and brown rice, so the green signal (G) is used instead of the red signal (R). G) may also be used.

Therefore, the red signal (
When extracting a region where the value obtained by subtracting the blue signal (B) from R) is larger than the set value for noise removal, the area is determined to be brown rice if it is paddy, regardless of individual differences. The difference is much larger than in the case of rice, and it is possible to determine whether it is paddy or brown rice with approximately 100% probability (see Figure 5).

In other words, the area of the region where the value obtained by subtracting the blue signal (B) from the red signal (R) is larger than the set value for noise removal is larger than the set threshold for distinguishing between brown rice and paddy. By determining whether or not it is, it is possible to simultaneously normalize the brightness of color information and determine whether it is paddy or brown rice.

However, in order to speed up image processing, when capturing an image captured by the image sensor (S1), in parallel with the quantization of the luminance signal (Y), the red signal (R) is quantized to the blue signal. The value obtained by subtracting the signal (B) is calculated as the luminance signal (Y
) is quantized at the same pixel density and stored in an image memory (not shown) in advance.

If it is determined to be paddy, then re-extract the twig stalk region (Xb) (see Figure 8) that corresponds only to the twig stalk from the region (X), calculate its length, and If the length is longer than a set value, it is determined to have twigs and stems.

Next, a process for determining the length of the twig stalk will be explained.

Since the twig stalk is much narrower than the main body of the grain (A),
The number of pixels belonging to the twig area (Xb) is smaller than the number of pixels belonging to the grain area (Xa).

Therefore, when calculating the frequency distribution of pixels having the same coordinate values on each of the X-axis and y-axis of the imaging field of view (B), the number of pixels at the center of the grain area (Xa) on each of the X-axis and y-axis is determined. The number of pixels becomes approximately the maximum, and the number of pixels gradually decreases toward the outer end of the region (X) (
(See Figure 8).

Therefore, the frequency distribution of pixels having the same coordinate value on each of the X-axis and the y-axis is determined, and on the side where the distribution width is larger, the frequency is smaller than the set frequency corresponding to the size of the paddy, and In addition, the center value of the coordinates where the width is the maximum is determined as the value corresponding to the center of the grain area (Xa).

Next, from the center of the grain region (Xa) toward the outer end of the distribution at a distance, find a point where the frequency is first smaller than the set frequency set corresponding to the width of the twig stalk. , the boundary between the grain region (Xa) and the twig stalk region (Xb) (
This will be obtained as a).

However, if the twig stalk is long, the end of the twig stalk may be imaged in a curved state, as shown in FIG. Area (Xa
), it is not possible to distinguish between the heavy boundary (a) and the false boundary (C) having the same coordinate values based only on the frequency distribution, and the frequency distribution is such that the grain Twig area (Xb) overlapping grain area (Xa)
), there is a risk of losing sight of the end (b).

Therefore, when determining the frequency distribution, for example, by sequentially searching from the smallest coordinate value (X or y) on the screen to the largest one, as shown in FIG. If there is no pixel belonging to the boundary, an overlap flag indicating that there is no pixel having the same coordinate value is set to 0, and a position determination flag for the boundary (a) is set to a value corresponding to nothing. If there are pixels that belong to the area (X) and are consecutively adjacent to each other, the overlap flag is set to 1, and the position determination flag is set to a value corresponding to paddy. . Next, if there are two pixels with the same coordinate value separated by an interval, an overlap flag indicating that these pixels overlap is set to 2, and the position determination flag is set to 2. Set to the value corresponding to .

After that, the coordinate value of the narrower width obtained from the frequency distribution is checked based on the set values of the overlap flag and the position determination flag, and the boundary obtained from the frequency distribution is determined as the typical boundary. (a) or a false boundary (C), and determine the coordinate values of the starting point of the twig-stem area (Xb).

After determining the coordinate values of the starting point of the twig stoma region (Xb),
The search is performed in order until there are no pixels adjacent to the pixel having the coordinate value, and the coordinate value of the final pixel is set as the end point of the twig-stem region (Xb).

After determining the starting point and ending point of the twig-stem area (Xb),
In order to find the length between the starting point and the ending point, the pixels belonging to the twig-stem region (Xb) are differentiated sequentially from the starting point toward the ending point to thin the line to a width of one pixel, The number of pixels, that is, the length of these thinned pixels is determined.

However, as shown in Figure 9, if the extracted pixels are adjacent in a direction parallel to the X-axis or y-axis, 1 is added as the unit distance, and if they are adjacent in a diagonal direction, , which corresponds to the diagonal distance of the square as a unit distance,
By adding 4, the length of the twig can be accurately determined even if the twig is curved.

However, if the twig stoma is curved in a direction intersecting the coordinate direction to be searched, it is determined that the adjacency of pixels belonging to the twig stoma region (Xb) is interrupted by the thinning process by the differentiation. In that case, there is a possibility that
Based on the values of the shape flag and the position determination flag, the value (n) of the difference in the number of pixels in the intersecting direction is added to the unit distance (l or 1.4).

After determining the length of the twig stalk, based on the length value,
For example, if it is less than 8M (equivalent to 15 dots in terms of number of pixels), which is a range that does not affect the drying process, etc., it is determined that it is a twig infarction.

Next, the operation of the control device (10) will be explained based on the flowchart shown in FIG.

Based on the information of the encoder (S2), the grain (
When grain A) is located approximately at the center of the imaging field of view of the image sensor (S1), the imaging information of the image sensor (S1) is captured and the grain (A)
In order to separate the grain from the background, it is quantized at a set pixel density and a region (X) corresponding only to the grain (A) is extracted (see FIG. 8).

After extracting the region (X), first, in order to prevent misclassification due to foreign matter such as dust, the number of pixels belonging to the region (X) is significantly smaller than the size of the grain (A). If the area is smaller than the value corresponding to the set pixel count (set to 20 dots), the imaged object is considered to be dust. It is determined that it is a foreign object such as a foreign object, and is removed from the processing target.

If the grain (A) has a twig stump, its size will be larger than when the twig stump is not attached, so it is determined whether the grain (A) has a twig stump attached or not. In order to
It is determined whether the number of pixels is greater than or equal to the set number of pixels (40 dots), and if it is greater than or equal to the set number of pixels, it is determined that the pixel has twigs.

When the area of the region (X) is less than the set number of pixels, the difference between red information (R) and blue information (B) of the three primary color information output from the image sensor (S1) Based on the value, it is determined whether the area of the extracted area is less than or equal to a set value (set to 15 dots), and if it is less than the set value, it is determined that it is brown rice, and if the area is larger than the set value. If so, it is determined that it is paddy.

If the paddy is determined to be paddy, next, if the length of the twig stalk determined as described above is less than the set value (15 dots), it is determined that the paddy is without the twig stalk. death,
If it exceeds the set value, it will be determined that it has twigs and stems.

Although not described in detail, the process described above is automatically repeated based on the information of the encoder (S2) until the number of samples reaches the set value.

Example C] In the above example, the grains to be inspected (A) are held by suction on the outer periphery of the rotating body (2), thereby making it possible to image each grain at intervals. Although the case is shown as an example, for example, the grains (A) may be continuously conveyed using a belt-type conveyor or the like, with the grains being spaced apart from each other. - The specific configuration of each part, such as means for imaging each particle, can be changed in various ways.

Further, in the above embodiment, the present invention is applied to a device for inspecting grain after threshing, but the present invention can be applied to various devices that need to distinguish between paddy and brown rice. For example, in a sorting device for sorting and recovering brown rice from grains after hulling, it can also be used as a means for detecting the sorting state.

By the way, in the sorting equipment that separates and collects grains after hulling into brown rice and paddy, it is desirable that the contamination rate of paddy in the sorted and collected brown rice is 0.3% or less. It is necessary to be able to distinguish between brown rice and paddy grain by grain, and by using the present invention, accurate discrimination can be made.

Incidentally, although reference numerals are written in the claims section for convenient comparison with the drawings, the present invention is not limited to the structure shown in the accompanying drawings.

[Brief explanation of drawings]

The drawings show an embodiment of the grain inspection device according to the present invention, FIG. 1 is a block diagram of the control configuration, FIG. 2 is a flowchart of control operation, and FIGS. 3 to 11 are explanatory diagrams of image processing.
FIG. 12 is a cutaway front view of the inspection device, and FIG. 13 is a sectional view of the same side. (S,)...Imaging means, (R)...Red information, (G)...Green information, (B)...
... Blue information, (101) ... Discrimination means.

Claims (1)

[Claims] An imaging means (S_1) for imaging grains, and an imaging means (S_1) for imaging grains;
A grain inspection device comprising a determining means (101) for determining whether the grain is paddy or brown rice based on the imaging information of S_1), the imaging means (S_1) comprising at least The discriminating means (101) is configured in color so as to output one of red information (R) and green information (G) of the three primary color components and blue information (B), and the discriminating means (101) and the area where the difference value between one of the green information (G) and the blue information (B) is greater than or equal to the set value, if the area is larger than the set value, the grain is paddy. And if the area is smaller than the set value,
A grain inspection device configured to determine that the grain is brown rice.