KR101288858B1 - Color sorting method of color sorter for grains - Google Patents

Abstract

PURPOSE: A method to detect faults of a color sorting machine for grains is provided to prevent malfunction where a blank section is recognized as the fault by omitting a fault detecting process with respect to the blank section through a pre-processing process. CONSTITUTION: A method to detect faults of a color sorting machine for grains comprises the steps of: setting a target area to detect the fault on color flat surface displaying with data of hue (H), saturation (S), and lightness (L) (S101); continuously filming the surface of granularity flowing through a supply unit and a discharging unit by a camera (S102); continuously obtaining the data of H, S, and L from an image signal continuously filming the surface of the granularity (S103); determining a section where the average of H, S, and L signals continued for a predetermined section belongs in a HSL section (S104); and performing predetermined operation to sort the fault when H, S, and L of the granularity belongs in the target are. [Reference numerals] (AA) Start; (BB) No; (CC) Yes; (DD) End; (S101) Set a target area; (S102) Film granularity; (S103) Obtain signals of H, S, and L; (S104) Determine a section where the average of each continued vibration signal of the H, S, and L belongs; (S105) All of the H, S, and L signals belong to the target area?; (S106) Perform a selecting movement

Description

Defect detection method of color sorter for grains {color sorting method of color sorter for grains}

The present invention relates to a defect detection method of a color sorter for grains, and more particularly, whether or not a granular material is defective by using a target area set on a color plane represented by hue, saturation, and brightness to detect a full color defect. It relates to a defect detection method of the color sorter for grain to determine.

The color sorter is a device for sorting foreign matter or defective products from the granular material by analyzing the surface color of the granular material. The color sorter is used to sort foreign matters or defective products against grains such as rice and barley, as well as various granular materials such as tea, coffee beans, beans, and salt.

1 is a view for explaining the operating state of the color sorter according to the prior art.

Referring to Figure 1, the color sorter according to the prior art (feeder) (1) for supplying the granular material, the discharge portion (chute) to drop the granular material supplied through the supply unit 1 along a certain trajectory or path (2) The ejection of the granules determined to be defective based on the screening result of the detection unit 3 and the detection unit 3 which detects the foreign matter or defective product through the color analysis of the granules falling through the discharge unit 2 is ejected through the air injection. and an ejector 4 for ejecting and separating from the normal granular material.

Meanwhile, the conventional color sorter acquires R, G, and B data on the granular object by using a camera, and calculates an OR by comparing the results of comparing the preset selection range with respect to each of the acquired R, G, and B data. Foreign substances or defective products are detected by AND or AND operation. However, this method can only perform gray level defect detection for each of the colors of R, G, and B, or there is a problem in that defective products or foreign substances are missed.

Figure 2 shows an example of the detection unit of the color sorter according to the prior art.

Referring to FIG. 2, the detector 3 outputs an image signal photographing the surface of the granular material in synchronization with a clock generator 31 that generates a clock and a clock output from the clock generator 31. A signal processor 33 and a signal processor 33 for signal processing the video signal output from the camera 32 and outputting R, G, and B data in synchronization with a clock output from the camera 32 and the clock generator 31. Through the logical sum operation of the signals output from the plurality of comparators 34a, 34b, 34c and the plurality of comparators 34a, 34b, 34c comparing each of the R, G, and B data output through An OR gate 35 for outputting a selection signal for controlling the operation of the injection unit 4 is included.

As shown in FIG. 2, when the OR signal is output by performing an OR operation on the results of comparing the R, G, and B data with the selection range, gray levels (Gray) for each of the R, G, and B colors are output. There is a problem that can only be performed to detect the failure level. This is because, by performing an OR operation on the results of comparing the R, G, and B data with the selection range, it is inevitable to generate a selection signal independent of each of the R, G, and B colors. In other words, the selection range set for each of the R, G, and B data is not present in the three-dimensional space of the general RGB color model, but only in the R, G, and B axes. Only gray level detection is possible for each of the B colors. However, the actual granular material may not only exist defects of R, G, and B single colors, but may also have numerous color defects, in which three colors are added at various ratios. Taking coffee beans as an example, there are various colors such as dark green to light green, light green, blue green, defense color, and ocher. Therefore, when the R, G, and B data obtained from the granular material are compared with the respective selection ranges, and the logical sum operation is performed, defects formed by combinations of R, G, and B colors are not detected, and only R on the RGB color model is detected. Defects can only be detected on the colors on the axis, G and B axis lines.

Figure 3 shows another example of the detection unit of the color sorter according to the prior art.

Referring to FIG. 3, the detector 3 includes a clock generator 31 for generating a clock and a camera 32 for outputting an image signal photographing the surface of a granular material in synchronization with a clock output from the clock generator 31. The signal processor 33 and the signal processor 33 output the R, G, and B data by signal processing the video signal output from the camera 32 in synchronization with the clock output from the clock generator 31. The injection unit by performing an AND operation on the signals output from the plurality of comparators 34a, 34b, 34c and the plurality of comparators 34a, 34b, 34c for comparing each of the R, G, and B data to a predetermined selection range. And an AND gate 36 for outputting a selection signal for controlling the operation of (4).

As shown in FIG. 3, in the case of outputting a selection signal by performing an AND operation on the results of comparing each of the R, G, and B data with the selection range, unlike the case of performing the OR operation, the R, G data is different. In addition to the gray level defect detection for each of the B colors, the defect detection for the combination of the R, G, and B colors is possible. In this case, an arbitrary cuboid space is formed on the RGB color model based on the selection range set for each of the R, G, and B colors, and all colors outside the formed cuboid space are recognized as defective. However, in this case, since each of the R, G, and B colors includes a brightness component and has a high correlation with each other, a color corresponding to a defect may be included in a cubic space defined as good, so that an unintended color is recognized as good. Problem occurs.

An object of the present invention is to provide a defect detection method for improving the screening reliability of the color sorter for grain.

In accordance with an aspect of the present invention, a failure detection method includes: setting a target area used for failure detection on a color plane represented by hue, saturation, and lightness; Calculating a selection range for each of the hue, saturation, and brightness based on the set boundary of the target area; Obtaining hue, saturation and brightness data from an image signal photographing the surface of the granular material; Comparing the obtained hue, saturation and brightness data with the calculated selection range, respectively; Performing a logical AND operation on the comparison results of comparing the obtained hue, saturation, and brightness data with the calculated selection range, respectively; And separating the granular material recognized as defective from the normal granular material based on the result of the AND operation.

According to the present invention, screening reliability is improved by setting a target region for defect detection using a color model represented by hue, saturation, and brightness instead of RGB, and performing defect detection using hue, saturation, and brightness of granular materials. It is effective to let. In addition, by omitting the defect detection process for the blank section through the preprocessing process, there is an effect of preventing a malfunction in recognizing the blank section as a defect and reducing unnecessary power consumption in the blank section.

1 is a view showing the operating state of the color sorter according to the prior art.
2 and 3 show examples of the detection unit of the color sorter according to the prior art.
4 shows an RGB color model.
5 shows an HSL color model.
6 is a structural diagram showing a color sorter for grains according to an embodiment of the present invention.
7 shows the HSL color plane.
8 illustrates an example of dividing the H, S, and L data into a plurality of regions in the color plane.
9 to 12 show examples of signal waveforms corresponding to specific regions on the color plane of FIG. 8.
FIG. 13 illustrates an ideal signal waveform in a blank section in which no granular material is photographed.
14 is a flowchart illustrating a failure detection method according to an exemplary embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and like reference numerals designate like parts throughout the specification.

Throughout the specification, when a part is said to "include" a certain component, it means that it can further include other components, without excluding other components unless specifically stated otherwise.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

In addition, the suffix "module" and " part "for constituent elements used in the following description are given or mixed in consideration of ease of specification, and do not have their own meaning or role.

For convenience of explanation, this document describes a color model divided into hue, saturation, and lightness, using an HSL (Hue, Saturation, Lightness) color model as an example. However, it is clear that the present invention is not limited thereto. According to the present invention, other color models such as HIS (Hue, Saturation, Intensity) color model and HSV (Hue, Saturation, Value) color model may be used as the color model that is divided into hue, saturation, and brightness.

Hereinafter, before describing the color sorter for grain according to an embodiment of the present invention, a method of converting an RGB color model, an HSL color model, and an RGB color model to an HSL color model will be described.

4 illustrates an RGB color model, and FIG. 5 illustrates an HSL color model.

Referring to FIG. 4, the RGB color model is represented by a three-dimensional space composed of an R axis, a G axis, and a B axis.

Referring to FIG. 5, the HSL color model is composed of hue (Hue, H), saturation (S), and lightness (L). Color is data that distinguishes color types such as red, orange, and yellow, and saturation is data representing clear or muddy information. In addition, the brightness is data indicating the degree of brightness. In the HSL color model, the brightest index white is set to 1.0 and the black to 0.0, and the brightness of all other colors is present between these two colors. On the other hand, in the HSL color model, hue, saturation, and brightness do not affect each other and exist as independent axes. Accordingly, when the HSL color model is applied to color selection, the quality range set for each of hue, saturation, and brightness may operate independently without affecting each other. For example, if the brightness range 0.4-0.6 (max 1.0), the color range 100-140 (max 360), and the saturation range 0.8-1.0 (max 1.0) are good, the green colors are all good. . In addition, when the HSL color model is used for color selection, since a brightness value exists in the middle of a separate Z-axis, a range of achromatic colors that vary from white to black can be prevented from being set as good products.

On the other hand, in general, image data photographed through a color camera is converted into R, G, B data through a signal processing unit and output. Therefore, in order to use the H, S, and L data for color selection, a conversion process for converting the R, G, and B data into H, S, and L data is required.

R, G, and B data can be converted into H, S, and L data using Equation 1 to be described later. Equation 1 to be described later is a well known technique in the field of image processing, detailed description thereof will be omitted.

Using Equation 1 above, when the R, G, and B data are output in 24 bits each by 8 bits after signal processing, it is necessary to calculate the HSL [24-bit] value for the number 2 ^{24 in} all cases. There will be a significant time delay.

In general, in the case of grain, the time required to drive the sprayer 4 by analyzing the image taken by the camera on the surface of the granular material is within 20 ms. Accordingly, the color analyzer for detecting defects of granular materials introduced at high speed, such as grains, needs to process all operations until photographing the granular objects with a camera and removing the defective products with the sprayer 4 within 20 ms. To this end, it is necessary to minimize the delay in converting the RGB color model to the HSL color model.

Hereinafter will be described with reference to the accompanying drawings for the color sorter for grains according to an embodiment of the present invention.

6 is a structural diagram showing a color sorter for grains according to an embodiment of the present invention. On the other hand, the components shown in Figure 6 are not essential, so the color sorter for grain may be implemented to have more or less components than those shown in Figure 6.

According to an embodiment of the present invention, the color sorter 100 for grain is a trajectory for supplying the granular material supplied through the supply unit 1 and the supply unit 1, as shown in FIG. 1 described above. Alternatively, it is determined that the defect is based on the screening result of the detection unit 3 and the detection unit 3 which detects the foreign matter or defective product by analyzing the color of the granular material falling through the discharge unit 2 and the discharge unit 2 falling along the path. And an injection unit 4 for separating the granular material from the normal granular material through air injection.

6, the detector 3 includes a clock generator 310, a camera 320, a signal processor 330, an HSL look up table 340, and a plurality of comparators 351 and 352. 353 and AND gate 360.

The clock generator 310 generates a clock necessary for the operation of the detector 2.

The camera 320 operates in synchronization with the clock output from the clock generator 310, and outputs an image signal photographing the surface of the granular material. According to one embodiment of the present invention, the grain color sorter 100 is a tri-linear CCD (scan) capable of scanning more than 9000 times per second to support the photographing without missing the granular material flowing at high speed ( A line scan camera 320 to which a Charge-Coupled Device sensor is applied may be used.

The signal processor 330 includes an analog-digital converter (ADC). The signal processor 330 operates in synchronization with a clock output from the clock generator 310, and converts an analog image signal output from the camera 320 into digital R, G, and B data using an ADC.

The HSL lookup table 340 is a memory that stores and manages H, S, and L data corresponding to all R, G, and B data combinations. When the R, G, and B data are input through the signal processor 330, Outputs H, S, and L data corresponding to the input R, G, and B data. That is, the R, G, and B data output through the signal processor 330 are input as an address, and the H, S, and L data corresponding to the received address are output. Here, H, S, and L data according to R, G, and B data may be previously calculated using Equation 1, and only the calculated result may be stored in the HSL lookup table 340.

Meanwhile, according to an embodiment of the present disclosure, the HSL lookup table 340 may be implemented as a flash memory. Accordingly, the time required to convert the R, G, and B data into the H, S, and L data using the HSL lookup table corresponds to the time to read the data from the flash memory, which corresponds to several tens of microseconds to several tens of microseconds. It is a very short time.

The H comparator 351 compares the H data output through the HSL lookup table 340 with a preset color selection range, and outputs a comparison result as an output signal. For example, in the color model of FIG. 5 described above, when the color screening range is 100 (lower limit value) to 140 (upper limit value), the H comparator 351 outputs 1 or 0 when the input H data enters the color screening range. do.

The S comparator 352 compares the S data output through the HSL lookup table 340 with a preset chroma screening range, and outputs a comparison result as an output signal. For example, in the aforementioned color model of FIG. 5, when the saturation screening range is 0.8 (lower limit value) to 1.0 (upper limit value), the S comparator 351 outputs 1 or 0 when the input S data falls within the saturation screening range. do.

The L comparator 353 compares the L data output through the HSL lookup table 340 with a predetermined brightness selection range and outputs a comparison result as an output signal. For example, in the above-described color model of FIG. 5, when the brightness selection range is 0.4 (lower limit value) to 0.6 (upper limit value), the L comparator 351 outputs 1 or 0 when the input L data falls within the brightness selection range. do.

The AND gate 360 performs an AND operation on the output results of the comparators 351, 352, and 353 for the H, S, and L data, respectively, and outputs a selection signal for driving the injection unit 4 to eject the granular material. do.

Hereinafter, before describing a defect detection method using a grain color sorter according to an embodiment of the present invention, a method of setting a color boundary for defect detection will be described with reference to necessary drawings. do.

7 shows the HSL color plane.

Referring to FIG. 7, the color model of FIG. 5 may be represented as a color plane in which H and S data correspond to a horizontal axis and a vertical axis, respectively, and L data is separately located. Accordingly, the color is determined by the H-S coordinates, and the brightness of the determined color depends on the L data. Meanwhile, according to an exemplary embodiment of the present disclosure, a target area for detecting a defect may be set on the full color color plane illustrated in FIG. 7.

In the following description, for convenience of description, each of the H, S, and L data is divided into three regions in the color plane of FIG. 7, and the H, S, and L data output through the HSL lookup table 340 are each divided into 8 bits. A case of outputting data is described as an example. However, the present invention is not limited thereto, and the H, S, and L data may be divided into any of various sections, and the H, S, and L data may be output as 16 bits, 32 bits, and the like instead of 8 bits.

8 illustrates an example of dividing the H, S, and L data into a plurality of regions in the color plane.

Referring to FIG. 8A, the H data is divided into three regions A, B, and C according to a value. In addition, the L data is also divided into three areas (a, b, c) according to the value. The S data is also divided into three regions 1, 2 and 3 according to the value.

On the other hand, as described above, the number of regions that can be represented by the combination of H, S, and L data each divided into three regions is 33, that is, 27, as shown in FIG. If you list all 27 areas, A1a, A1b, A1c, A2a, A2b, A2c, A3a, A3b, A3c, B1a, B1b, B1c, B2a, B2b, B2c, B3a, B3b, B3c, C1a, C1b, C1c , C2a, C2b, C2c, C3a, C3b, C3c.

9 to 12 illustrate examples of signal waveforms corresponding to specific regions on a color plane divided into a plurality of regions as illustrated in FIG. 8C.

FIG. 9 illustrates H, S, and L signal waveforms corresponding to A ○○ area. Referring to FIG. 9, it is understood that S, L signals exist throughout the entire area, but H signals exist only in the A area. Can be.

FIG. 10 illustrates H, S, and L signal waveforms corresponding to a ○ 3 ○ region. Referring to FIG. 10, it is understood that the H and L signals exist throughout the entire region, but the S signal exists only in the three regions. Can be.

FIG. 11 illustrates H, S, and L signal waveforms corresponding to an area ○○ b. Referring to FIG. 11, it is understood that H, S signals exist throughout the entire area, but L signals exist only in the b area. Can be.

12 illustrates H, S, and L signal waveforms corresponding to the A3b region. Referring to FIG. 12, in order to select only the A3b region in the color plane, each of the H, S, and L data needs to satisfy all of the conditions 0 <H <85, 170 <S <255, and 85 <L <170. That is, in order to recognize the color corresponding to the A3b region as a defect in the color plane, all of the conditions 0 <H <85, 170 <S <255, 85 <L <170 must be satisfied. Therefore, after comparing each of the H, S, and L data satisfying the above conditions, it is necessary to perform an AND operation on the comparison result.

According to an embodiment of the present invention, after setting a target area for defect recognition on a color plane represented by H, S, and L data, an upper limit value and a lower limit value of each of the H, S, and L data corresponding to the boundary of the target area is set. The selection range for each of the H, S, and L data may be set by calculating the.

Meanwhile, an ideal signal waveform in a blank section in which the camera 32 does not photograph the granular material is shown in FIG. 13. Therefore, in an embodiment of the present invention, in order to prevent the blank section from being recognized as a failure, in the blank section, the detection of the failure is stopped, or the preprocessing process is performed so that the selection signal is output as a good product. For example, the input or output of the comparator may be controlled such that all the comparison results of each comparator output a signal corresponding to good quality.

14 is a flowchart illustrating a failure detection method of a color sorter for grains according to an embodiment of the present invention.

Referring to FIG. 14, the grain color sorter 100 sets a target area for defect detection on a color plane represented by H, S, and L data (S101). The grain color sorter 100 divides each of H, S, and L into a plurality of sections on the color plane. This target area may be any rectangle satisfying the setting of 0 <H <85,170 <S <255,85 <L <170. In addition, any one of the divided sections is selected for each of the H, S, and L signals, and a common area including all the selected sections is set as the target area.

Subsequently, when the grain color sorter 100 starts to detect defects, the grain color sorter 100 continuously photographs the surface of the granular material introduced through the supply unit 1 and the discharge unit 2 through the camera 320 (S102). In addition, H, S and L data are continuously obtained from the image signal of continuously photographing the surface of the granular material (S103). The color sorter 100 for grains is obtained through the camera 320 using the signal processor 330. A video signal obtained by continuously photographing the granular object to be screened is converted into R, G, and B data. In addition, the HSL lookup table 340 converts the R, G, and B data continuously output through the signal processor 330 into H, S, and L data.

Meanwhile, according to an exemplary embodiment of the present disclosure, an image signal may be photographed over a plurality of frames with respect to one granular object due to high speed photographing when the granular object to be screened is photographed. As such, when image capturing is performed over several frames for one granular material, the grain color sorter 100 continuously receives H, S, and L signals for a predetermined period to determine whether the granular material is defective. Need to be acquired.

Accordingly, the grain color sorter 100 continuously acquires H, S, and L signals for a predetermined section, and the H, S, and L signals obtained during the predetermined section may be output in the form of a vibration signal, thereby more accurately determining a defect. It is necessary to calculate the average for the H, S and L signals. In addition, the grain color sorter 100 determines a section in which the average of the H, S, and L signals consecutive for a predetermined section belongs to the HSL section (S104).

In addition, it is determined whether a section to which the averages of the H, S, and L signals belong to the target area set in step S101 (S105). In this case, the color sorter 100 for grains performs a predetermined defective sorting operation (S106).

Alternatively, the grain color sorter 100 continuously acquires H, S, and L signals for a predetermined period, detects whether the obtained H, S, and L data values belong to the set target region, and belongs to the target region. In this case, it is possible to increase the count and determine that the granular material is defective when the count number exceeds a preset criterion (for example, three times), that is, when a signal belonging to the defective criteria is continuously detected. Do.

In the step S104 and S105, the color sorter 100 for grains H, S obtained for the granular material in the selection section set for each of the H, S and L signals using a plurality of comparators 351, 352, 353 And the L signal may be determined. The H, S, and L signals may be determined by performing an AND operation on the comparison results output from the plurality of comparators 351, 352, and 353 using the AND gate 360. Can be determined whether or not all are included in the target area set in step S101. Here, the AND gate 4 outputs different calculation results for the case where the granular material is defective or the case of good quality. Subsequently, the result of the AND operation is output as a selection signal for driving the injection unit 4, and when the injection unit 4 outputs a selection signal indicating that the injection unit 4 is defective, the granular material determined to be defective is determined. Is ejected to separate it from the granular material identified as good.

According to an embodiment of the present disclosure, when the target area set in step S101 is a failure detection area, that is, when the inside of the set target area is recognized as a failure, inside defect detection is performed, and the set target area is normally detected. In the case of an area, that is, when the inside of the set target area is recognized as normal and the outside is regarded as bad, both outside defect detection is possible. In the case of inside defect detection, each comparator outputs a signal corresponding to a failure when the input data is within the preset selection range, and in the case of outside failure detection, each comparator outputs the input data outside the preset selection range. Outputs a signal corresponding to a defective surface.

Meanwhile, according to an embodiment of the present invention, the grain color sorter 100 determines whether the image signal currently input corresponds to a blank section in which the granular material does not exist before determining whether the granular material for the screening object is defective. Can be performed. The grain color sorter 100 determines whether each of the H, S, and L data is included in the section set as the blank section, and when all the H, S, and L data are included in the section set as the blank section, the granular material is present. It is determined that the blank section does not. In addition, if it is determined that the blank section is determined, the defect detection process performed in steps S105 and S104 is omitted, or a screening signal indicating normality is output to the injection unit 4. On the other hand, when the blank section is not, the control unit ( 180 performs a failure detection process over steps S105 and S104.

According to an embodiment of the present invention described above, the grain color sorter sets a target area for defect detection using a color model represented by hue, saturation, and brightness instead of RGB, and uses the color, saturation, and brightness of the granular material. By performing the defect detection, there is an effect of improving the screening reliability. In addition, by omitting the defect detection process for the blank section through the preprocessing process, there is an effect of preventing a malfunction in recognizing the blank section as a defect and reducing unnecessary power consumption in the blank section.

The embodiments of the present invention described above are not only implemented by the apparatus and method but may be implemented through a program for realizing the function corresponding to the configuration of the embodiment of the present invention or a recording medium on which the program is recorded, The embodiments can be easily implemented by those skilled in the art from the description of the embodiments described above.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. The present invention is not limited to the drawings, and all or some of the embodiments may be selectively combined so that various modifications may be made.

Claims (5)

A common region including each of the hue, saturation, and brightness signal sections divided into a plurality of sections, selecting one of the divided sections for each of hue, saturation, and brightness, and including all of the selected sections. Setting the target area;
A plurality of hue, saturation, and brightness data are stored as a look-up table, and when RGB data of an image signal is input as an address, a selection target is prized using a memory that outputs hue, saturation, and brightness data corresponding to the input RGB data. Continuously acquiring hue, saturation, and lightness signals from an image signal of continuously photographing water;
Vibration signal form continuously obtained from the image signal photographing the granular object through a plurality of comparators for outputting a comparison result between the hue, saturation and brightness data output from the memory and a preset hue, saturation and brightness selection range Determining a section to which an average of each of hue, saturation, and lightness signals belongs;
Determining a section to which the average of the hue, saturation, and brightness signals belong, based on a result of the logical AND operation of the output values of the plurality of comparators, and determining whether the hue, saturation, and brightness of the granular material are included in the target region; ; And
If the hue, saturation and brightness of the granular material is included in the target area, performing a predetermined failure screening operation,
When the hue, saturation and brightness signals belong to a predetermined reference abnormal failure detection area, the count is increased, and when the count continuously exceeds the predetermined reference value, a defect sorting operation is performed. Defect detection method. The method of claim 1,
The set target region is a defect detection region, and the defect sorting operation is an operation of ejecting the granular material. delete delete delete

Images