Seed selection method by hyperspectral content and classification device by seed ingredient content

Abstract

The present invention relates to a method for sorting seeds by component content using hyperspectral The automatic weighing unit comprises an automatic weighing step that weighs seeds according to weight, consisting of seeds classified by content and seeds that meet the content, by repeatedly reclassifying seeds that do not meet the above-mentioned content.

Description

Technology Field

[0001] The present invention relates to a seed sorting method based on seed component content and a seed classification device based on seed component content, and more specifically, to a seed sorting method based on seed component content and a seed classification device based on seed component content utilizing hyperspectral Background Technology

[0003] In agricultural research, research related to seed composition is very important as it is closely linked to food security, including food.

[0004] Conventional seed sorting facilities employ a process that automatically weighs seeds after removing contaminants, classifying by size, removing foreign substances, using washing machines, and sorting out defective seeds. However, this sorting process does not consider the composition of the seeds; consequently, seeds with varying compositions are mixed and distributed for food processing, making it difficult to achieve uniform quality.

[0005] Recently, classification by component content has become an essential step across all fields, from seed sorting to food processing, and much research is being conducted, including the analysis and utilization of wavelength bands using hyperspectral sensors.

[0006] However, utilizing hyperspectral sensors in industrial settings is difficult to achieve commercialization due to the long time required for research costs and durations for experiments and analysis.

[0007] Therefore, there is a need for technology to utilize such hyperspectral technology in industrial seed sorting applications. Prior art literature

[0009] Patent Document: Registered Patent No. 10-2208447 (2021.01.21) The problem to be solved

[0010] The present invention has been devised to solve the aforementioned problems. The invention provides seeds weighed according to their respective component content through a sorting technology that utilizes a hyperspectral sensor to classify by component content, comprising raw material input, removal of foreign substances, spectral image analysis of seeds, classification of seeds by component content, and weighing operations. It also enables uniform purification operations starting from the raw grain to derive uniform food content and aims to provide classifications of various crops and their respective component contents by modifying the verification model.

[0011] In addition, the present invention provides a sorting function based on component content with a wide range of applications, and more specifically, by utilizing it on various seeds such as rice, soybeans, wheat, barley, and coffee, it enables classification by component including functional substances such as proteins, lipids, and carbohydrates, thereby providing a technology that can be utilized in various ways in the high-protein tofu, soy milk, baking industry, flour, and coffee industry. means of solving the problem

[0013] A method for sorting seeds by component content using hyperspectral

[0014] According to another embodiment of the present invention, the spectral imaging step may involve the spectral imaging unit capturing the spectral image at a wavelength set to correspond to the component content of the seed.

[0015] According to another embodiment of the present invention, the spectroscopic image capturing step may allow the spectroscopic image capturing unit to capture the spectroscopic image at a wavelength selected according to the component content in the wavelength band of 400 nm to 2500 nm.

[0016] According to another embodiment of the present invention, the result classification unit step may classify according to the components of the protein contained in the soybean by using the result of analyzing the spectroscopic image of the soybean that the result classification unit took of the soybean seed.

[0017] According to another embodiment of the present invention, the step of repeating the reclassification may reset the PLSR model of the data analysis unit, and the result classification unit may repeat the reclassification of soybeans with insufficient protein content to classify them into high protein, medium protein, and low protein.

[0018] A seed component content classification device according to one embodiment of the present invention comprises: a raw material purification unit that removes contaminants and foreign substances from the raw material and selects uniform seeds through seed size classification and washing; a spectroscopic image capturing unit that captures spectroscopic images of seeds free-falling through a raw material descent chute; a data analysis unit that analyzes the spectroscopic images of the seeds captured through a PLSR (Partial Least Square Regression) model; and a result classification unit that classifies seeds with insufficient component content and seeds with sufficient component content by selecting seeds with insufficient component content through an air compression and injection method using the results of analyzing the spectroscopic images. The data analysis unit resets the PLSR model, and the result classification unit repeats the reclassification of the seeds with insufficient component content.

[0019] According to another embodiment of the present invention, the apparatus may further comprise an automatic weighing unit that weighs seeds according to weight based on their component content, comprising seeds classified by repeating the reclassification of seeds with insufficient component content and seeds that have reached the required content.

[0020] According to another embodiment of the present invention, the spectroscopic imaging unit can capture the spectroscopic image at a wavelength set to correspond to the component content of the seed.

[0021] According to another embodiment of the present invention, the spectroscopic imaging unit can capture the spectroscopic image at a wavelength selected according to the component content in the wavelength range of 400 nm to 2500 nm.

[0022] According to another embodiment of the present invention, the result classification unit can classify according to the components of the protein contained in the soybean by using the result of analyzing the spectroscopic image of the soybean that is the seed.

[0023] According to another embodiment of the present invention, the result classification unit can classify soybeans with insufficient protein content into high protein, medium protein, and low protein by repeating the reclassification of soybeans with insufficient protein content as the data analysis unit resets the PLSR model. Effects of the invention

[0025] According to the present invention, through a sorting technology that utilizes a hyperspectral sensor to perform classification by component content, seeds weighed according to their component content can be provided through the input of raw materials, removal of foreign substances, spectral image analysis of seeds, classification of seeds according to component content, and weighing operation. Since uniform purification can be performed starting from the raw grain, uniform food content can be derived, and by modifying the verification model, various crops and classifications of their respective component contents can be provided.

[0026] In addition, according to the present invention, a sorting function based on component content is provided, thereby providing a wide range of applications. More specifically, since it can be applied to various seeds such as rice, soybeans, wheat, barley, and coffee, and classified by component including functional substances such as proteins, lipids, and carbohydrates, it can provide a technology that can be utilized in various ways in the high-protein tofu, soy milk, baking industry, flour, and coffee industry. Brief explanation of the drawing

[0028] FIG. 1 is a diagram illustrating the overall process of a seed component content sorting method using hyperspectral spectroscopy according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating a sorting method based on seed component content using hyperspectral spectroscopy according to an embodiment of the present invention. FIG. 3 is a diagram illustrating a sorting method based on seed component content using hyperspectral spectroscopy according to an embodiment of the present invention. FIG. 4 is a diagram illustrating a data analysis method according to an embodiment of the present invention. FIG. 5 is a diagram illustrating the statistical results of a PLSR model according to an embodiment of the present invention. FIG. 6 is a diagram illustrating prediction data of a PLSR model according to an embodiment of the present invention. FIG. 7 is a drawing illustrating a method for reclassifying seeds according to an embodiment of the present invention. FIGS. 8 and 9 are drawings illustrating the results of component classification of a sorting method based on seed component content using hyperspectral spectroscopy according to an embodiment of the present invention. FIG. 10 is a diagram showing the configuration of a classification device by seed component content according to an embodiment of the present invention. FIG. 11 is a drawing illustrating a classification device by seed component content according to an embodiment of the present invention. Specific details for implementing the invention

[0029] The present invention is capable of various modifications and may have various embodiments, and specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and it should be understood that it includes all modifications, equivalents, and substitutions that fall within the spirit and scope of the invention.

[0030] However, in describing the embodiments, if it is determined that a detailed description of related known functions or configurations could unnecessarily obscure the essence of the invention, such detailed description is omitted. Furthermore, the sizes of each component in the drawings may be exaggerated for illustrative purposes and do not represent the actual sizes applied.

[0031] Furthermore, throughout the specification, when a component is referred to as being "connected" or "joined" with another component, it should be understood that the component may be directly connected or joined to the other component, but unless specifically stated otherwise, it may also be connected or joined through an intermediate component. Additionally, throughout the specification, when a part is described as "including" a component, unless specifically stated otherwise, this means that it may include additional components rather than excluding other components.

[0032] FIG. 1 is a diagram illustrating the overall process of a seed component content sorting method using hyperspectral spectroscopy according to an embodiment of the present invention.

[0033] From now on, with reference to FIG. 1, a sorting method based on seed component content using hyperspectral spectroscopy according to an embodiment of the present invention will be explained.

[0034] In a sorting method for seed component content using hyperspectral

[0035] Afterwards, the inflow and outflow are managed by weighing the raw materials (S13), the raw materials are sorted by sorting the soybeans (S14), and stones contained in the raw materials are removed using a soybean stone separator (S15). In addition, after separating the outer shells and foreign substances from the raw materials (S16), foreign substances such as soil and dust are removed from the surface of the raw materials using a polisher (S17).

[0036] Afterwards, off-white foreign substances mixed in the raw material are sorted using a color sorter (S18), the component content of the raw grain is classified using hyperspectral analysis of a content analyzer (S19), the grain is weighed using an automatic weighing machine, and then packaged using a sewing machine and a plastic sealer (S20).

[0037] Accordingly, the seed sorting method utilizing hyperspectral

[0038] FIG. 2 is a flowchart illustrating a sorting method based on seed component content using hyperspectral

[0039] In addition, FIG. 4 is a diagram illustrating a data analysis method according to an embodiment of the present invention, FIG. 5 is a diagram illustrating the statistical results of a PLSR model according to an embodiment of the present invention, and FIG. 6 is a diagram illustrating the prediction data of a PLSR model according to an embodiment of the present invention.

[0040] In addition, FIG. 7 is a diagram illustrating a method for reclassifying seeds according to an embodiment of the present invention, and FIGS. 8 and 9 are diagrams illustrating the results of a sorting method based on seed component content using hyperspectral spectroscopy according to an embodiment of the present invention.

[0041] From now on, a sorting method based on seed component content using hyperspectral spectroscopy according to an embodiment of the present invention will be explained with reference to FIGS. 2 to 9.

[0042] First, the raw material purification unit removes contaminants and foreign substances from the raw material and selects uniform seeds through seed size classification and washing (S110). That is, the raw material is fed into the raw material purification unit of the seed component content classification device shown in Fig. 3 (a), and contaminants and foreign substances are removed, and uniform seeds are selected through seed size classification and skin washing.

[0043] Afterwards, as shown in Fig. 3(b), the spectroscopic imaging unit captures the spectroscopic image of the seed free-falling through the raw material descent chute (S120). To explain in more detail, the selected seed is moved to the spectroscopic imaging unit and captures the spectroscopic image while free-falling along a 10-20 variable line with a width of 10-20 mm.

[0044] At this time, as shown in Fig. 3(c), the spectroscopic imaging unit captures the spectroscopic image at a wavelength set to correspond to the component content of the seed, and at this time, the spectroscopic imaging unit can capture the spectroscopic image at a wavelength selected to match the component content in the wavelength range of 400 nm to 2500 nm.

[0045] Afterwards, as shown in Fig. 3(d), the data analysis unit analyzes the spectroscopic image of the seed using a PLSR (Partial Least Square Regression) model (S130). At this time, the captured data is verified by the data analysis unit using the PLSR model within 1 second to have an accuracy of at least 80% in component content.

[0046] At this time, referring to Fig. 4, the shooting and analysis of the spectroscopic image are explained in more detail. First, the spectroscopic image captured by the spectroscopic image capturing unit corrects the collected hyperspectral image using HIS Hypercube, selects the ROI area of ​​the corrected image to extract the spectroscopic image (spectrum data), and the data analysis unit analyzes it using PLSR.

[0047] Figure 5 illustrates the statistical results of such a PLSR model, and Figure 6 illustrates the predicted data for the crude protein content of soybean seeds of the PLSR model, with Figure 6 (a) showing the predicted data of SWIR-HIS and Figure 6 (b) showing the predicted data of HELIOS EQ32.

[0048] Afterwards, as shown in (e) of FIG. 3, the result classification unit uses the results of analyzing the spectroscopic image to select seeds with insufficient component content among the seeds by means of air compression and injection, and classifies them into seeds with insufficient component content and seeds with sufficient component content (S140).

[0049] At this time, the result classification unit can select seeds with insufficient component content within 0.5 seconds using an air compression and injection method according to accuracy, and classify them into seeds with insufficient component content and seeds with sufficient component content.

[0050] Afterwards, the resetting of the PLSR model in the data analysis unit and the reclassification of the seeds with insufficient component content in the result classification unit are repeated (S150).

[0051] To explain in more detail, the result classification unit classifies the soybeans according to the protein components contained in the soybeans by using the results of analyzing the spectroscopic image of the soybean seed. At this time, the PLSR model of the data analysis unit is reset, and the result classification unit can classify the soybeans with insufficient protein content into high protein, medium protein, and low protein by repeating the reclassification of the soybeans.

[0052] That is, by referring to Figure 7, the accuracy of the selection can be verified by confirming the correlation between the selected high-protein soybeans and the actual component content. To this end, the soybeans are classified into three stages—high protein, medium protein, and low protein—based on the protein content of the soybeans, and a total of three selections are performed. In the case of high protein, the soybeans are selected twice to ensure high-quality soybean raw materials, while in the case of medium protein and low protein, the soybeans are selected only once each.

[0053] Figures 8 and 9 illustrate the results of sorting by component in this manner.

[0054] Accordingly, as shown in (e) of FIG. 3, the automatic weighing unit repeatedly reclassifies the seeds with insufficient component content and weighs the seeds according to component content, consisting of the classified seeds and the seeds that meet the content, and packages the seeds according to component content that have been weighed in this way (S160).

[0055] Thus, according to the seed component content sorting method of the present invention, seeds weighed according to seed component content can be provided through the input of raw materials, removal of foreign substances, spectral image analysis of seeds, classification of seeds according to component content, and weighing operation.

[0056] In addition, according to the present invention, since a uniform purification process can be performed from the raw grain, a uniform food content can be derived, and by modifying the verification model, a classification of various crops and their respective component contents can be provided.

[0057] FIG. 10 is a configuration diagram of a seed component content classification device according to an embodiment of the present invention, and FIG. 11 is a diagram for explaining a seed component content classification device according to an embodiment of the present invention.

[0058] From now on, with reference to FIGS. 10 and FIGS. 11, a classification device according to an embodiment of the present invention will be described.

[0059] A seed component content classification device (100) according to one embodiment of the present invention comprises a raw material purification unit (110), a spectroscopic image capturing unit (120), a data analysis unit (130), a result classification unit (140), and an automatic weighing unit (150).

[0060] The above raw material purification unit (110) removes contaminants and foreign substances from the raw material and selects uniform seeds through seed size classification and washing, and the above spectroscopic image capturing unit (120) captures spectroscopic images of seeds free-falling through the raw material descent chute (121).

[0061] At this time, illumination is irradiated through a halogen lamp (122), and the spectroscopic image capturing unit (120) captures the spectroscopic image at a wavelength set to correspond to the component content of the seed, and more specifically, can capture the spectroscopic image at a wavelength selected to match the component content in the wavelength range of 400 nm to 2500 nm.

[0062] The data analysis unit (130) analyzes the spectroscopic image of the seed through a PLSR (Partial Least Square Regression) model.

[0063] In addition, the above-mentioned result classification unit (140) uses the results of analyzing the above-mentioned spectroscopic image to select seeds with insufficient component content among the seeds by means of air compression and injection, and classifies them into seeds with insufficient component content and seeds with sufficient component content.

[0064] At this time, the data analysis unit (130) can reset the PLSR model, and accordingly, the result classification unit (140) can repeat the reclassification of seeds with insufficient component content.

[0065] As a more specific example, the result classification unit (140) classifies the soybeans according to the protein components contained in the soybeans by using the results of analyzing the spectroscopic image of the soybeans that are the seeds. At this time, by resetting the PLSR model of the data analysis unit (130), the result classification unit (140) can repeatedly reclassify soybeans with insufficient protein content into high protein, medium protein, and low protein.

[0066] Accordingly, the automatic weighing unit (150) can repeatedly reclassify the seeds with insufficient component content and weigh the seeds according to component content, consisting of the classified seeds and the seeds that have reached the content, so that they can be packaged.

[0067] As such, the seed component content classification device according to the present invention can perform high-speed, large-scale classification of seeds by component content using hyperspectral wavelengths, and by utilizing spectral characteristics to verify component content prediction models such as protein and lipids in the raw grain state and by compressing and extruding air, it is possible to remove seeds with insufficient component content, and thus the sorting power of the raw grain is over 95%, allowing for processing of over 100 kg per hour, so it can be performed at high speed and large scale and utilized in rice processing complexes (RPCs) and soybean processing complexes (SPCs).

[0068] In addition, the seed component content classification device according to the present invention provides a sorting function based on component content, thus offering a wide range of applications. More specifically, it can be utilized for various seeds such as rice, soybeans, wheat, barley, and coffee, and can classify them by component, including functional substances such as proteins, lipids, and carbohydrates, allowing for diverse applications such as high-protein tofu, soy milk, the baking industry, wheat flour, and the coffee industry.

[0069] In the detailed description of the present invention as described above, specific embodiments have been described. However, various modifications are possible within the scope of the present invention. The technical concept of the present invention should not be limited to the aforementioned embodiments, but should be defined by the claims as well as equivalents thereof.

Claims

Claim 1 A method for sorting seeds by component content using hyperspectral Claim 2 A method for sorting seeds by component content using hyperspectral spectral spectral spectral imaging according to claim 1, wherein the spectral imaging unit captures the spectral image at a wavelength set to correspond to the component content of the seed. Claim 3 A method for sorting seeds by component content using hyperspectral spectral spectral spectral imaging according to claim 2, wherein the spectral imaging unit captures the spectral image at a wavelength selected according to the component content in the wavelength band of 400 nm to 2500 nm. Claim 4 A method for sorting seeds by component content using hyperspectral Claim 5 A method for sorting seeds by component content using hyperspectral spectral spectral positivity, wherein the step of repeating the reclassification is characterized by resetting the PLSR model of the data analysis unit and the result classification unit repeating the reclassification of soybeans with insufficient protein component content to classify them into high protein, medium protein, and low protein. Claim 6 A seed component content classification device comprising: a raw material purification unit that removes contaminants and foreign substances from the raw material and selects uniform seeds through seed size classification and washing; a spectroscopic image capturing unit that captures spectroscopic images of seeds free-falling through a raw material descent chute; a data analysis unit that analyzes the spectroscopic images of the seeds captured through a PLSR (Partial Least Square Regression) model; and a result classification unit that classifies seeds with insufficient component content and seeds with sufficient component content by selecting seeds with insufficient component content through an air compression and injection method using the results of analyzing the spectroscopic images, wherein the data analysis unit resets the PLSR model and the result classification unit repeats the reclassification of the seeds with insufficient component content. Claim 7 A seed sorting device utilizing hyperspectral spectral spectral content, characterized by further including an automatic weighing unit according to weight for weighing seeds by component content, comprising seeds classified by repeating the reclassification of seeds with insufficient component content and seeds that have reached the content. Claim 8 A seed sorting device utilizing hyperspectral spectral content, wherein, in claim 6, the spectroscopic image capturing unit captures the spectroscopic image at a wavelength set to correspond to the component content of the seed. Claim 9 A seed component content sorting device utilizing hyperspectral spectral spectral spectral imaging according to claim 8, wherein the spectroscopic imaging unit captures the spectroscopic image at a wavelength selected according to the component content in the wavelength band of 400 nm to 2500 nm. Claim 10 A seed component content sorting device utilizing hyperspectral spectral spectral 6, wherein the result classification unit classifies according to the protein components contained in the soybean by using the result of analyzing a spectroscopic image of the soybean seed. Claim 11 A seed sorting device utilizing hyperspectral spectral content, wherein, in claim 10, the result classification unit repeatedly reclassifies soybeans with insufficient protein content into high protein, medium protein, and low protein as the data analysis unit resets the PLSR model.

Images