Seaweed quality evaluation method and quality evaluation system of the same

Abstract

To provide a seaweed quality evaluation technique capable of easily and rapidly evaluating a quality of seaweed in a visible-light environment such as a fishing port or a fish market.SOLUTION: A seaweed quality evaluation system 10 evaluates a quality of seaweed, and includes a database construction unit 20, a simple measurement device 40, a camera-equipped mobile terminal 50, a network via the Internet, and an analysis unit 30. The database construction unit 20 includes a camera, a biochemical index measurement unit, a chlorophyll meter, a color tone measurement unit, a database configured to include physical quantities of seaweed, biochemical indices, SPAD values, b* values, and image data, and an evaluation model learning unit that extracts feature quantities from the image data by using machine learning and learns an evaluation model by associating the feature quantities with the physical quantities of seaweed, the biochemical indices, the SPAD values, and the b* values.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a seaweed quality evaluation technique for evaluating the quality of seaweed. [Background technology]

[0002] Regions blessed with rich marine environments and nutrient-rich ocean currents, such as the Sanriku region, are major producers of a variety of seaweeds, including wakame, konbu, matsumo, and hijiki. The region's seaweed industry represents a model of environmentally sustainable fishing, with local fishing communities harvesting seaweed while carefully protecting natural resources. Economically, seaweeds such as wakame and konbu play an important role, becoming major exports for the region. In recent years, advances in cultivation techniques and the development of new farming methods have made it possible to produce high-quality seaweed, contributing to the local economy, ecosystems, and even Japanese food culture.

[0003] A sales system for seaweed is well established, and when fishermen ship their seaweed, inspectors from each fishing cooperative grade and evaluate the quality, maintaining consistent and high quality. For example, in order to ensure a stable supply of high-quality wakame to consumers, local fisheries associations promote the branding of locally produced wakame through a system in which first-grade wakame caught is certified as brand wakame. However, the grading criteria under the seaweed inspection standards are based on the subjective knowledge gained from many years of experience by inspectors (mainly fishing cooperative employees) at the production site and purchasers, and there are very few examples of quantitative methods for comparing and evaluating the color and physical quantities of seaweed.

[0004] In particular, inspector judgment is a traditional approach to seaweed quality assessment and has played an important role for many years. However, this method is highly subjective, resulting in variations in quality standards among inspectors and challenges in consistency and reproducibility. Furthermore, inspector skills depend on experience and knowledge, which are difficult to transfer to new generations, affecting the continuity of quality control. Furthermore, the transfer of these evaluation skills to the next generation of personnel is a time-consuming process that is inefficient in mass-production environments and may result in production delays. Traditional judgment is based on visual and tactile perception and lacks evaluation indicators that ensure other important quality indicators, such as chemical quality and nutritional composition. It is susceptible to inspector fatigue and personal bias, which can lead to inconsistent evaluation accuracy due to long evaluation work periods.

[0005] In addition, evaluation systems based on expert judgement have limited scalability, making it difficult to quickly respond to large-scale production or new market demands. Furthermore, the quality of seaweed processed products is greatly influenced by the variety of raw seaweed, the production area, the marine environment, and the production time, so there are no unified evaluation standards. Traditionally, evaluations have relied on the experience and sensory abilities of inspectors, but the number of experienced inspectors is decreasing, so in recent years there has been a demand for the introduction of evaluation technologies that support objectivity and speed through mechanization and quantification (see Patent Document 1).

[0006] The method for assessing the quality of laver described in Patent Document 1 measures the amount of volatile sulfur-containing compounds contained in laver and assesses the quality of laver based on the measured value. Specifically, the proportion of methyl mercaptan in all volatile sulfur-containing compounds contained in laver is measured to determine the degree of quality deterioration of laver.

[0007] However, in the method for assessing the quality of laver described in Patent Document 1, the proportion of methyl mercaptan in the total volatile sulfur-containing compounds of the laver to be evaluated cannot be measured simply and quickly on the spot at the fishing port where the laver is landed or at the fish market. Moreover, the method for assessing the quality of laver described in Patent Document 1 cannot be applied to the quality evaluation of seaweed. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Application Publication No. 62-79759 Summary of the Invention [Problem to be solved by the invention]

[0009] In view of the above, an object of the present invention is to provide a seaweed quality evaluation technique that can easily and quickly evaluate the quality of seaweed under visible light environments such as fishing ports and fish markets. [Means for solving the problem]

[0010] [1] A method for evaluating the quality of seaweed, comprising: A database construction process for constructing a database from the physical quantities of the seaweed obtained by measuring the seaweed for learning, the biochemical indicators obtained by measuring the seaweed for learning, the SPAD value obtained by measuring the chlorophyll content of the seaweed for learning, the b* value indicating the hue and saturation obtained by measuring the seaweed for learning, and learning image data obtained by photographing the seaweed for learning; An evaluation model learning process in which feature quantities representing color, reflected light amount, and surface shape are extracted from the image data as training data of the database using machine learning, and an evaluation model is learned by associating the seaweed thickness, the biochemical index, the SPAD value, and the b* value; The method is characterized by comprising a seaweed quality evaluation process for evaluating the quality of the seaweed to be evaluated by estimating the evaluation model that is closest to the physical quantities of the seaweed to be evaluated and the captured image data for evaluation.

[0011] According to this configuration, a database construction process is included in which a database is constructed from the physical quantities of the seaweed obtained by measuring the learning seaweed, the biochemical indices obtained by measuring the learning seaweed, the SPAD values ​​obtained by measuring the chlorophyll content of the learning seaweed, the b* values ​​indicating hue and saturation obtained by measuring the learning seaweed, and learning image data of the learning seaweed.Therefore, rather than relying on the intuition of an expert craftsman, the physical quantities of the seaweed obtained by measuring the seaweed, the biochemical indices, the SPAD values, the b* values, and the image data can be matched, and the quality of the seaweed, including its freshness, can be evaluated using the biochemical indices, SPAD values, and b* values. The method includes an evaluation model training process that trains an evaluation model from the physical quantities of the seaweed and feature quantities representing the SPAD value, which indicates chlorophyll content, and the b* value, which indicates hue and saturation, extracted using machine learning from image data as training data in a database, and a seaweed quality evaluation process that evaluates the quality of the seaweed by estimating the most similar evaluation model from the physical quantities of the seaweed to be evaluated and the photographed image data for evaluation.This enables easy and rapid evaluation of seaweed quality using artificial intelligence in visible light environments such as fishing ports and fish markets.In this embodiment, the physical quantities are defined separately from the SPAD value, L* value, a* value, b* value, etc., and the physical quantities are at least one of the thickness of the wakame seaweed and its breaking strength.

[0012] [2] Preferably, the database construction step includes: The biochemical indicators include one or more of the content of fucoxanthin, which is a polysaccharide, and the content of alginic acid, and the method includes a biochemical indicator measurement sub-step of measuring these biochemical indicators using a measuring device.

[0013] According to this configuration, the biochemical indicators include one or more of the content of fucoxanthin, which is a polysaccharide, and the content of alginic acid, and these biochemical indicators are obtained by measuring them using measuring equipment in the biochemical indicator measurement sub-process. Therefore, there is no individual variation as occurs in sensory tests, and the same evaluation results can be obtained regardless of who evaluates the seaweed, making it possible to achieve stable and highly accurate quality evaluation.

[0014] [3] Preferably, in the seaweed quality evaluation step, a camera-equipped mobile terminal, a simple measuring device, and an evaluation terminal are used, The simplified measuring device includes a portable main body, a plurality of slits formed in the main body with different widths for inserting the seaweed to measure the thickness, a seaweed holding section provided in the main body for pulling and holding the cut pieces of the seaweed, and a color chart of different standard colors arranged in a plurality of regions on the surface of the main body. The physical quantity of the seaweed to be subjected to quality evaluation is measured using the slit, and the seaweed for evaluation is photographed together with the color chart using the mobile terminal with a camera. The quality of the seaweed is evaluated by performing arithmetic processing on the physical quantities of the seaweed measured by the evaluation terminal and the photographed image data for evaluation.

[0015] According to this configuration, the quality evaluation process uses a mobile device with a camera, allowing for inexpensive and easy photography of the seaweed to be evaluated. Furthermore, the quality evaluation process uses a simple measuring device, which is easy to carry, and the thickness of the seaweed to be evaluated is measured using a slit in the simple measuring device, allowing for easy on-the-spot measurement of the seaweed thickness. Furthermore, the seaweed to be evaluated is photographed using a color chart containing multiple standard colors for the simple measuring device. Therefore, even if the image data is taken outdoors or indoors with different light levels, it can be used as the basis for color adjustment, allowing for correction of differences in the color temperature of the lighting, resulting in accurate color adjustment and improved accuracy in seaweed quality evaluation.

[0016] [4] Preferably, a seaweed quality evaluation system for evaluating the quality of seaweed, The system includes a database construction unit, a simple measuring device, a mobile terminal with a camera, an internet network, and an analysis unit, The database construction unit includes a camera for photographing the seaweed, a physical quantity measurement unit for measuring physical quantities of the seaweed, a biochemical index measurement unit for measuring biochemical indexes including at least one of the content of fucoxanthin, which is a polysaccharide of the seaweed, and the content of alginic acid, a chlorophyll meter for measuring the chlorophyll content of the seaweed and indicating a SPAD value, a color measurement unit for measuring a b* value indicating the hue and saturation of the seaweed, and a thickness of the seaweed obtained by measuring the seaweed for learning, the biochemical indexes obtained by measuring the seaweed for learning, and a learning index measurement unit for measuring the chlorophyll content of the seaweed. The system includes a database including a SPAD value obtained by measuring the chlorophyll content of the seaweed for training, a b* value indicating hue and saturation obtained by measuring the seaweed for training, and training image data obtained by photographing the seaweed for training; and an evaluation model learning unit that uses machine learning to extract features representing color, reflected light amount, and surface shape from the image data as training data in the database, and learns an evaluation model by associating the physical quantity, biochemical index, SPAD value, and b* value of the seaweed; The simplified measuring device includes a portable main body, a plurality of slits formed in the main body with different widths for inserting the seaweed to measure the thickness, a seaweed holding section provided in the main body for pulling and holding the cut pieces of the seaweed, and a color chart of different standard colors arranged in a plurality of regions on the surface of the main body. The mobile terminal with a camera takes an image of the seaweed to generate the image data, transmits the image data and data on physical quantities of the seaweed, and displays an evaluation result of the seaweed, The analysis unit evaluates the quality of the seaweed that is the subject of quality evaluation by estimating the evaluation model that is closest to the physical quantity data of the seaweed that is the subject of quality evaluation received via the network and the photographed image data for evaluation, and transmits the evaluation results to the camera-equipped mobile terminal.

[0017] According to this configuration, the database construction unit constructs a database from the thickness of the training seaweed obtained by measuring the training seaweed, the biochemical index obtained by measuring the training seaweed, the SPAD value obtained by measuring the chlorophyll content of the training seaweed, the b* value indicating hue and saturation obtained by measuring the training seaweed, and training image data obtained by photographing the training seaweed.Therefore, rather than relying on the intuition of an expert craftsman, the physical quantities of the seaweed obtained by measuring the seaweed, the biochemical index, SPAD value, b* value, and image data can be matched, and the quality of the seaweed, including its freshness, can be evaluated using the biochemical index, SPAD value, and b* value.

[0018] Furthermore, since the analysis unit evaluates the physical quantity information and image data of the seaweed acquired by the simple measuring device and the mobile device with a camera via an internet network, only simple portable equipment is required at the location where the seaweed to be evaluated is located. Furthermore, the analysis unit uses an evaluation model trained from the physical quantities of the seaweed and feature values ​​expressing the SPAD value, which indicates chlorophyll content, and the b* value, which indicates hue and saturation, extracted using machine learning from image data as training data in a database, to estimate the evaluation model that is closest to the physical quantities of the seaweed to be evaluated and the photographed image data for evaluation. Therefore, the quality of the seaweed to be evaluated can be easily and quickly evaluated using artificial intelligence in visible light environments such as fishing ports and fish markets. [Effects of the Invention]

[0019] It is possible to provide a seaweed quality evaluation technique that can easily and quickly evaluate the quality of seaweed under visible light environments such as fishing ports and fish markets. [Brief explanation of the drawings]

[0020] [Figure 1] FIG. 1 is a diagram illustrating an overview of database construction and evaluation model learning according to the present invention. [Figure 2] 1 is a diagram illustrating a simplified measuring device of the present invention. [Figure 3] FIG. 1 is a diagram illustrating an outline of wakame quality evaluation. [Figure 4]FIG. 10 is a diagram illustrating an example of wakame seaweed. [Figure 5] FIG. 1 is a diagram illustrating the parts of wakame seaweed. [Figure 6] FIG. 10 is a diagram illustrating advance preparation of wakame seaweed. [Figure 7] FIG. 1 is a diagram illustrating measurement items. [Figure 8] FIG. 1 is a diagram illustrating color tone (L* value, a* value, b* value). [Figure 9] FIG. 1 is a diagram illustrating the SPAD value (an index of chlorophyll content). [Figure 10] 10A and 10B are examples of images of seaweed photographed at different color temperatures. [Figure 11] 1 is a graph showing the breaking strength of wakame seaweed. [Figure 12] FIG. 1 is a diagram illustrating the content of polyphenols in pigment components. [Figure 13] FIG. 1 is a diagram illustrating the fucoxanthin content of salted wakame seaweed and frozen wakame seaweed. [Figure 14] FIG. 1 is a diagram illustrating the contents of fucoidan and alginic acid in salted wakame seaweed. [Figure 15] FIG. 1 is a diagram illustrating the correlation between the b* value, SPAD value, and thickness of frozen wakame seaweed. [Figure 16] FIG. 1 is a diagram illustrating the correlation between the b* value, SPAD value, and thickness of salted wakame seaweed. [Figure 17] FIG. 10 is a diagram illustrating the correlation of all data. [Figure 18] FIG. 10 is a diagram illustrating a focus point of correlation. DETAILED DESCRIPTION OF THE INVENTION

[0021] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The following describes an embodiment of the present invention with reference to the accompanying drawings, in which each device of a fresh food quality assessment system is shown conceptually (schematically). [Example]

[0022] As an example of the application of seaweed, wakame seaweed will be described below as an evaluation target. As shown in Figures 1 to 3, a wakame seaweed quality evaluation system 10 for evaluating the quality of wakame seaweed 60 includes a database construction unit 20, an internet network 11, an analysis unit 30, a simple measurement device 40, and a mobile terminal 50 with a camera.

[0023] The database construction unit 20 includes a camera 22 that photographs the wakame seaweed 60 together with a color chart 21, a physical quantity measurement unit 23 that measures the thickness of the wakame seaweed 60, a biochemical indicator measurement unit 24 that measures biochemical indicators including one or more of the content of fucoxanthin, which is a polysaccharide, and the content of alginic acid in the wakame seaweed 60, a chlorophyll meter 25 that measures the chlorophyll content of the wakame seaweed 60 and indicates the SPAD value, a color measurement unit 26 that measures the b* value, which indicates the hue and saturation of the wakame seaweed 60, and a rheometer 27 that measures the force required to break the wakame seaweed 60 (breaking strength) and the distance stretched to break. In this example, the physical quantity is defined separately from the SPAD value, L* value, a* value, b* value, etc., and the physical quantity is at least one of the thickness of the wakame seaweed and the breaking strength.

[0024] The database construction unit 20 also includes a database 28 containing the thickness of the training wakame seaweed 60 obtained by measuring the training wakame seaweed 60, the biochemical indices obtained by measuring the training wakame seaweed 60, the SPAD value obtained by measuring the chlorophyll content of the training wakame seaweed 60, the b* value among the L*, a*, and b* values ​​indicating the hue and saturation obtained by measuring the training wakame seaweed 60, and training image data obtained by photographing the training wakame seaweed 60. The database 28 also includes an evaluation model training unit 29 that uses machine learning to extract features representing color, reflected light amount, and surface shape from the image data serving as training data in the database 28, and trains an evaluation model by associating the thickness (physical quantity), strength (physical quantity), biochemical indices, SPAD value, and b* value with the wakame seaweed. The database 28 is a database created by analyzing the correlations between all of the data on the physical quantities (thickness and strength), biochemical indices, SPAD value, b* value, and image data of the wakame seaweed.

[0025] In the embodiment, the color tone measuring unit 26 is a spectrophotometer, but is not limited to this and may be a colorimeter, color difference meter, or the like.

[0026] The simple measuring device 40 also includes a portable main body 41, a plurality of slits 42 formed in the main body 41 with different widths into which the seaweed 60 is fitted to measure its thickness, a magnetic thickness indicator 42a movably arranged near the slits 42 and indicating the thickness of the seaweed 60, a pair of seaweed holding parts 43 provided on the main body 41, one of which is supported by a tension spring 44 and which pull and hold the cut pieces of seaweed 60, a spring 44 for measuring the strength (measurement of strength, which is a physical quantity) of holding and cutting the seaweed 60, a color chart 21 of different standard colors arranged in a plurality of areas on the surface of the main body 41, a length measuring part 45 for measuring the length of the seaweed 60, a magnetic length indicator 45a movably arranged on the length measuring part 45 and indicating the length of the seaweed 60, and a strength measuring part 46 for measuring the breaking strength of the seaweed 60.

[0027] The camera-equipped mobile terminal 50 is a so-called smartphone, which has a camera and a transmitting / receiving device, photographs the wakame seaweed 60 to generate image data, transmits the image data and data on the physical quantities (thickness, strength) of the wakame seaweed 60 to the analysis unit (server) 30 via the network 11, and receives and displays the evaluation results of the wakame seaweed 60 graded by the analysis unit 30 via the network 11. Note that, although the physical quantities are both thickness and strength in the embodiment, the physical quantities may be thickness only or strength only.

[0028] The analysis unit 30 is a terminal known as a server, which analyzes the correlation between data on physical quantities (thickness, strength) of the wakame 60 to be subjected to quality evaluation received via the network 11 and the captured image data for evaluation, compares the data with the accumulated database 28 to estimate the most similar evaluation model, evaluates (grades) the quality of the wakame 60 to be subjected to quality evaluation, and transmits the evaluation results to the camera-equipped mobile terminal 50. Note that in the embodiment, the physical quantity data is data on both thickness and strength, but the physical quantity may be data on thickness only or data on strength only. Furthermore, the analysis unit 30 may use the average values ​​of the thickness and strength measurement results at multiple locations on the wakame 60 as the physical quantity data.

[0029] The analysis unit 30 also has artificial intelligence (AI), which uses machine learning to extract features representing the color, reflected light amount, and surface shape of the wakame 60 from image data used as training data in the database 28, and learns a judgment model by associating these features with the thickness, biochemical index, SPAD value, and b* value of the wakame 60.

[0030] The artificial intelligence of the analysis unit 30 may be a general one. Generally, machine learning refers to a technique for learning patterns and judgment criteria from data and predicting and judging unknown things based on the learned patterns and judgment criteria, as well as an analytical technique related to artificial intelligence. Deep learning, which is a subtype of machine learning, may also be used. Deep learning is an extension of the neural network analysis method, which is a more basic and broader machine learning method, and is a method that enables highly accurate analysis and utilization. In one embodiment of the present invention, so-called supervised learning, which is machine learning in which teacher data corresponding to the correct answer is provided, is used. Note that examples of supervised learning analysis methods include regression analysis and decision trees.

[0031] Neural networks are also applied to the field of image recognition, and neural networks with two or more intermediate layers (hidden layers) are known as deep learning. Deep learning is a method of determining what to focus on in given data, extracting features that characterize the data without human instruction on specific features, and learning the rules and regularities contained in the data.

[0032] The thickness of the wakame 60 to be subjected to quality evaluation may be input numerically into the camera-equipped mobile terminal 50, but is not limited to this. The thickness indicator 42a may be placed at a position indicating the thickness of the wakame 60 near the slit 42 of the simple measuring device 40, and the wakame 60 may be photographed while being held, and the analysis unit 30 may read the position of the thickness indicator 42a from the image data to obtain thickness information of the wakame 60.

[0033] Similarly, the length of the seaweed 60 may be input numerically into the camera-equipped mobile terminal 50, but is not limited to this. For example, the length indicator 45a may be placed at a position indicating the length of the seaweed 60 in the length measuring section 45 of the simple measuring device 40, and the seaweed 60 may be photographed while being held, and the analysis section 30 may read the position of the length indicator 45a from the image data to obtain length information for the seaweed 60.

[0034] The strength measuring unit 46 of the simplified measuring device 40 measures the breaking strength by pulling the seaweed 60 while it is held by the seaweed holding unit (clip) 43 and reading the memory position of the strength measuring unit 46 at the position where the seaweed 60 breaks. When the seaweed 60 breaks, the seaweed holding unit (clip) 43 on the strength measuring unit 46 side stops at that position without moving.

[0035] Next, the functions of the wakame quality evaluation system 10 will be described. In the color analysis, 60 samples of wakame (seaweed) are placed on a flat surface next to the color chart 21, and the color is photographed and analyzed using a mobile device with a camera (smartphone) 50.

[0036] In the strength measurement, a specific mechanism consisting of a tension spring, a wakame holding part 43 and a memory is used to measure the strength, elasticity and force required to break 60 samples of wakame (seaweed).

[0037] In thickness measurement, a metal measuring device with slits (grooves) 42 of different widths is used, and a wakame (seaweed) 60 sample is inserted into the slits (grooves) 42, and the thickness is physically measured by the width of the slits (grooves) 42.

[0038] For length measurement, 60 samples of wakame (seaweed) are placed along a fixed scale (ruler) and their lengths are read.

[0039] In data processing, 60 wakame (seaweed) samples, color chart 21, and measurement data pointers and scales are photographed with a mobile device (smartphone) 50 with a camera and sent to an analysis unit (server) 30. The analysis unit (server) 30 performs image recognition, records the measurement data, and returns the evaluation (analysis) results to the user's mobile device (smartphone) 50 with a camera.

[0040] In cloud data management, measured data is uploaded to the cloud, enabling large-scale data analysis.

[0041] Next, the procedure for color analysis among the procedures for operating the wakame quality evaluation system 10 and the simple measurement device 40 will be described. To prepare 60 samples of wakame (seaweed), 60 samples of wakame (seaweed) to be measured are selected and cut to a predetermined size using a cutting tool 51 such as scissors or a die cutter.

[0042] Regarding fixing the wakame (seaweed) 60 sample, the cut wakame (seaweed) 60 sample is fixed with a wakame holder 43 such as a clip or clamp attached to the simple measuring device 40 .

[0043] The color chart 21 is placed within the shooting range. This color chart 21 includes various standard colors and serves as a reference for color adjustment.

[0044] Regarding the placement of the 60 wakame (seaweed) samples, the 60 wakame (seaweed) samples are placed flat near the color chart 21 and adjusted so that the entire 60 wakame (seaweed) samples fit within the image.

[0045] When taking a photograph using the mobile terminal 50 with a camera, a photograph is taken with the camera of the mobile terminal (smartphone) 50 with a camera so that both the wakame (seaweed) 60 sample and the color chart 21 are visible.

[0046] Regarding the transfer of image data, the photograph (image data) is transferred to the analysis unit (server) 30. This may be done via so-called Wi-Fi or mobile data communication.

[0047] Regarding color analysis, the transferred image data is analyzed by a color analysis tool provided in the analysis unit 30. Using the color of the color chart 21 as a reference, the color change and consistency of the wakame (seaweed) 60 are evaluated.

[0048] The analysis results are interpreted by the analysis unit 30, which determines the quality and maturity of the wakame (seaweed) 60 based on the information obtained from the color analysis. The analysis unit 30 analyzes specific color changes and provides information on the quality and maturity of the wakame (seaweed) 60 based on the analysis results. For example, specific color changes can be indicators of the quality and maturity of the wakame (seaweed) 60.

[0049] Next, the procedure for measuring the intensity will be described. To prepare 60 samples of wakame (seaweed), 60 samples of wakame (seaweed) to be measured are selected and cut to a predetermined size using a cutting tool 51 such as scissors or a die cutter.

[0050] To attach the wakame (seaweed) 60 sample, the cut wakame (seaweed) 60 sample is fixed with a wakame holder 43 such as a clip or clamp attached to the simple measuring device 40. It is important to position the wakame (seaweed) 60 sample so that it is pulled evenly.

[0051] To carry out the measurements, a spring balance is used to measure the strength and viscoelastic properties of the wakame (seaweed) 60 sample. During the measurement, the spring balance is stretched at a constant speed, and the response at the moment the wakame (seaweed) 60 sample breaks is recorded. This method allows for a quantitative evaluation of the physical strength and elasticity of the wakame (seaweed) 60 sample.

[0052] To read the data, read the pointer or scale displayed on the screen. If necessary, record the reading.

[0053] Next, the operating procedure for thickness measurement will be described. To prepare 60 samples of wakame (seaweed), 60 samples of wakame (seaweed) to be measured are selected and cut to a predetermined size using a cutting tool 51 such as scissors or a die cutter.

[0054] To perform the measurement, align the appropriate slit (groove) 42 on the metal measuring device with the area of ​​the wakame (seaweed) 60 sample where you want to measure the thickness. Ensure that the wakame (seaweed) 60 sample fits snugly into the slit (groove) 42. If the wakame (seaweed) 60 sample does not fit completely into the slit (groove) 42, try the next larger slit (groove) 42 until the appropriate thickness is obtained.

[0055] To read the data, the scale on the metal measuring device with the slit 42 is referenced, and the thickness of the wakame (seaweed) 60 sample aligned with the slit (groove) 42 is read. If necessary, the read value is recorded. Note that the thickness of the wakame (seaweed) 60 can be measured using a laser scanner 23, but other methods may also be used.

[0056] Next, the operating procedures for color analysis, intensity measurement, and thickness measurement will be described. To prepare 60 samples of wakame (seaweed), 60 samples of wakame (seaweed) to be measured are selected and cut to a predetermined size using a cutting tool 51 such as scissors or a die cutter.

[0057] To perform the measurement, align the appropriate slit (groove) 42 on the metal measuring device with the area of ​​the wakame (seaweed) 60 sample where you want to measure the thickness. Ensure that the wakame (seaweed) 60 sample fits snugly into the slit (groove) 42. If the wakame (seaweed) 60 sample does not fit completely into the slit (groove) 42, try the next larger slit (groove) 42 until the appropriate thickness is obtained.

[0058] Regarding the movement of the pointer, the thickness indicator 42a is moved to the determined scale.

[0059] To attach the wakame (seaweed) 60 sample, the cut wakame (seaweed) 60 sample is fixed with a wakame holder 43 such as a clip or clamp attached to the simple measuring device 40. It is important to position the wakame (seaweed) 60 sample so that it is pulled evenly.

[0060] To carry out the strength measurements, a spring balance is used to measure the strength and viscoelastic properties of the wakame (seaweed) 60 samples. During the measurements, the spring balance is stretched at a constant speed and the response at the moment when the wakame (seaweed) 60 samples break is recorded.

[0061] To check the pointer, check the scale indicated by the pointer at the moment when 60 samples of wakame (seaweed) broke.

[0062] Regarding fixing the wakame (seaweed) 60 sample, the torn wakame (seaweed) 60 sample is fixed with a wakame holder 43 such as a clip or clamp attached to the simple measuring device 40 .

[0063] The color chart 21 is placed within the shooting range. This color chart 21 includes various standard colors and serves as a reference for color adjustment.

[0064] Regarding the placement of the wakame (seaweed) 60 sample, place the wakame (seaweed) 60 sample flat near the color chart 21 and adjust it so that the entire wakame (seaweed) 60 sample and all the pointers and scales fit properly within the image.

[0065] When photographing using the mobile terminal (smartphone) 50 with a camera, the camera of the mobile terminal (smartphone) 50 with a camera is used to photograph 60 samples of wakame (seaweed), the color chart 21, and all the pointers and scales.

[0066] Regarding the transfer of image data, the photograph (image data) taken is transferred to the analysis unit (server) 30 via so-called Wi-Fi or mobile data communication.

[0067] For color analysis, the transferred image data is analyzed by a color analysis tool in the analysis unit 30. Using the colors of the color chart 21 as a reference, the color changes and consistency of 60 samples of wakame (seaweed) are evaluated.

[0068] Regarding the reading of the pointer and scale, the transferred image data is automatically read by the scale recognition tool of the analysis unit 30.

[0069] Regarding the interpretation of the analysis results, the quality and grade of the 60 wakame (seaweed) samples are evaluated based on the information obtained from the color analysis, thickness, and strength using the analysis unit 30. Based on this, information regarding the quality and grade of the 60 wakame (seaweed) samples is provided.

[0070] Regarding the transfer of result data, information regarding the measurement results, quality, and maturity level is transferred to the user's camera-equipped mobile terminal (smartphone) 50 via so-called Wi-Fi or mobile data communication.

[0071] Through the above process, quantitative information about wakame (seaweed) 60 can be obtained, which can be used for quality control and sorting.

[0072] Next, an example of the wakame seaweed 60 to be used will be described. As shown in Figure 4, the sample wakame (seaweed) 60 used was Sanriku wakame 60 harvested from February to May 2022, stored at a salting temperature of 4°C, and boiled during the processing process (hereinafter referred to as salted), and Sanriku wakame 60 harvested from February to May 2022, stored in a freezer at -30°C, and not boiled (hereinafter referred to as frozen).

[0073] Next, the parts of the wakame seaweed 60 will be described. As shown in Figure 5, measurements were taken near the stem and at the top, middle, and bottom of the leaf, and the average was calculated.

[0074] Next, we will explain the preparations for wakame. As shown in FIG. 6, wakame seaweed 60 is rehydrated with 20 ml of distilled water equal to the mass of wakame seaweed 60, and the rehydrated wakame seaweed 60 is cut into pieces of 3 cm x 2 cm using a cutting tool 51, with 20 pieces prepared for each section.

[0075] Next, the measurement items will be explained. As shown in Figure 7, physical quantities such as the thickness and strength of the wakame 60, biochemical indicators such as polyphenols and fucoxanthin, the SPAD value which is an indicator of chlorophyll content, image data including the wakame 60, the breaking strength of the wakame 60, and the color tone of the wakame 60 (L* value, a* value, b* value) are measured.

[0076] Next, the color tone (L* value, a* value, b* value) will be explained. As shown in FIG. 8, the L* value indicates brightness, the positive side of the a* value indicates red, the negative side indicates green, and the positive side of the b* value indicates yellow, and the negative side indicates blue.

[0077] Next, we will explain the SPAD value (an indicator of chlorophyll content). As shown in Figure 9, the calculation was performed using 650 nm, the wavelength of light absorbed by chlorophyll, and 940 nm, the wavelength of light not absorbed, as a control. Wakame 60 is rich in chlorophyll and is known to have antioxidant and anti-cancer effects.

[0078] Next, examples of images of wakame seaweed photographed at each color temperature will be described. As shown in Figure 10, image data was obtained by using a visible light camera to cut out 10mm x 30mm pieces of leaves and stems of wakame (seaweed) 60 obtained through different processing methods (blanching, salting) and storage conditions (temperature, time), and photographing them under different color temperatures (3300K, 3700K, 4100K, 4500K, 4900K, 5300K, 5600K). Simultaneously with the creation of the image data, measurements of the physical quantities, optical properties, and biochemical indices of wakame (seaweed) 60 were taken, and the obtained data were associated with the image data of wakame (seaweed) 60.

[0079] Next, a graph showing the breaking strength of wakame 60 will be described. As shown in Figure 11, specific physical quantities, such as strength and viscoelastic properties, were measured using a CR-100 Rheometer (Sun Scientific Co., Ltd.) to record the breaking strength (g) of individual leaf pieces (10 mm x 30 mm) when pulled at a speed of 300 mm per minute. Measurements were further divided into upper, middle, and lower sections for 60 wakame (seaweed) leaves and stems, and leaf thickness was measured at the measurement site using a digital thickness gauge (Neoteck). Similarly, measurements were further divided into upper, middle, and lower sections for the seaweed leaves and stems. The strength and thickness results suggest that thickness and strength characteristics vary significantly depending on the place of origin, suggesting their usefulness as indicators for distinguishing between origins and grades.

[0080] For optical properties, chlorophyll was measured using a chlorophyll meter SPAD-502Plus (Konica Minolta), and L*, a*, and b* color space values ​​were calculated using a handheld colorimeter NR-12A (Nippon Denshoku Industries Co., Ltd.).

[0081] Freeze-dried powdered seaweed was used for component analysis of biochemical indicators. Fucoidan and alginic acid were measured after acid or alkaline extraction using standard methods. The Folin-Ciocalteu method was used to measure total polyphenol concentration. Fucoxanthin was extracted using the same method as polyphenols, and the absorbance value at 449 nm was measured to calculate the content. The results of the biochemical analysis of seaweed from each production area were separated into leaves and stems, and the average values ​​for the top, middle, and bottom parts are shown.

[0082] Next, the content of polyphenols, which are pigment components, will be described. As shown in FIG. 12, the thallus of both salted and frozen wakame 60 contains a large amount of polyphenol, a pigment component.

[0083] Next, the fucoxanthin contents of salted wakame and frozen wakame will be explained. As shown in FIG. 13, the thallus of both salted and frozen wakame 60 contains a large amount of the pigment component fucoxanthin.

[0084] Next, we will explain the contents of fucoidan and alginic acid in salted wakame. As shown in FIG. 14, both salted and frozen wakame 60 contain sufficient amounts of the functional components fucoidan and alginic acid.

[0085] Next, we explain the correlation between the b* value, SPAD value, and thickness of frozen wakame. As shown in Figure 15, there is a high correlation between the b* value and the SPAD value, and there is also a high correlation between thickness and strength. From data including images, it is possible to obtain data on the b* value and thickness, and to determine SPAD (chlorophyll content) and strength (texture).

[0086] Next, we will explain the correlation between the b* value, SPAD value, and thickness of salted wakame. As shown in FIG. 16, it can be seen that there is a high correlation between the b* value and the SPAD value of salted wakame.

[0087] Next, the correlation of the entire data will be described. As shown in Figure 17, the overall data shows that there is a correlation between SPAD value and strength (texture).

[0088] Next, the focus of correlation will be described. As shown in Figure 18, there is a correlation between the b* value and the SPAD value, and between thickness and strength. It has been found that by obtaining color tone (L* value, a* value, b* value) and thickness data from image data, it is possible to evaluate the chlorophyll content and texture.

[0089] Next, learning of a type discrimination model using a mechanical technique in the above-described wakame quality evaluation system 10 and wakame quality evaluation method will be described.

[0090] Data selection involves selecting a large amount of data from the database to be used to train a model for discriminating between 60 types of wakame (seaweed). This process includes images, physical quantities, optical properties, and biochemical indicators of seaweed of different types and qualities. The selection process is carried out carefully, as the quality and diversity of the data directly affect the performance of the model.

[0091] In data preprocessing, the collected data must be preprocessed appropriately. At this stage, the analysis unit 30 performs processes such as image resizing, normalization, and noise reduction based on the color chart information in the image stored in the database.

[0092] In the feature extraction, features that are important for discrimination are extracted from the image by the analysis unit 30. In this process, particular importance is placed on features such as the color and texture of the wakame (seaweed) 60.

[0093] Model selection involves choosing an appropriate machine learning model. To distinguish between 60 types of wakame (seaweed), various models are candidates, including convolutional neural networks (CNNs), support vector machines (SVMs), decision trees, random forests, and linear discriminant analysis (LDA). The architecture and parameters of these models are selected to suit the specific task.

[0094] In model training, the model is trained using training data. In the training process, the goal is for the model to learn patterns from the data, and parameters are adjusted to minimize the loss function. It is also important to apply regularization techniques to prevent overfitting.

[0095] Model evaluation involves assessing the model's performance using test data to ensure proper operation and making adjustments as necessary.

[0096] Model deployment involves integrating models that have proven to perform satisfactorily into the actual selection process, which involves testing and tuning in a real environment.

[0097] After deployment, models must be maintained and improved to adapt to changes in the data and new challenges, including procedures such as collecting new data and retraining the model.

[0098] Learning a model to identify 60 types of wakame (seaweed) is a crucial step in achieving accurate species identification. Appropriate data collection, preprocessing, feature extraction, model selection, training, evaluation, deployment, and ongoing maintenance are essential elements for building a high-performance sorting model. As an example, as shown in Figure 9, we extracted only color features in the L*, a*, and b* color spaces from image data and used the most basic linear discriminant analysis (LDA) to prototype a system for identifying the origin of seaweed. This method achieved an accuracy rate of approximately 0.7566.

[0099] Next, the formulation of the grading criteria will be explained. There are several important steps that must be taken to develop grading standards. These standards are used to evaluate the quality, consistency, and market value of a particular wakame (seaweed). Below are the steps to developing grading standards:

[0100] Market research and standard setting involves analyzing consumer demand, market preferences, and competitor product standards. Standards are set taking into account the uses of wakame (seaweed) 60 and consumer expectations.

[0101] The selection of quality indicators will include key quality indicators of wakame (seaweed) 60, including physical quantities (strength, thickness, texture, etc.), optical properties (color, transparency, etc.), and biochemical indicators (nutritional components, storage conditions, etc.).

[0102] Data analysis will involve analyzing data collected from a seaweed database constructed under various production areas, processing methods, and storage conditions, and evaluating correlations with quality indicators.

[0103] Standard values ​​are established for each indicator, including the minimum quality level that must be met and the range of optimal quality levels.

[0104] Grade classification categorizes seaweed into different grades based on benchmark values ​​(e.g., premium, standard, economy).

[0105] Piloting and evaluation involves applying the standards to actual samples, conducting a pilot test, evaluating the effectiveness and applicability of the standards, and adjusting them as necessary.

[0106] Stakeholder consultations will be conducted with producers, processors, retailers and consumers to ensure the appropriateness of the standards. The standards will be periodically reviewed and updated as necessary in response to market changes, consumer preferences and technological advances.

[0107] Modeling of the judgment criteria involves modeling the criteria for judging the grade standards of wakame (seaweed) 60 and implementing the model in the actual judgment process.

[0108] As such, it is important that the formulation of grading standards be transparent, consistent, and objective, and that they adapt to market needs and technological advances.

[0109] Next, the determination of the grade standard using the judgment model will be explained. After determining the type of wakame (seaweed) 60, the step of determining the grade standard of that type of seaweed using the determination model criteria proceeds as follows.

[0110] The species are identified using a discriminant model, which identifies 60 species of wakame (seaweed).

[0111] The application of grading criteria involves applying predefined grading criteria to 60 identified species of wakame (seaweed), which are based on multiple parameters, including physical quantities, optical properties, and biochemical indicators.

[0112] The data comparison involves comparing actual data (physical quantities, optical properties, biochemical indicators, etc. obtained from the developed equipment) of wakame (seaweed) 60 with the grading standard criteria.

[0113] The grade assignment is based on the comparison results and assigns the appropriate grade to wakame (seaweed) 60.

[0114] Recording results and feedback: Recording the grades assigned. Collect feedback on whether the grades were appropriate and use it to improve the model.

[0115] Sorting and classification involves sorting 60 pieces of wakame (seaweed) and classifying them into different categories based on the grades determined.

[0116] The results are displayed on the user's camera-equipped mobile terminal (smartphone) 50 via Wi-Fi or mobile data communication.

[0117] Next, the spectral analysis of wakame 60 using an optical filter will be described. When applying spectral analysis to compare the components of wakame seaweed, optical filters can be used. The selected optical filter is placed on the wakame seaweed, and visual observation or camera photography is performed. If different types of seaweed exhibit distinct differences in absorption and reflection characteristics in specific wavelength ranges of visible light or near-infrared light, the combination of appropriate optical filters and spectral analysis techniques makes it possible to comparatively analyze seaweed components. The simple device is equipped with appropriate optical filters.

[0118] However, it should be noted that optical filters can be broadly divided into two types: absorption and interference, and that they differ in their operating principles, transmittance characteristics, and response to environmental factors. Absorption filters are made of colored glass or gelatin filters that selectively absorb specific wavelength ranges, and have the advantage of being relatively inexpensive and easy to use. On the other hand, interference filters use the interference of light through multilayer films to transmit only specific wavelengths, resulting in high transmittance and sharp transmission characteristics, but they tend to be more expensive.

[0119] When selecting an optical filter, it is necessary to consider the characteristics of the seaweed to be analyzed and the experimental environment. For example, phycoerythrin, which is found in large amounts in red algae, has an absorption peak around 495 nm, while chlorophyll a in green algae has an absorption peak around 675 nm. By selecting an optical filter with these wavelength ranges in mind, it is possible to detect the target components more effectively. It is also important to consider the environmental resistance of the optical filter. (In particular, interference filters have large fluctuations in their characteristics due to changes in temperature and humidity, so controlling the experimental environment is important for stable measurements.)

[0120] When using a simple analyzer, optical filters are usually preset, but an appropriate optical filter should be selected taking into consideration the analysis target and experimental conditions to perform an accurate comparative analysis of seaweed components.

[0121] Follow the steps below to install it into the device. In preparing the sample, the wakame sample to be measured is selected and cut to a predetermined size using scissors or a die cutter.

[0122] When placing the sample, place the cut wakame sample flat on the measurement stage attached to the device, taking care to ensure that the surface of the sample is uniform.

[0123] When selecting and arranging an optical filter, select an appropriate optical filter taking into consideration the absorption spectrum of the pigment components of wakame (e.g., chlorophyll a, phycoerythrin, etc.). Place the selected filter on the optical path of the device.

[0124] Measurement: The device's light source is turned on and the wakame sample is irradiated with light. The light passes through the filter, is reflected and transmitted by the sample, and is photographed with the smartphone camera.

[0125] To read the data, the light spectrum data captured by the smartphone camera is transmitted to a server via Wi-Fi or mobile data communication.

[0126] In spectral analysis, light spectrum data is analyzed using a specialized color analysis tool. Using the colors on a color chart as a reference, the pigment composition of the wakame is compared and analyzed based on the shape and intensity of the spectrum.

[0127] In interpreting the analytical results, the obtained spectral data is analyzed and the spectra of different samples are compared to clarify the differences in components.

[0128] Next, the measurement of the physical quantities of wakame will be explained. Using the simplified analytical equipment, tests on the physical properties of wakame seaweed include tensile tests, compression tests, shear tests, stem and leaf separation tests, and friction tests. The tests analyze the tensile strength, recovery force, shear strength, stem and leaf separation force, and friction coefficient of salted wakame seaweed under different loading speeds and thickness conditions.

[0129] As a result, the average tensile strength, recovery force, shear strength, and stem-leaf separation strength of the wakame stems and leaves, as well as the friction coefficients between salted wakame and stainless steel, silicone rubber, and vulcanized rubber, are determined. Increasing the thickness of salted wakame improves its mechanical properties, while increasing the loading rate leads to a decrease in its mechanical properties.

[0130] Next, the procedure for measuring the physical quantities of wakame is summarized as follows. (1) Sample preparation a. Select the wakame sample to be measured and cut it to the specified size (e.g., 10 mm x 30 mm) using scissors or a mold cutter. b. Prepare the leaves and stems separately. c. Measure and record the thickness of the sample.

[0131] (2) Tensile test a. Fix the cut wakame sample with the clips or clamps provided on the device. b. Apply a tensile force at a constant rate and record the maximum load until breakage occurs. c. Measurements are taken at the top, middle and bottom of each leaf and stem.

[0132] (3) Compression test a. Place the wakame sample on a flat table and apply a compressive force at a constant rate. b. Record the load when the sample reaches a certain strain (e.g., 50%). c. Measure the thickness of the sample after the load is removed to evaluate the recovery force.

[0133] (4) Shear test a. Fix the wakame sample with a holding part such as a clip or clamp. b. Set the hardness tester probe so that it touches the surface of the wakame. c. Press the probe of the hardness tester against the surface of the wakame and apply pressure. d. Apply pressure at a constant force for a certain period of time until the wakame is crushed. e. Record the hardness tester reading (maximum load) when the wakame is crushed.

[0134] (5) Stem and leaf separation test a. Secure the seaweed at the boundary between the stem and the leaf with a clip. b. Apply a tensile force at a constant rate and record the maximum load required to separate the stem and leaf.

[0135] (6) Friction test a. Place the wakame sample on a flat table and place plates made of different materials (stainless steel, silicone rubber, vulcanized rubber, etc.) on top of it. b. Apply a horizontal force at a constant rate and record the load when the plate begins to slide. c. Calculate the coefficient of friction.

[0136] (7) Data analysis a. The load-displacement curve obtained from each test is photographed with a smartphone camera, and the images are transferred to a server via Wi-Fi or mobile data communication. b. The transferred images are analyzed using a dedicated physical quantity analysis tool developed to calculate physical quantities such as tensile strength, recovery force, shear strength, stem and leaf separation force, and friction coefficient. c. Evaluate the change in physical quantities due to sample thickness and loading rate. d. Compare the differences in physical quantities between stem and leaf parts.

[0137] Next, a measurement method using spectroscopy as a physical quantity will be described. There are many different types of spectroscopy, each with its own unique properties and therefore used for different applications.

[0138] Absorption spectroscopy: A method of irradiating a sample with light and measuring the amount of light absorbed by the sample at each wavelength. The types of light used include ultraviolet, visible, and infrared. By analyzing the intensity of the light transmitted or reflected by the sample, the concentration and properties of the substance can be estimated.

[0139] Fluorescence spectroscopy: A method of detecting the fluorescence emitted by a substance by irradiating it with light in the short wavelength region, such as ultraviolet light. The fluorescence that occurs is specific to the substance, and this characteristic is used to analyze the substance qualitatively and quantitatively.

[0140] Raman spectroscopy: A technique in which a laser beam is irradiated onto a sample and the wavelength and intensity of the light scattered from the sample are analyzed. Raman scattering occurs when photons interact with the molecules of a substance, resulting in an exchange of energy, and the resulting observed energy change (Raman shift) provides detailed information about the molecular vibration and rotational state of the substance. Raman spectroscopy is particularly well suited to analyzing samples containing water, such as wakame seaweed, and is useful for identifying components in aqueous solutions.

[0141] Spectroscopy can also be applied to the measurement of wakame. The following spectroscopic methods are sometimes used to analyze the components and quality of wakame.

[0142] Absorption spectroscopy: The pigments and other compounds in wakame absorb light at specific wavelengths, allowing the concentration of these components to be measured. Ultraviolet and visible light are used to quantify the amount of pigments and antioxidants in wakame.

[0143] Fluorescence spectroscopy: The presence and amount of specific components in wakame can be determined by measuring the fluorescence emitted by those components after absorbing light. For example, the photosynthetic pigment chlorophyll can be analyzed using this method.

[0144] Raman spectroscopy: It can be used to obtain detailed information about the molecular structure of wakame. Raman spectroscopy is effective in studying the chemical bonds and intermolecular interactions within wakame tissue. It can be used to analyze samples such as wakame, which has a high water content, and is useful for investigating its biochemical properties in detail.

[0145] Combining these methods will enable the quality evaluation of wakame and comprehensive analysis of its nutritional components. Spectroscopy is suitable for use in both production sites and laboratories because it is non-destructive and allows for rapid measurements.

[0146] The basic steps for using spectroscopy to measure the composition and quality of wakame are as follows:

[0147] Absorption spectroscopy: (1) Sample preparation: Select the wakame sample to be measured and cut it to the specified size using scissors or a metal cutter. (2) Sample placement: The cut wakame sample is placed flat on the measurement stage attached to the device. (3) Irradiation of the sample: ultraviolet, visible, or infrared light is irradiated onto the sample. (4) Data recording: The wavelength and intensity of the light absorbed by the sample are measured and recorded. (5) Analysis: The concentration of the wakame's components is calculated from the absorbed light data.

[0148] Fluorescence spectroscopy: (1) Sample preparation: Select the wakame sample to be measured and cut it to the specified size using scissors or a metal cutter. (2) Sample placement: The cut wakame sample is placed flat on the measurement stage attached to the device. (3) Illumination of the sample: Short wavelength light (usually ultraviolet) is irradiated onto the sample. (4) Fluorescence detection: Fluorescence emitted from the sample is captured by a detector. (5) Data recording: The intensity and wavelength of the resulting fluorescence are recorded. (6) Analysis: The fluorescence data is analyzed to identify the types and amounts of components in the wakame.

[0149] Raman spectroscopy: (1) Sample preparation: Select the wakame sample to be measured and cut it to the specified size using scissors or a metal cutter. (2) Sample placement: The cut wakame sample is placed flat on the measurement stage attached to the device. (3) Laser irradiation: A laser is irradiated onto the sample. (4) Detection of Raman scattered light: The scattered light (Raman shift) is captured by a detector. (5) Data recording: Raman shift data is acquired and recorded. (6) Analysis: The obtained Raman spectrum is analyzed, and the molecular structure and chemical properties of wakame are investigated in detail.

[0150] These spectroscopic methods serve as effective tools for component analysis, quality control, and nutritional evaluation of wakame. Each method focuses on a specific component or characteristic, and should be selected according to the purpose.

[0151] Next, the quality evaluation method for wakame seaweed using the quality evaluation system according to the present invention will be described, taking the above-mentioned wakame seaweed as an example of seaweed.

[0152] The wakame quality evaluation method for evaluating the quality of wakame includes a database construction step (STEP 1) for constructing a database 28 from the thickness of the training wakame 60 obtained by measuring the training wakame 60, the biochemical indexes obtained by measuring the training wakame 60, the SPAD value obtained by measuring the chlorophyll content of the training wakame 60, the b* value indicating hue and saturation obtained by measuring the training wakame 60, and training image data obtained by photographing the training wakame 60; an evaluation model training step (STEP 2) for extracting features representing color, reflected light amount, and surface shape from the image data used as training data in the database 28 using machine learning, and learning an evaluation model by associating them with the thickness, biochemical indexes, SPAD value, and b* value of the wakame 60; and a wakame quality evaluation step (STEP 3) for evaluating the quality of the wakame 60 to be evaluated by estimating the closest evaluation model from the thickness of the wakame 60 to be evaluated and the photographed evaluation image data.

[0153] In addition, the database construction process (STEP 1) includes a biochemical indicator measurement sub-process (STEP 1-2) in which the biochemical indicators include one or more of the content of fucoxanthin, which is a polysaccharide, and the content of alginic acid, and these biochemical indicators are measured using a measuring device to obtain the biochemical indicators.

[0154] In addition, in the wakame quality evaluation process (STEP 3), a camera-equipped mobile terminal 50, a simple measuring device 40, and an evaluation terminal (analysis unit, server) 30 are used. The simple measuring device 40 has a portable main body 41, multiple slits 42 formed in different widths in the main body 41 and into which wakame 60 is inserted to measure the thickness, a wakame holding section 43 provided in the main body 41 to pull and hold pieces of cut wakame 60, and color charts 21 of different standard colors arranged in multiple areas on the surface of the main body 41. The slits 42 are used to measure the thickness of the wakame 60 to be evaluated for quality, and the camera-equipped mobile terminal 50 takes a photograph of the wakame 60 held in the wakame holding section 43 together with the color chart 21 for evaluation. The evaluation terminal (analysis unit, server) 30 processes the measured thickness of the wakame 60 and the photographed image data for evaluation to evaluate the quality of the wakame 60.

[0155] Next, the operation and effect of the wakame quality evaluation method and the wakame quality evaluation system 10 described above will be explained.

[0156] According to an embodiment of the present invention, a database construction process is provided in which database 28 is constructed from the physical quantities of learning wakame 60 obtained by measuring learning wakame 60, the biochemical indices obtained by measuring learning wakame 60, the SPAD value obtained by measuring the chlorophyll content of learning wakame 60, the b* value indicating hue and saturation obtained by measuring learning wakame 60, and learning image data obtained by photographing learning wakame 60.Therefore, the quality of wakame 60, including its freshness, can be evaluated using the biochemical indices, SPAD value, and b* value by correlating the thickness, biochemical indices, SPAD value, b* value, and image data of wakame 60 obtained by measuring wakame 60, rather than relying on the intuition of an expert craftsman. The system includes an evaluation model learning process for learning an evaluation model from the physical quantities of the wakame 60, the SPAD value indicating the chlorophyll content extracted using machine learning from image data as training data in the database 28, and features expressing the b* value indicating hue and saturation, and a quality evaluation process for the wakame 60 for evaluating the quality of the wakame 60 to be evaluated by estimating the closest evaluation model from the thickness of the wakame 60 to be evaluated and the photographed image data for evaluation.This makes it possible to easily and quickly evaluate the quality of the wakame 60 using artificial intelligence in a visible light environment such as a fishing port or fish market.

[0157] Furthermore, the biochemical indicators include one or more of the content of fucoxanthin, which is a polysaccharide, and the content of alginic acid, and these biochemical indicators are obtained by measuring them using measuring equipment in the biochemical indicator measurement sub-step. Therefore, there is no individual variation that occurs in sensory tests, and the same evaluation results can be obtained regardless of who evaluates the wakame, making it possible to achieve stable and highly accurate quality evaluation.

[0158] Furthermore, the quality evaluation process uses a camera-equipped mobile terminal 50, allowing for inexpensive and easy photography of the wakame seaweed to be evaluated. Furthermore, the quality evaluation process uses a simple measuring device 40, which is easily portable and measures the thickness of the wakame seaweed 60 to be evaluated using a slit 42 in the simple measuring device 40, allowing for easy on-the-spot measurement of the thickness of the wakame seaweed 60. Furthermore, the wakame seaweed 60 to be evaluated is photographed using the color chart 21 containing multiple standard colors in the simple measuring device 40. Therefore, even if image data is taken outdoors or indoors where the amount of light varies, differences in the color temperature of the lighting can be corrected for, allowing for accurate color adjustment and improving the accuracy of the quality evaluation of the wakame seaweed 60.

[0159] Furthermore, the database construction unit 20 constructs the database 28 from the physical quantities of the learning wakame 60 obtained by measuring the learning wakame 60, the biochemical indices obtained by measuring the learning wakame 60, the SPAD value obtained by measuring the chlorophyll content of the learning wakame 60, the b* value indicating hue and saturation obtained by measuring the learning wakame 60, and learning image data obtained by photographing the learning wakame 60.Therefore, the quality of the wakame 60, including its freshness, can be evaluated using the biochemical indices, SPAD value, and b* value by correlating the physical quantities, biochemical indices, SPAD value, b* value, and image data of the wakame 60 obtained by measuring the wakame 60, rather than relying on the senses of an expert craftsman.

[0160] Furthermore, the analysis unit 30 evaluates the physical quantity information and image data of the wakame 60 acquired by the simple measuring device 40 and the camera-equipped mobile terminal 50 via the Internet network 11, so that only simple portable equipment is required at the location of the wakame 60 to be evaluated. Furthermore, the analysis unit 30 uses an evaluation model learned from the thickness of the wakame 60 and feature quantities expressing the SPAD value indicating chlorophyll content and the b* value indicating hue and saturation extracted using machine learning from image data as training data in the database 28, and estimates the evaluation model that is closest to the thickness of the wakame 60 to be evaluated and the photographed image data for evaluation, thereby evaluating the quality of the wakame 60 to be evaluated. Therefore, the quality of the wakame 60 can be easily and quickly evaluated using artificial intelligence in a visible light environment such as a fishing port or a fish market.

[0161] In the examples, the seaweed 60 was from Sanriku, but it is not limited thereto, and it may be from Miyagi Prefecture, Tokushima Prefecture, or any other seaweed. Furthermore, the present invention is not limited to seaweed as an evaluation target, but can be applied to seaweed in general. In other words, the present invention is not limited to the examples as long as the effects and advantages of the present invention are achieved. [Industrial Applicability]

[0162] The present invention is suitable for a seaweed quality evaluation technique that evaluates the quality of seaweed from image data of the seaweed. [Explanation of symbols]

[0163] 10...Wakame quality evaluation system, 11...Network, 20...Database construction unit, 21...Color chart, 22...Camera, 23...Physical quantity measurement unit (laser scanner, etc.), 24...Biochemical indicator measurement unit (biochemical indicators, fucoxanthin, alginic acid), 25...Chlorophyll meter, 26...Color measurement unit, 28...Database, 29...Evaluation model learning unit, 30...Analysis unit (server, evaluation terminal), 40...Simple measurement device, 41...Main body, 42...Slit (groove), 43...Wakame holding unit, 45...Length measurement unit, 46...Strength measurement unit, 50...Mobile terminal with camera (smartphone), 51...Cutting tool (scissors, mold), 60...Wakame (seaweed).

Claims

1. A method for evaluating the quality of seaweed, comprising: A database construction process for constructing a database from the physical quantities of the seaweed obtained by measuring the seaweed for learning, the biochemical indices obtained by measuring the seaweed for learning, the SPAD value obtained by measuring the chlorophyll content of the seaweed for learning, the b* value indicating the hue and saturation obtained by measuring the seaweed for learning, and learning image data obtained by photographing the seaweed for learning; An evaluation model learning process in which feature quantities representing color, reflected light amount, and surface shape are extracted from the image data as training data of the database using machine learning, and an evaluation model is learned by associating the seaweed thickness, the biochemical index, the SPAD value, and the b* value; A seaweed quality evaluation method characterized by comprising a seaweed quality evaluation process for evaluating the quality of the seaweed to be evaluated by estimating the evaluation model that is closest to the physical quantities of the seaweed to be evaluated from the photographed image data for evaluation.

2. The method for evaluating the quality of seaweed according to claim 1, The database construction step includes: The biochemical indicators include one or more of the content of fucoxanthin, a polysaccharide, and the content of alginic acid, and the method for evaluating the quality of seaweed includes a biochemical indicator measurement sub-step in which these biochemical indicators are measured using a measuring instrument.

3. The method for evaluating the quality of seaweed according to claim 1 or claim 2, In the seaweed quality evaluation process, a camera-equipped mobile terminal, a simple measuring device, and an evaluation terminal are used, The simplified measuring device includes a portable main body, a plurality of slits formed in the main body with different widths for inserting the seaweed to measure the thickness, a seaweed holding section provided in the main body for pulling and holding the cut pieces of the seaweed, and a color chart of different standard colors arranged in a plurality of regions on the surface of the main body. The b of the seaweed to be subjected to quality evaluation is measured using the slit, and the seaweed for evaluation is photographed together with the color chart by the mobile terminal with a camera, the seaweed held in the seaweed holding unit, A seaweed quality evaluation method that evaluates the quality of the seaweed by performing arithmetic processing on the physical quantities of the seaweed measured by the evaluation terminal and the photographed image data for evaluation.

4. A seaweed quality evaluation system for evaluating the quality of seaweed, The system includes a database construction unit, a simple measuring device, a mobile terminal with a camera, an internet network, and an analysis unit, The database construction unit includes a camera for photographing the seaweed, a physical quantity measurement unit for measuring the physical quantities of the seaweed, a biochemical index measurement unit for measuring biochemical indexes including at least one of the content of fucoxanthin, which is a polysaccharide of the seaweed, and the content of alginic acid, a chlorophyll meter for measuring the chlorophyll content of the seaweed and indicating the SPAD value, a color measurement unit for measuring the b* value indicating the hue and saturation of the seaweed, and the physical quantities of the seaweed obtained by measuring the seaweed for learning, the biochemical indexes obtained by measuring the seaweed for learning, The system includes a database configured to include a SPAD value obtained by measuring the chlorophyll content of the seaweed for training, a b* value indicating hue and saturation obtained by measuring the seaweed for training, and training image data obtained by photographing the seaweed for training; and an evaluation model learning unit that uses machine learning to extract features representing color, reflected light amount, and surface shape from the image data as training data of the database, and learns an evaluation model by associating the physical quantity, biochemical index, SPAD value, and b* value of the seaweed, The simplified measuring device includes a portable main body, a plurality of slits formed in the main body with different widths for inserting the seaweed to measure the thickness, a seaweed holding section provided in the main body for pulling and holding the cut pieces of the seaweed, and a color chart of different standard colors arranged in a plurality of regions on the surface of the main body. The mobile terminal with a camera takes an image of the seaweed to generate the image data, transmits the image data and data on physical quantities of the seaweed, and displays an evaluation result of the seaweed, The analysis unit evaluates the quality of the seaweed to be evaluated by estimating the evaluation model that is closest to the physical quantity data of the seaweed to be evaluated received via the network and the photographed image data for evaluation, and transmits the evaluation results to the camera-equipped mobile terminal.

Images