Freshness determination program, freshness determination method, freshness determination device, quality determination program, quality determination method, and quality determination device

Abstract

Provided are a freshness determination program, a freshness determination method, and a freshness determination device for non-destructively determining the freshness of a frozen object with accuracy. The present invention causes a computer to execute a process that acquires waveform data of ultrasonic waves with a frequency of 1 MHz or less that have been passed through a frozen object and inputs the waveform data into a machine learning model to determine the freshness of the object.

Description

Freshness determination program, freshness determination method, and freshness determination device

[0001] The present invention relates to a freshness determination program, a freshness determination method, and a freshness determination device.

[0002] One indicator of the quality of fish such as tuna is freshness. Here, freshness refers to the degree of rigor mortis. For example, the freshness of frozen tuna varies depending on when the tuna began freezing and the condition of the freezing equipment.

[0003] Conventionally, methods for detecting the freshness of fish include testing with freshness test strips, analyzing ATP (adenosine triphosphate) using chromatography, and inspecting cross sections during tail-cutting selection. However, these methods are invasive to the frozen tuna, posing a challenge to maintaining the commercial value of the tuna. Furthermore, because frozen tuna is hard, specialized equipment is required for processing, and there is a risk of significant damage during processing. Therefore, invasive inspection methods also pose a challenge due to the difficulty of processing frozen tuna. Furthermore, a technology for evaluating fish quality using ultrasound has been proposed, which uses medical ultrasound equipment with a frequency range of 1 MHz to 100 MHz to perform non-destructive testing of fresh fish.

[0004] Japanese Patent Application Laid-Open No. 2007-155692

[0005] Shigeo Ehira and Hitoshi Uchiyama, "Fish Freshness Determination Method and Measurement Accuracy of the FTP Method," [online], April 26, 2013, Cooking Science, [Retrieved November 16, 2022], Internet <URL: https: / / www.jpo.go.jp / system / laws / rule / guideline / patent / handbook_shinsa / document / index / 01.pdf#page=28>. "Fish Freshness (K Value) Test Method - High-Performance Liquid Chromatography," [online], March 31, 2022, Ministry of Agriculture, Forestry and Fisheries, [Retrieved November 16, 2022], Internet <URL: https: / / public-comment.e-gov.go.jp / servlet / PcmFileDownload?seqNo=0000229410>.

[0006] However, when ultrasonic waves in a frequency band above 1 MHz are propagated through frozen tuna, attenuation of nearly 10 dB occurs per cm. Therefore, when inspecting frozen tuna using ultrasonic waves in a frequency band above 1 MHz, the attenuation coefficient is very large, making it difficult to obtain appropriate waveform data of the reflected waves, making it difficult to perform non-destructive inspection of frozen tuna. As such, with conventional technology, it has been difficult to determine the freshness of frozen objects non-destructively.

[0007] The disclosed technology has been developed in consideration of the above, and aims to provide a freshness determination program, a freshness determination method, and a freshness determination device that can non-destructively and accurately determine the freshness of a frozen object.

[0008] In one aspect, the freshness determination program, freshness determination method, and freshness determination device disclosed in the present application acquire waveform data of ultrasonic waves having a frequency of 1 MHz or less that are propagated through a frozen object, input the waveform data into a machine learning model, and cause a computer to execute a process to determine the freshness of the object.

[0009] According to one aspect of the freshness determination program, freshness determination method, and freshness determination device disclosed in the present application, it is possible to accurately determine the freshness of a frozen object in a non-destructive manner.

[0010] FIG. 1 is a schematic diagram of a freshness determination system according to an embodiment. FIG. 2 is a block diagram of a freshness determination device according to an embodiment. FIG. 3 is a diagram showing an example of the measurement position of a beamline used in freshness determination. FIG. 4 is a diagram showing an example of a measurement position designation screen. FIG. 5 is a diagram showing a comparison between the waveform of normal frozen tuna and the waveform of frozen tuna of poor freshness. FIG. 6 is a diagram showing an example of a power spectrum. FIG. 7 is a diagram showing an example of a measurement in progress screen. FIG. 8 is a diagram showing an example of a comprehensive result display screen. FIG. 9A is a first diagram showing a representative example of waveform data of a reflected wave obtained from normal frozen tuna. FIG. 9B is a second diagram showing a representative example of waveform data of a reflected wave obtained from normal frozen tuna. FIG. 10A is a first diagram showing a representative example of waveform data of a reflected wave obtained from frozen tuna of poor freshness. FIG. 10B is a second diagram showing a representative example of waveform data of a reflected wave obtained from frozen tuna of poor freshness. FIG. 11 is a flowchart of freshness determination processing according to an embodiment. FIG. 12 is a hardware configuration diagram.

[0011] The following describes in detail embodiments of the program, method, and device for determining freshness disclosed herein with reference to the accompanying drawings. Note that the program, method, and device for determining freshness disclosed herein are not limited to the embodiments.

[0012] FIG. 1 is a schematic diagram of a freshness determination system according to an embodiment. The freshness determination system 1 includes a freshness determination device 10, an ultrasonic inspection device 20, and an ultrasonic probe 30. The freshness determination system 1 determines the freshness of a frozen object. The object includes, for example, fish and meat that undergo rigor mortis. In this embodiment, the freshness determination system 1 determines the freshness of frozen tuna F as the frozen object. The freshness determination device 10 is connected to the ultrasonic inspection device 20.

[0013] The ultrasonic probe 30 is a linear array probe in which an ultrasonic wave emitting section and a reflected wave receiving section are arranged in the same location. The ultrasonic probe 30 receives information on ultrasonic wave emission settings, such as frequency, power, and direction, from the ultrasonic inspection device 20. The ultrasonic probe 30 then irradiates ultrasonic waves onto the frozen tuna F in accordance with the specified emission settings, and scans the object using an echo method in which the ultrasonic waves reflected by the frozen tuna F are received again.

[0014] The ultrasonic inspection device 20 receives an instruction to send ultrasonic waves from the freshness determination device 10. The ultrasonic inspection device 20 then transmits information on ultrasonic irradiation settings, such as frequency, power, and direction, to the ultrasonic probe 30, causes the ultrasonic probe 30 to irradiate the frozen tuna F with ultrasonic waves, and the ultrasonic probe 30 acquires a reflected signal of the ultrasonic waves that have propagated through the frozen tuna F. Here, since the freshness determination system 1 according to this embodiment scans frozen fish such as frozen tuna F, the ultrasonic inspection device 20 outputs ultrasonic waves in the low frequency band of 20 kHz to 1 MHz. The ultrasonic inspection device 20 transmits data on the reflected ultrasonic waves obtained by the ultrasonic probe 30 to the freshness determination device 10.

[0015] The freshness determination device 10 sends an instruction to send out ultrasound to the ultrasound inspection device 20, and receives reflected wave data obtained by scanning the frozen tuna F from the ultrasound inspection device 20. The freshness determination device 10 then inputs waveform data obtained from the acquired reflected wave data into a trained machine learning model to determine whether the frozen tuna F is of poor freshness at the measurement position. The freshness determination device 10 then notifies the user P of the final determination result regarding the freshness of the frozen tuna F. The user P is, for example, a buyer of frozen tuna. The freshness determination device 10 will be described in detail below. As shown in FIG. 2 , the freshness determination device 10 has an input / output unit 11, a control unit 12, and a memory unit 13.

[0016] The input / output unit 11 has an output device such as a monitor and an input device such as a keyboard and a mouse. The input / output unit 11 receives instructions from the control unit 12 and displays a specified screen on the output device. The input / output unit 11 also outputs data and commands input by the user P using the input device to the control unit 12.

[0017] The control unit 12 issues an instruction to the ultrasonic inspection device 20 to transmit ultrasonic waves, and the control unit 12 receives, for example, data on reflected waves of ultrasonic waves with a frequency of 500 kHz irradiated onto the frozen tuna F from the ultrasonic inspection device 20, and uses the waveform data of the reflected waves to determine the freshness of the frozen tuna F. As shown in Fig. 2, the control unit 12 has a measurement position designation unit 121, a measurement instruction unit 122, a result notification unit 123, a measurement failure detection unit 124, a data acquisition unit 125, and a freshness failure determination unit 126. While ultrasonic waves with a frequency of 500 kHz have been given as an example, the frequency is not limited to this as long as it is between 20 kHz and 1 MHz, and ultrasonic waves of multiple frequencies may also be used.

[0018] The measurement position designation unit 121 stores in advance information on measurement positions on the frozen tuna F to which ultrasonic waves are applied. FIG. 3 is a diagram showing an example of measurement positions used in freshness determination. FIG. 3 schematically shows three cross sections #1 to #3 of the frozen tuna F with the dorsal fin facing up. Cross section #1 is the tail side of the frozen tuna F, and cross section #3 is the head side of the frozen tuna F. Measurement positions to be brought into contact with the ultrasonic probe 30 are designated for each of the cross sections #1 to #3. For example, for cross section #1, ULv0, ULm0, ULh0, URv0, URm0, URh0, DLv0, DLm0, DLh0, DRv0, DRm0, and DRh0 indicate the measurement positions to be brought into contact with the ultrasonic probe 30. In FIG. 3, the upper side of the measurement positions as viewed from the tail side of the frozen tuna F is designated as U, and the lower side as D. Furthermore, the measurement position is represented by R on the right side when viewed from the tail side of the frozen tuna F and L on the left side. Furthermore, the position of the beamline is represented by v on the vertical direction of the frozen tuna F, m on the middle, and h on the horizontal direction. For example, URh0 corresponds to the measurement position where irradiation is from above and to the right horizontally when viewed from the tail side in cross section #1 of the tail side of the frozen tuna F. Here, the direction from the head to the tail of the frozen tuna F is called the front-to-back direction, and the direction from the dorsal fin to the pectoral fin is called the up-and-down direction.

[0019] The measurement position designation unit 121 stores, for example, ULv0-2, ULm0-2, ULh0-2, URv0-2, URm0-2, URh0-2, DLv0-2, DLm0-2, DLh0-2, DRv0-2, DRm0-2, and DRh0-2 as measurement positions, as shown in Figure 3. It is preferable that a total of four measurement positions be provided on each side of the frozen tuna F, with at least two points aligned in the front-to-back direction and two lines aligned in the up-and-down direction. It is also preferable that the measurement positions include a position near the tail of the frozen tuna F, at least a position closer to the tail than the center in the front-to-back direction.

[0020] The measurement position designation unit 121 also has a selection algorithm that selects all of the measurement positions by selecting them one by one from the measurement positions stored in the unit. For example, the selection algorithm selects measurement positions one by one from the back row of the top row of measurement positions lined up front, back, top, and bottom on one side of the frozen tuna F, and when it reaches the front row, it returns to the back of the next row below, repeating this process until it reaches the front row of the bottom row. When the freshness determination process is started, the measurement position designation unit 121 determines the first measurement position from the measurement positions in accordance with the selection algorithm and sets it as the probe contact position.

[0021] The measurement position designation unit 121 then displays a measurement position designation screen 201 showing the probe contact position as shown in Fig. 4 on the display device of the input / output unit 11. Fig. 4 is a diagram showing an example of the measurement position designation screen. In this way, the measurement position designation unit 121 shows the user P the probe contact position on the frozen tuna F where the ultrasonic probe 30 is to come into contact.

[0022] For example, the measurement position designation unit 121 places an image of frozen tuna F on the measurement position designation screen 201 as shown in Figure 4, and then places each measurement position on the image of the frozen tuna F. Furthermore, the measurement position designation unit 121 adds a predetermined pattern to a probe contact position 211 among the measurement positions placed on the measurement position designation screen 201. The measurement position designation unit 121 also places a display example 212 of a pattern indicating the probe contact position on the measurement position designation screen 201. In addition, the measurement position designation unit 121 places a measurement start button 213 on the measurement position designation screen 201.

[0023] The user P checks the probe contact position by referring to the measurement position designation screen 201, and brings the ultrasound probe 30 into contact with the designated probe contact position. The user P then presses the measurement start button 213 provided on the measurement position designation screen 201 to instruct the start of measurement.

[0024] If there are no measurement defects in the scan by the ultrasound probe 30 after the measurement has been performed, the measurement position designation unit 121 receives an instruction to designate the next probe contact position from the result notification unit 123. Upon receiving the instruction from the result notification unit 123, the measurement position designation unit 121 determines the next measurement position from among the measurement positions in accordance with the selection algorithm and sets it as the probe contact position. Then, the measurement position designation unit 121 displays a measurement position designation screen 201 indicating the next probe contact position on the display device of the input / output unit 11.

[0025] Thereafter, the measurement position designation unit 121 receives a notification from the result notification unit 123 that the result notification has been made. Then, the measurement position designation unit 121 determines whether freshness determination has been completed at all measurement positions. If there are any measurement positions for which freshness determination has not been performed, the measurement position designation unit 121 determines that the inspection is not complete and selects the next measurement position.

[0026] On the other hand, when the freshness determination at all measurement positions is completed, the measurement position designation unit 121 determines that the inspection is completed. When the freshness determination at all measurement positions is completed, the measurement position designation unit 121 notifies the result notification unit 123 of the completion of the inspection.

[0027] When the user P presses the measurement start button 213 provided on the measurement position specification screen 201, the measurement instruction unit 122 receives an instruction to start measurement from the input / output unit 11. Then, the measurement instruction unit 122 instructs the ultrasound inspection device 20 to emit ultrasound waves.

[0028] The data acquisition unit 125 receives from the ultrasonic inspection device 20 data on the reflected waves from the frozen tuna F of ultrasonic waves with a frequency of 500 kHz irradiated from the ultrasonic probe 30. The data acquisition unit 125 then outputs the acquired reflected wave data to the measurement failure detection unit 124.

[0029] 5 is a diagram showing a comparison between the waveform of normal frozen tuna and the waveform of frozen tuna with poor freshness. Graphs 301 and 302 show the waveform of normal frozen tuna with no freshness problems. Graphs 303 and 304 show the waveform of frozen tuna with poor freshness. Graphs 301 to 304 show the elapsed time between the transmission and reception of ultrasonic waves on the horizontal axis, and the amplitude of the reflected wave on the vertical axis. The data acquisition unit 125 acquires data on the reflected waves that form waveforms such as those shown in graphs 301 to 304, and outputs the data to the measurement error detection unit 124.

[0030] The measurement failure detection unit 124 has a trained machine learning model that receives reflected wave data as input and outputs information on measurement failures such as peeling, dirt, or poor contact. Here, "peeling" refers to a state in which the skin of the frozen tuna F has peeled back due to being picked at, for example. "Dirt" refers to a state in which the frozen tuna F is soiled by blood or the like. "Poor contact" refers to a state in which the ultrasonic probe 30 is not in accurate contact with the surface of the frozen tuna F. In either case, the freshness determination device 10 is unable to obtain appropriate waveform data, making it difficult to accurately determine the freshness of the frozen tuna F. This machine learning model may be stored in the memory unit 13, for example. In this case, the measurement failure detection unit 124 uses the trained machine learning model stored in the memory unit 13.

[0031] The measurement failure detection unit 124 receives, from the data acquisition unit 125, data on the reflected waves from the frozen tuna F of the ultrasonic waves having a frequency of 500 MHz irradiated from the ultrasonic probe 30. When the input of the reflected wave data begins, the measurement failure detection unit 124 notifies the result notification unit 123 of the start of measurement.

[0032] Next, the measurement failure detection unit 124 inputs the acquired reflected wave data into a machine learning model it holds to determine whether a measurement failure has occurred. That is, the measurement failure detection unit 124 determines whether there is poor contact of the ultrasonic probe 30 with the object, such as poor contact, curling, or dirt, or whether the object is damaged or dirty. If no measurement failure has occurred, the measurement failure detection unit 124 outputs the reflected wave data to the result notification unit 123 and the freshness failure determination unit 126. On the other hand, if a measurement failure has occurred, the measurement failure detection unit 124 outputs the measurement failure information output from the machine learning model to the result notification unit 123.

[0033] The poor freshness determination unit 126 has a trained machine learning model for freshness determination that inputs waveform data and outputs poor freshness information. The machine learning model for freshness determination can use a support vector machine (SVM) or a neural network. Furthermore, the poor freshness information may be displayed in stages or as a continuous value. The machine learning model for freshness determination possessed by the poor freshness determination unit 126 is trained using, for example, a pair of waveform data of frozen tuna and the results of chemical analysis of the frozen tuna as training data. Alternatively, the machine learning model for freshness determination may be trained using, for example, a pair of waveform data of frozen tuna and the results of a human assessment of the freshness of the frozen tuna as training data. This machine learning model for freshness determination may be stored, for example, in the memory unit 13. In this case, the poor freshness determination unit 126 uses the trained machine learning model stored in the memory unit 13.

[0034] If no measurement failure has occurred, the poor freshness determination unit 126 receives input from the measurement failure detection unit 124 of data on reflected waves from the frozen tuna F of ultrasound with a frequency of 500 kHz irradiated from the ultrasound probe 30, and obtains waveform data of the reflected waves. The poor freshness determination unit 126 then inputs the waveform data into a trained machine learning model for freshness determination, and determines whether the frozen tuna F is poor in freshness at the measurement position at that time. In other words, the poor freshness determination unit 126 inputs waveform data obtained by ultrasound with a frequency between 20 kHz and 1 MHz that is propagated through the frozen tuna F, which is a frozen object, into the machine learning model, and determines the freshness of the frozen tuna F.

[0035] Rigor mortis changes the Young's modulus and viscosity coefficient, which represent the elasticity of the fish body. Both Young's modulus and viscosity coefficient are physical properties that affect sound propagation. When fish are poorly fresh, the amplitude of the reflection from the backbone is larger than in normal cases due to changes in Young's modulus, etc. In other words, in graphs 303 and 304 shown in Figure 5, the area near the maximum amplitude is the reflected wave from the backbone. Furthermore, when fish are poorly fresh, the reflected wave from the area deeper than the backbone is likely to attenuate due to changes in viscosity, etc. In graphs 303 and 304, the area after the area near the maximum amplitude, i.e., the right side of the page, is the reflected wave from the area deeper than the backbone. Compared to graphs 301 and 302, graphs 303 and 304 have a larger difference between the maximum amplitude portion of the waveform and the portion with longer elapsed time of transmission and reception. In other words, the amplitude decreases rapidly from the maximum amplitude portion in the direction of increasing elapsed time of transmission and reception. As such, when fish are poorly fresh, the reflected wave decreases rapidly from the backbone toward the inner part compared to normal cases. Therefore, depending on this condition, the freshness defect determining unit 126 can determine the freshness from the waveform data.

[0036] More specifically, the poor freshness determination unit 126 according to this embodiment determines the freshness of the frozen tuna F using the following method. The poor freshness determination unit 126 calculates the power spectrum by Fourier transforming the waveform data of the reflected wave. FIG. 6 is a diagram showing an example of a power spectrum. Graph 311 in FIG. 6 is waveform data converted into the power spectrum of normal frozen tuna F, and graph 312 is waveform data converted into the power spectrum of poorly fresh frozen tuna F. In graphs 311 and 312, the vertical axis represents intensity and the horizontal axis represents frequency.

[0037] Next, the poor freshness determination unit 126 inputs the waveform data represented by the calculated power spectrum into a machine learning model and obtains poor freshness information output from the machine learning model for determining poor freshness. The power spectrum is frequency-dependent data, and is less affected by the intensity of the irradiated signal and the range in which reflected waves are collected. Therefore, by using the power spectrum, the poor freshness determination unit 126 can use data from different sections in which spectra are collected to train and evaluate the machine learning model for determining poor freshness. Furthermore, by using the power spectrum, the poor freshness determination unit 126 can obtain consistent data from the machine learning model for determining poor freshness, regardless of the start and end points of the waveform. In other words, using the power spectrum makes it easier for the poor freshness determination unit 126 to determine poor freshness.

[0038] The freshness defect determination unit 126 stores the result of the freshness determination of the frozen tuna F at the measurement position at that time in the memory unit 13. At this time, the freshness defect determination unit 126 may notify the result notification unit 123 of the result of the freshness determination of the frozen tuna F at the measurement position at that time.

[0039] The result notification unit 123 receives a notification of the start of measurement from the freshness defect determination unit 126. Then, the result notification unit 123 displays information notifying that the measurement has started on the display device of the input / output unit 11, thereby notifying the user P that the measurement is in progress.

[0040] If a measurement failure occurs, the result notification unit 123 receives input of information about the measurement failure from the measurement failure detection unit 124. The measurement failure information includes, for example, information about the occurrence of curling, dirt, poor contact, etc. Then, the result notification unit 123 displays the information about the measurement failure on the display device of the input / output unit 11, and notifies the user P of the measurement failure.

[0041] Furthermore, if no measurement failure occurs, the result notification unit 123 receives input of reflected wave data from the measurement failure detection unit 124. Then, the result notification unit 123 displays information indicating that waveform data is being acquired on the display device of the input / output unit 11, thereby notifying the user P that waveform data is being acquired.

[0042] 7 is a diagram showing an example of a measurement in progress screen. The result notifying unit 123 displays, for example, a measurement in progress screen 202 shown in FIG. 7 on the display device of the input / output unit 11. If a waveform is being acquired, the result notifying unit 123 highlights a display field 221 on the measurement in progress screen 202. At this time, the result notifying unit 123 may display information 116 of the waveform being acquired on the measurement in progress screen 202.

[0043] Furthermore, if a measurement defect occurs, the result notification unit 123 highlights the display fields 222 to 224 according to the content of the measurement defect. In this way, the result notification unit 123 notifies the user P of the information about the measurement defect. Furthermore, the freshness defect determination unit 126 may provide a display field 225 on the measurement in progress screen 202 that indicates that waveform data is being measured.

[0044] After the freshness determination by the freshness defect determination unit 126, the result notification unit 123 obtains the result of the freshness determination of the frozen tuna F at the measurement position at that time from the memory unit 13. Next, the result notification unit 123 notifies the user P of the result of the freshness determination of the frozen tuna F at the measurement position at that time by displaying it on the display device of the input / output unit 11.

[0045] Furthermore, when freshness determination at all measurement positions is completed, the result notification unit 123 receives notification of inspection completion from the measurement position designation unit 121. Next, the result notification unit 123 obtains the results of the freshness determination of the frozen tuna F at each measurement position from the memory unit 13. Then, the result notification unit 123 notifies the user P of the results of the freshness determination of the frozen tuna F at each measurement position by displaying them on the display device of the input / output unit 11. Furthermore, the result notification unit 123 may notify an evaluation of the overall freshness of the frozen tuna F.

[0046] 8 is a diagram showing an example of a comprehensive result display screen. After completing freshness determination at all measurement positions, the freshness defect determination unit 126 causes the display device of the input / output unit 11 to display, for example, a comprehensive result display screen 203 shown in FIG.

[0047] In this embodiment, the poor freshness determination unit 126 places an image of the frozen tuna F on the overall result display screen 203 as shown in FIG. 8 , and places each measurement position on the image of the frozen tuna F. Furthermore, the measurement position designation unit 121 displays the results of the freshness determination at each measurement position placed on the overall result display screen 203 by adding a specific pattern. On the overall result display screen 203 of FIG. 8 , the poor freshness determination unit 126 adds a pattern indicating whether each part corresponding to each measurement position is of poor freshness or normal. Furthermore, the measurement position designation unit 121 places display examples 232 and 233 of patterns indicating the results of the freshness determination on the overall result display screen 203. Furthermore, as described above, the measurement position designation unit 121 may display an evaluation of the overall freshness of the frozen tuna F on the overall result display screen 203.

[0048] Fig. 9A is a first diagram showing a representative example of waveform data of reflected waves obtained from normal frozen tuna. Fig. 9B is a second diagram showing a representative example of waveform data of reflected waves obtained from normal frozen tuna. Fig. 10A is a first diagram showing a representative example of waveform data of reflected waves obtained from frozen tuna with poor freshness. Fig. 10B is a second diagram showing a representative example of waveform data of reflected waves obtained from frozen tuna with poor freshness. In each of the graphs in Figs. 9A, 9B, 10A and 10B, the horizontal axis represents the elapsed time of transmission and reception, and the vertical axis represents amplitude.

[0049] 10A and 10B, in the case of poor freshness, the amplitude of the reflected wave from a position deeper than the backbone is suddenly smaller than the amplitude of the reflection from the backbone, which is a clear difference from the graphs of Figures 9A and 9B. In other words, because characteristic waveform data exists in the reflected wave for poor freshness, the freshness determination device 10 can distinguish between normal and poor freshness by using a machine learning model for determining poor freshness that has been trained using this waveform data, making it possible to accurately determine freshness.

[0050] 11 is a flowchart of the freshness determination process according to the embodiment. Next, the flow of the freshness determination process by the freshness determination system 1 according to the embodiment will be described with reference to FIG.

[0051] The measurement position designation unit 121 selects one measurement position from the plurality of measurement positions as the probe contact position. The measurement position designation unit 121 displays a measurement position designation screen 201 indicating the probe contact position on the display device of the input / output unit 11, and instructs the user P on the probe contact position on the frozen tuna F where the ultrasonic probe 30 is to be brought into contact (step S1).

[0052] The user P checks the probe contact position by referring to the measurement position designation screen 201, and brings the ultrasonic probe 30 into contact with the designated probe contact position (step S2).

[0053] The user P instructs the start of measurement by pressing the measurement start button 213 provided on the measurement position specification screen 201. When the user P presses the measurement start button 213 provided on the measurement position specification screen 201, the measurement instruction unit 122 receives an instruction to start measurement from the input / output unit 11. Then, the measurement instruction unit 122 instructs the ultrasound inspection device 20 to emit ultrasound waves (step S3).

[0054] The ultrasonic inspection device 20 receives an instruction to transmit ultrasonic waves from the freshness determination device 10. Then, the ultrasonic inspection device 20 transmits information on ultrasonic wave irradiation settings, such as frequency, power, and direction, to the ultrasonic probe 30, and the ultrasonic probe 30 irradiates the frozen tuna F with ultrasonic waves (step S4). The ultrasonic inspection device 20 then acquires, with the ultrasonic probe 30, a reflected signal of the ultrasonic waves that have propagated through the frozen tuna F.

[0055] The measurement failure detection unit 124 receives, from the data acquisition unit 125, data on the reflected waves of ultrasound from the frozen tuna F irradiated from the ultrasound probe 30. The measurement failure detection unit 124 then inputs the acquired reflected wave data into a machine learning model it holds, and determines whether or not a measurement failure has been detected (step S5).

[0056] If a measurement failure is detected (step S5: Yes), the measurement failure detection unit 124 outputs information about the measurement failure to the result notification unit 123. The result notification unit 123 displays the measurement failure information input from the measurement failure detection unit 124 on the measurement in progress screen 202 displayed on the display device of the input / output unit 11, thereby notifying the user P of the measurement failure. Upon receiving the measurement failure information, the user P corrects the position of the ultrasound probe 30 so that an appropriate measurement can be performed (step S6). Thereafter, the freshness determination process returns to step S3.

[0057] On the other hand, if no measurement defect is detected (step S5: No), the measurement defect detection unit 124 outputs the reflected wave data to the freshness defect determination unit 126. The freshness defect determination unit 126 receives input of data on the reflected waves from the frozen tuna F of the ultrasonic waves irradiated from the ultrasonic probe 30 from the measurement defect detection unit 124, and obtains waveform data of the reflected waves. The freshness defect determination unit 126 then uses the waveform data in a machine learning model for freshness determination to determine whether the frozen tuna F is poor in freshness at the measurement position at that time (step S7).

[0058] After the freshness determination by the poor freshness determination unit 126, the result notification unit 123 obtains the result of the freshness determination of the frozen tuna F at the measurement position at that time from the memory unit 13. Next, the result notification unit 123 notifies the user P of the result of the freshness determination of the frozen tuna F at the measurement position at that time by displaying it on the display device of the input / output unit 11 (step S8).

[0059] After the result notification unit 123 notifies the measurement position at that time, the measurement position designation unit 121 determines whether the inspection is complete (step S9). If there are measurement positions for which freshness determination has not been performed and the inspection is not complete (step S9: No), the freshness determination process returns to step S1.

[0060] On the other hand, if freshness determination has been completed at all measurement positions and the inspection has ended (Step S9: Yes), the result notification unit 123 obtains the results of the freshness determination of the frozen tuna F at each measurement position from the memory unit 13. Then, the result notification unit 123 displays an overall result including the results of the freshness determination of the frozen tuna F at each measurement position on the display device of the input / output unit 11 (Step S10). The user P can check the overall result and understand the freshness of the frozen tuna F.

[0061] As described above, the freshness determination device according to this embodiment receives waveform data of reflected waves of ultrasound irradiated onto a frozen object and determines the freshness of the frozen object using a machine learning model. This allows the freshness determination device according to this embodiment to determine the freshness of the frozen object non-destructively.

[0062] Furthermore, if a measurement error occurs due to curling, dirt, or poor contact of the ultrasonic probe, the freshness determination device according to this embodiment notifies the user of the details of the measurement error, prompts them to correct the contact position of the ultrasonic probe, and then performs scanning again to repeatedly acquire ultrasonic waveforms and determine freshness. This allows appropriate waveform data to be acquired, making it possible to accurately determine the freshness of frozen objects.

[0063] Next, a second embodiment will be described. The freshness determination device 10 according to this embodiment differs from the first embodiment in that it determines freshness using data included in a predetermined section of the waveform data of the reflected ultrasonic wave. The freshness determination device 10 according to this embodiment is also shown in Figure 2. In the following description, explanations of the operations of the same parts as those in the first embodiment will be omitted.

[0064] Because the freshness of frozen tuna F is not a localized change but a change throughout the entire flesh of the frozen tuna F, it is preferable to determine freshness using reflected waves from a certain section. On the other hand, the waveform of ultrasonic waves attenuates as they move away from the surface of the frozen tuna F. Furthermore, ultrasonic waveforms that have been there for a long time reflect the effects of multiple reflections, and there is a possibility that information about the flesh is missing. Also, while reflected waves contain information about the flesh along the path they have passed, and freshness can be determined from this information, reflected waves from earlier times contain little information about the entire flesh. For this reason, it is preferable to use reflected waves from a certain section away from the surface to determine freshness.

[0065] Furthermore, the propagation distance of the ultrasonic waves can be calculated from the sound speed of the waveform data, and it can be seen that a large reflected wave is generated at the position of the backbone. Therefore, as can be seen in Figures 9 and 10, compared to normal frozen tuna F, the waveform data of the reflected wave of frozen tuna F with poor freshness is characterized by a larger amplitude at the position of the backbone than at other positions, and a rapid decrease in amplitude from the position of the backbone. Therefore, it is preferable to use the waveform data of the reflected wave in the section including the backbone to determine freshness. Therefore, the freshness determination device 10 according to this embodiment performs freshness determination as follows.

[0066] The freshness defect determination unit 126 identifies the position of the reflected wave from the backbone in the acquired waveform data of the reflected wave. For example, the freshness defect determination unit 126 may determine the position where the amplitude of the reflected wave is maximum as the reflected wave from the backbone, or may determine the center of a certain section where the average value of the amplitude within the section is maximum as the position of the backbone.

[0067] Next, the freshness defect determination unit 126 acquires data within a predetermined distance from the backbone in the direction of ultrasonic irradiation in the waveform data of the reflected wave. For example, the freshness defect determination unit 126 acquires data for a predetermined section before and after the position of the backbone in the waveform data of the reflected wave in the direction of ultrasonic irradiation. Specifically, the freshness defect determination unit 126 determines the position of the backbone from the peak of the amplitude and acquires data for a section of 40 msec before and after the peak of the amplitude in terms of the elapsed time of transmission and reception. The freshness defect determination unit 126 then uses the data for the predetermined section from the acquired position of the backbone to determine the freshness of the frozen tuna F.

[0068] In this embodiment, the section is set to 40 msec before and after the position of the backbone, but it is preferable that the section of the ultrasonic irradiation direction be determined according to the object. Furthermore, the section of the ultrasonic irradiation direction does not have to be an equal distance from the position of the backbone. For example, the freshness defect determination unit 126 may calculate the section of the ultrasonic irradiation direction so that the length from the back of the backbone, i.e., in the direction of ultrasonic propagation as viewed from the backbone, is shorter than the length in front of the backbone, i.e., in the direction opposite to the direction of ultrasonic propagation as viewed from the backbone.

[0069] Furthermore, in this embodiment, the freshness determination device 10 acquired waveform data for the entire range of the direction of ultrasonic irradiation of the frozen tuna F, and performed freshness determination using data from the acquired waveform data within a range of a predetermined distance around the backbone in the direction of ultrasonic irradiation. However, this is not limiting, and for example, the freshness determination device 10 may control the ultrasound inspection device 20 to scan so as to acquire reflected waves within a range of a predetermined distance around the backbone in the direction of ultrasonic irradiation. In this case, the freshness determination device 10 may use a pre-specified position as the position of the backbone, or it may acquire waveform data once, identify the position of the backbone from the waveform data, and specify the scanning range for the ultrasound inspection device 20 using the identified position of the backbone.

[0070] As described above, the freshness determination device according to this embodiment determines the freshness of a frozen object using a predetermined range of data from the waveform data of the reflected wave, including the backbone in the direction of ultrasonic irradiation. This allows the freshness determination process to be performed while excluding waveform portions of the reflected wave that are not important for freshness determination, thereby reducing the processing load. Furthermore, by excluding waveform portions of the reflected wave that are not important for freshness determination, the influence of such reflected waves on freshness determination can be reduced, enabling more accurate freshness determination.

[0071] (Hardware) Next, an example of the hardware configuration of the freshness determination device 10 will be described. Fig. 12 is a diagram showing an example of the hardware configuration of the freshness determination device. As shown in Fig. 12, the freshness determination device 10 has a processor 91, a memory 92, a hard disk 93, and a communication device 94. The processor 91 is also connected to the memory 92, the hard disk 93, and the communication device 94 via a bus.

[0072] The communication device 94 is a communication interface with an external device. For example, the communication device 94 relays communication between the processor 91 and the ultrasonic inspection device 20. The communication device 94 also realizes the function of the data acquisition unit 125.

[0073] The hard disk 93 is an auxiliary storage device. The hard disk 93 realizes the functions of the storage unit 13 illustrated in Fig. 1. The hard disk 93 also stores various programs including a program for realizing the functions of the control unit 10 illustrated in Fig. 1.

[0074] The processor 91 reads out various programs stored on the hard disk 93, expands them into the memory 92, and executes them. As a result, the processor 91 realizes the functions of the control unit 12, which includes a measurement position designation unit 121, a measurement instruction unit 122, a result notification unit 123, a measurement failure detection unit 124, and a freshness failure determination unit 126.

[0075] Furthermore, the freshness determination device 10 can also realize the same functions as those of the above-described embodiment by reading the program from a recording medium using a media reading device and executing the read program. Note that the program referred to here is not limited to being executed by a specific freshness determination device 10. For example, the present invention can also be applied in the same way when another computer or server executes the program, or when these execute the program in cooperation with each other.

[0076] This program can be distributed via a network such as the Internet. In addition, this program can be recorded on a computer-readable recording medium such as a hard disk, a flexible disk (FD), a CD-ROM, a magneto-optical disk (MO), or a digital versatile disk (DVD), and can be read out from the recording medium and executed by a computer.

[0077] REFERENCE SIGNS LIST 1 Freshness determination system 10 Freshness determination device 20 Ultrasonic inspection device 30 Ultrasonic probe 11 Input / output unit 12 Control unit 13 Memory unit 121 Measurement position designation unit 122 Measurement instruction unit 123 Result notification unit 124 Measurement failure detection unit 125 Data acquisition unit 126 Freshness failure determination unit

Claims

1. Acquiring waveform data of ultrasonic waves having a frequency of 1 MHz or less propagated through the frozen object; inputting the waveform data into a machine learning model to determine the freshness of the object; A freshness determination program that causes a computer to execute a process.

2. The process of acquiring the waveform data includes a process of acquiring the waveform data including a wave of ultrasound reflected by a central bone in the object.

2. The freshness determination program according to claim 1, wherein the freshness determination program is a program for determining freshness of a food product.

3. The process of acquiring the waveform data includes a process of acquiring waveform data of reflected waves from a range of a predetermined distance across the backbone in the direction of ultrasonic irradiation.

3. The freshness determination program according to claim 2.

4. Based on the waveform data, poor contact of the ultrasonic probe with the object, damage to the object, and dirt on the object are detected.

2. The freshness determination program according to claim 1, further comprising causing the computer to execute a process.

5. The process of inputting the waveform data into the machine learning model includes a process of calculating a power spectrum of the waveform data and inputting the calculated power spectrum into the machine learning model.

2. The freshness determination program according to claim 1,

6. Acquiring waveform data of ultrasonic waves having a frequency of 1 MHz or less propagated through the frozen object; The waveform data is input into a machine learning model to determine the freshness of the object. A freshness determination method characterized in that the processing is executed by a computer.

7. A control unit that acquires waveform data of ultrasonic waves having a frequency of 1 MHz or less that are propagated through a frozen object, inputs the waveform data into a machine learning model, and determines the freshness of the object. A freshness determination device comprising:

8. Acquiring waveform data of ultrasonic waves having a frequency of 1 MHz or less propagated through the frozen object; inputting the waveform data into a machine learning model to determine the quality of the object; A quality assessment program that causes a computer to execute a process.

9. Acquiring waveform data of ultrasonic waves having a frequency of 1 MHz or less propagated through the frozen object; Inputting the waveform data into a machine learning model to determine the quality of the object. A quality assessment method characterized in that the processing is executed by a computer.

10. a control unit that acquires waveform data of ultrasonic waves having a frequency of 1 MHz or less that are propagated through the frozen object, inputs the waveform data into a machine learning model, and determines the quality of the object; A quality determination device comprising: