Crop cultivation support methods, crop cultivation support programs, and crop cultivation support devices

Abstract

It outputs information for appropriately adjusting the amount of useful substances in the seeds or fruits of crops. [Solution] The present invention provides a crop cultivation assistance method in which a computer performs the following processes: acquiring first spectral data of the leaf portion of a crop and second spectral data of the fruit portion; inputting the acquired first spectral data into a first model that shows the relationship between the first spectral data and the amount of a specific compound contained in the leaves to estimate the amount of the specific compound corresponding to the acquired first spectral data; inputting the acquired second spectral data into a second model that shows the relationship between the second spectral data and the amount of a specific useful substance contained in the fruit to estimate the amount of the specific useful substance corresponding to the acquired second spectral data; and determining and outputting information on the stress to be applied to the crop based on the estimated amount of the specific compound and the estimated amount of the specific useful substance.

Description

[Technical Field]

[0001] The present invention relates to a crop cultivation assistance method, a crop cultivation assistance program, and a crop cultivation assistance apparatus. [Background Art]

[0002] As one of the radical treatment methods for Japanese cedar pollinosis, "oral immune tolerance", which can induce immune tolerance even by oral administration of antigenic proteins whose antigenic determinants (T cell epitopes) contained in antigen proteins or whose three-dimensional structure has been destroyed, has attracted attention as a safe immune tolerance induction method with extremely few side effects.

[0003] As one immune tolerance induction method, a genetically modified rice "Japanese cedar pollen rice" in which antigenic determinants are highly accumulated in rice grains is known. In addition to Japanese cedar pollinosis, vaccine rice in which antigen proteins for disease prevention are accumulated in rice is also known. When cultivating such rice, it is considered preferable that the higher the grain protein content, the greater the production amount of antigen protein.

[0004] Conventionally, techniques for evaluating the growth state of plants using hyperspectral data of plants are known (see, for example, Patent Document 1). Also, in order to increase the enriched components (lutein, sugar concentration) in the edible part of spinach, a method of applying water stress while promoting transpiration is known (see, for example, Patent Document 2). Furthermore, for improving the quality of fruits, techniques for sensing water stress indicators and analyzing the degree of water stress are known (see, for example, Patent Document 3). [Prior Art Documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Unexamined Patent Application Publication No. 2021-001777 [Patent Document 2] Japanese Unexamined Patent Application Publication No. 2020-14395 [Patent Document 3] Japanese Unexamined Patent Application Publication No. 2020-137415 [Summary of the Invention] [Problems that the invention aims to solve]

[0006] However, even with the conventional techniques described above, it may not be possible to adequately control the protein content of rice grains. Furthermore, not only the protein content of rice grains, but also the amount of useful substances contained in the grains or fruits of other crops is difficult to adjust appropriately using only conventional techniques.

[0007] Therefore, the present invention aims to provide a crop cultivation assistance method, a crop cultivation assistance program, and a crop cultivation assistance device that can output information for appropriately adjusting the amount of useful substances in the seeds or fruits of crops. [Means for solving the problem]

[0008] The present invention provides a crop cultivation assistance method in which a computer performs the following processes: acquiring first spectral data of the leaf portion of a crop and second spectral data of the fruit portion; inputting the acquired first spectral data into a first model that shows the relationship between the first spectral data and the amount of a specific compound contained in the leaves to estimate the amount of the specific compound corresponding to the acquired first spectral data; inputting the acquired second spectral data into a second model that shows the relationship between the second spectral data and the amount of a specific useful substance contained in the fruit to estimate the amount of the specific useful substance corresponding to the acquired second spectral data; and determining and outputting information on the stress to be applied to the crop based on the estimated amount of the specific compound and the estimated amount of the specific useful substance. [Effects of the Invention]

[0009] The crop cultivation assistance method, crop cultivation assistance program, and crop cultivation assistance device of the present invention have the effect of being able to output information for appropriately adjusting the amount of useful substances in the seeds or fruits of a crop. [Brief explanation of the drawing]

[0010] [Figure 1] Figure 1 is a schematic diagram showing the configuration of the cultivation support system according to the first embodiment. [Figure 2] Figure 2 shows the hardware configuration of the control device shown in Figure 1. [Figure 3] Figure 3 is a functional block diagram of the control device shown in Figure 1. [Figure 4] Figure 4 schematically illustrates the method by which the learning model generation unit generates a machine learning model used to estimate the amount of compounds in the leaves, and the method by which the leaf compound amount estimation unit estimates the amount of compounds in the leaves. [Figure 5] Figure 5 schematically illustrates the method by which the learning model generation unit generates a machine learning model used to estimate the seed protein content, and the method by which the seed protein content estimation unit estimates the seed protein content. [Figure 6] Figure 6 is a flowchart showing the processing of the control device in the first embodiment. [Figure 7] Figures 7(a) to 7(c) are diagrams illustrating the method for adjusting the nutrient solution EC. [Figure 8] Figure 8 is a schematic diagram showing the configuration of the cultivation support system according to Modification Example 1. [Figure 9] Figures 9(a) to 9(c) are diagrams illustrating the method for adjusting UV intensity according to Modification Example 1. [Figure 10] Figure 10 is a schematic diagram showing the configuration of the cultivation support system according to the modified example 2. [Figure 11] This is a flowchart showing the processing of the control device according to the second embodiment. [Figure 12] Figure 12(a) is a graph showing the grain protein content at harvest when EC control is not performed (comparative example) and when EC control is performed (second embodiment), and Figure 12(b) is a graph showing the grain protein content at harvest when UV irradiation control is not performed (comparative example) and when UV irradiation control is performed. Mode for Carrying Out the Invention

[0011] First Embodiment Hereinafter, a first embodiment of a cultivation support system will be described in detail based on FIGS. 1 to 7(c). FIG. 1 schematically shows the configuration of a cultivation support system 100 according to the first embodiment.

[0012] The cultivation support system 100 of the first embodiment is a system used in a closed plant factory for hydroponically cultivating crops (herein, rice) into which an antigen protein that is a causative substance of cedar pollinosis or a vaccine antigen protein has been introduced by genetic recombination. The grain protein content of general rice (paddy rice cultivation) is about 6 to 8%, and it is known that the content can be increased by applying stress to rice through measures such as increasing nitrogen fertilizer application (see, for example, "Inatsu (1995) Effect of Silicic Acid on Dry Matter Production and Protein Content of Polished Rice, Hokkaido Central Agricultural Experiment Station", "Maruyama S et al (2010) Effect of Nutrient Solution Concentration at the Heading Time on the Growth, Development, and Seed Storage Protein Content of Rice Plants in a Controlled Environment. Environ Control Biol 48(1), 17-24. doi: 10.2525 / ecb.48.17"). To produce a large amount of antigen protein, it is desirable to increase the grain protein content.

[0013] As shown in FIG. 1, the cultivation support system 100 of the first embodiment includes a first spectrometer, a second spectrometer 22, an irrigation device 30, and a control device 10 serving as a crop cultivation auxiliary device.

[0014] The first spectroscope 21 is a device that acquires spectral data (first spectral data) of rice leaf portions. The first spectroscope 21 is fixed at a height position corresponding to the leaf portion of rice (a position facing the leaf).

[0015] The second spectroscope 22 is a device that acquires spectral data (second spectral data) of rice grain portions. The second spectroscope 22 is fixed at a position where grains are highly likely to exist on rice (a position facing the grains). For details of the second spectroscope 22, it is disclosed in, for example, "Yasuyuki Hidaka, Eiji Kurihara, Kazunobu Hayashi, Takahiro Noda, Hiroshi Nishimura, Takao Sugiyama, Kengo Muramatsu, Kunio Sashida, 'Development of a Reflection-type Near-infrared Spectroscopic Protein Meter Mounted on a Self-threshing Combine', Journal of the Japanese Society of Agricultural Machinery, 72(6), 570-577, 2010".

[0016] The irrigation device 30 is a device that supplies nutrient solution to rice in hydroponic cultivation. Under instructions from the control device 10, the irrigation device 30 adjusts the EC (Electrical Conductivity) (i.e., nitrogen fertilization amount) of the nutrient solution supplied to rice, and also adjusts the timing and amount of irrigation.

[0017] The control device 10 determines whether to increase or decrease the EC of the nutrient solution (nutrient solution EC) supplied to rice based on the first and second spectral data acquired from the first spectroscope 21 and the second spectroscope 22, and outputs the determined information to the irrigation device 30, thereby controlling irrigation by the irrigation device 30.

[0018] Figure 2 shows the hardware configuration of the control device 10. As shown in Figure 2, the control device 10 includes a CPU 90, ROM 92, RAM 94, storage (in this case, an SSD or HDD) 96, a communication interface 97, and a portable storage medium drive 99, etc. Each of these components of the control device 10 is connected to a bus 98. In the control device 10, the CPU 90 executes programs (including crop cultivation support programs) stored in the ROM 92 or storage 96, or programs read from the portable storage medium 91 by the portable storage medium drive 99, thereby realizing the functions of each component shown in Figure 3. Note that the functions of each component in Figure 3 may be realized by integrated circuits such as ASICs (Application Specific Integrated Circuits) or FPGAs (Field Programmable Gate Arrays).

[0019] Figure 3 shows a functional block diagram of the control device 10. In the control device 10, the CPU 90 executes a program, and as shown in Figure 3, it functions as a first spectral data acquisition unit 61, a second spectral data acquisition unit 62, a leaf compound amount estimation unit 63 as a first estimation unit, a seed protein content estimation unit 64 as a second estimation unit, a stress control unit 65 as a processing unit, and a learning model generation unit 70. The first spectral data acquisition unit 61 and the second spectral data acquisition unit 62 realize the function of acquisition units that acquire the first and second spectral data.

[0020] The first spectral data acquisition unit 61 acquires spectral data of the leaf portion (first spectral data) from the first spectrometer 21 (see actual data in Figure 4) and inputs it to the leaf compound amount estimation unit 63. In the actual data in Figure 4, the horizontal axis is Wavelength [nm] and the vertical axis is Reflectance.

[0021] The second spectral data acquisition unit 62 acquires spectral data of the grain portion (second spectral data) from the second spectrometer 22 (see actual data in Figure 5) and inputs it to the grain protein content estimation unit 64. Note that, as with Figure 4, the horizontal axis of the actual data in Figure 5 is Wavelength [nm] and the vertical axis is Reflectance.

[0022] The leaf compound amount estimation unit 63 uses a machine learning model generated by the learning model generation unit 70 to estimate the amount of leaf compounds (e.g., anthocyanins and chlorophyll, which are leaf pigments) to estimate the amount of compounds in the leaf portion (leaf compound amount) from the first spectral data (real data). The leaf compound amount estimation unit 63 inputs the estimated leaf compound amount to the stress control unit 65.

[0023] The grain protein content estimation unit 64 estimates the grain protein content from the second spectral data (real data) using a machine learning model for estimating the grain protein content generated by the learning model generation unit 70. The grain protein content estimation unit 64 inputs the estimated grain protein content to the stress control unit 65.

[0024] The stress control unit 65 determines whether to increase or decrease the stress applied to the rice plant based on the estimated amount of leaf compounds input from the leaf compound amount estimation unit 63 and the estimated amount of grain protein content input from the grain protein content estimation unit 64. It then outputs the determined information to the irrigation device 30 and controls the irrigation device 30. It is known that when stress is applied to rice plants by nutrient solution EC, the effects of the stress appear relatively quickly in the amount of leaf compounds, and later in the grain protein content, which is the target product. Therefore, the stress control unit 65 in this first embodiment controls the irrigation device 30 and adjusts the nutrient solution EC based on these two indicators (amount of leaf compounds and grain protein content).

[0025] The learning model generation unit 70 generates a machine learning model (first model) used by the leaf compound amount estimation unit 63 when estimating the amount of leaf compounds, and a machine learning model (second model) used by the seed protein content estimation unit 64 when estimating the seed protein content, using machine learning techniques.

[0026] Figure 4 schematically shows the method by which the learning model generation unit 70 generates a machine learning model used to estimate the amount of compounds in the leaves, and the method by which the leaf compound amount estimation unit 63 estimates the amount of compounds in the leaves.

[0027] As shown in Figure 4, the learning model generation unit 70 generates a machine learning model for estimating the amount of compounds in leaves using a large amount of training data. Here, the training data is data that associates (annotates) the measured data of the first spectral data with the measured values ​​of the amount of compounds in leaves obtained by destructive testing, etc. By machine learning the large amount of training data, the learning model generation unit 70 generates a machine learning model for the leaf compound amount estimation unit 63 to estimate the amount of compounds in leaves (output data) from the first spectral data (actual data).

[0028] Figure 5 schematically illustrates the method by which the learning model generation unit 70 generates a machine learning model used to estimate the grain protein content, and the method by which the grain protein content estimation unit 64 estimates the grain protein content. As shown in Figure 5, the learning model generation unit 70 generates a machine learning model for estimating the grain protein content using a large amount of training data. Here, the training data is data that associates (annotates) the measured data of the second spectral data with the measured values ​​of the grain protein content obtained by destructive testing, etc. By machine learning the large amount of training data, the learning model generation unit 70 generates a machine learning model for the grain protein content estimation unit 64 to estimate the grain protein content (output data) from the second spectral data (actual data). Furthermore, regarding the method for estimating the grain protein content from spectral data, the techniques disclosed in "Ma J, Zheng B and He Y (2022) 'Applications of a Hyperspectral Imaging System Used to Estimate Wheat Grain Protein: A Review.' Front. Plant Sci. 13:837200. doi: 10.3389 / fpls.2022.837200" and "Yasuyuki Hidaka, Eiji Kurihara, Kazunobu Hayashi, Takahiro Noda, Hiroshi Nishimura, Takao Sugiyama, Kengo Muramatsu, Kunio Sashida, 'Development of a Self-Propelled Combine-Mounted Reflective Near-Infrared Spectroscopy Protein Meter', Journal of the Japanese Society of Agricultural Machinery 72(6), 570-577, 2010" may be adopted.

[0029] (Regarding the processing of the control device 10) Next, the processing of the control device 10 will be explained in detail according to the flowchart in Figure 6. As a prerequisite for the start of the processing in Figure 6, the rice plants have gone through the tillering stage after planting and reached the heading stage, and during a predetermined period d1 after heading, the nutrient solution EC is gradually increased as shown in Figure 7(a), and the nutrient solution EC reaches a predetermined value e [mS / cm].

[0030] When the process shown in Figure 6 begins, in step S10, the first spectral data acquisition unit 61 acquires first spectral data from the first spectrometer 21. The first spectral data acquisition unit 61 inputs the acquired first spectral data to the leaf compound amount estimation unit 63.

[0031] Next, in step S12, the leaf compound amount estimation unit 63 estimates the amount of leaf compounds based on the first spectral data input from the first spectral data acquisition unit 61. At this time, as shown in Figure 4, the leaf compound amount estimation unit 63 inputs the first spectral data into the machine learning model generated by the learning model generation unit 70 to obtain an output (estimated value) of the amount of leaf compounds. The leaf compound amount estimation unit 63 inputs the estimated value of the amount of leaf compounds to the stress control unit 65.

[0032] Next, in step S14, the stress control unit 65 determines whether the amount of leaf compounds (estimated value) is less than or equal to α. If the amount of a particular leaf compound is high, it means that the rice is in a state of high stress. α is a predetermined threshold (first threshold) used to determine whether or not the rice is in a state of high stress. If the determination in step S14 is affirmed, that is, if the rice is not in a state of high stress, the process proceeds to step S16.

[0033] When the process moves to step S16, the second spectral data acquisition unit 62 acquires second spectral data from the second spectrometer 22. The second spectral data acquisition unit 62 inputs the acquired second spectral data to the seed protein content estimation unit 64.

[0034] Next, in step S18, the grain protein content estimation unit 64 estimates the grain protein content based on the second spectral data input from the second spectral data acquisition unit 62. At this time, as shown in Figure 5, the grain protein content estimation unit 64 inputs the second spectral data into the machine learning model generated by the learning model generation unit 70 to obtain an output (estimated value) of the grain protein content.

[0035] Next, in step S20, the stress control unit 65 determines whether the grain protein content is above a predetermined threshold β (second threshold). If the determination in step S20 is negative, the stress control unit 65 proceeds to step S21. In step S21, the stress control unit 65 determines whether it is within a predetermined period from the start of processing shown in Figure 6. Here, the predetermined period refers to the period (initial period) during which, due to the growth characteristics of rice, the grain protein content does not reach the target even if the stress intensity is set high. The predetermined period is assumed to be set in advance. If the determination in step S21 is positive, the process proceeds to step S24, and the stress control unit 65 waits for one day. During this waiting period, the stress intensity is maintained.

[0036] On the other hand, if the judgment in step S21 is rejected, it means that the rice grain protein content is low even after the predetermined period has passed. In this case, the stress control unit 65 proceeds to step S22 and instructs the irrigation device 30 to increase the stress intensity. As a result, the irrigation device 30 increases the nutrient solution EC by Δe1, for example, as shown in Figure 7(b) for the number of days after heading d2. By increasing the nutrient solution EC, an increase in grain protein content can be expected. Note that Δe1 may be a predetermined value, or it may be a value corresponding to the estimated grain protein content (for example, a value corresponding to the difference between the threshold β and the estimated grain protein content).

[0037] After step S22, in step S24, the stress control unit 65 remains idle for one day.

[0038] After waiting for one day in step S24, the process returns to step S10. In this first embodiment, the process in Figure 6 is repeated on a daily basis, so there is a one-day wait in step S24 and step S28, which will be described later. However, the waiting time is arbitrary and may be longer or shorter than one day.

[0039] On the other hand, if the judgment in step S14 is rejected, that is, if the rice is in a state of high stress, the process proceeds to step S26, and the stress control unit 65 instructs the irrigation device 30 to reduce the stress intensity. As a result, the irrigation device 30 reduces the nutrient solution EC by Δe2, for example, as shown in Figure 7(c) for the number of days after heading d2. By reducing the nutrient solution EC, the stress on the rice can be reduced. Note that Δe2 may be a predetermined value, or it may be a value corresponding to the estimated amount of intraleaf compounds (for example, a value corresponding to the difference between the threshold α and the estimated amount of intraleaf compounds).

[0040] After step S26, the stress control unit 65 waits for one day in step S28, and after one day has elapsed, the process returns to step S10.

[0041] Furthermore, if step S14 is affirmed and the judgment in step S20 is also affirmed, that is, if the amount of leaf compounds is below the threshold α and the grain protein content is above the threshold β, the process proceeds to step S26. In this case, the stress control unit 65 instructs the irrigation device 30 to reduce the stress intensity so that the grain protein content does not increase any further. As a result, the irrigation device 30 reduces the nutrient solution EC by Δe2, for example, as shown in Figure 7(c) at d2 days after heading.

[0042] The above process is then repeated. For example, once the maturity stage is passed (before harvest time), the process shown in Figure 6 is forcibly terminated.

[0043] As mentioned above, the effects of stress applied to rice plants appear relatively quickly in the amount of leaf compounds, and more slowly in the grain protein content, which is the target product. Therefore, by performing the treatment shown in Figure 6, as in this first embodiment, it is possible to identify the stress intensity from the amount of leaf compounds in the initial stage and control the stress (adjust the nutrient solution EC), and thereafter control the stress (adjust the nutrient solution EC) based on the grain protein content. This allows for appropriate control of stress (nutrient solution EC) and enables rice plants to produce more grain protein.

[0044] As described in detail above, according to this first embodiment, the first and second spectral data acquisition units 61 and 62 acquire first spectral data of the leaf portion of the rice plant and second spectral data of the grain portion (S10, S16). The leaf compound amount estimation unit 63 inputs the acquired first spectral data to a machine learning model that shows the relationship between the first spectral data and the leaf compound amount to estimate the leaf compound amount (S12), and the grain protein content estimation unit 64 inputs the acquired second spectral data to a machine learning model that shows the relationship between the second spectral data and the grain protein content to estimate the grain protein content (S18). Then, the stress control unit 65 determines the stress to be applied to the rice plant based on the estimated leaf compound amount and grain protein content and outputs it to the irrigation device 30 (S22, S26). In this first embodiment, by monitoring both the amount of intraleaf compounds and the grain protein content, which have different response rates to stress, and adjusting the nutrient solution EC via the irrigation device 30 based on these, the amount of the target substance, grain protein, can be appropriately adjusted. When cultivating rice into which vaccine antigen protein has been introduced by genetic modification, as in this first embodiment, it is necessary to increase the production of vaccine antigen protein, and as described above, it is possible to produce more vaccine antigen protein by increasing the grain protein content of the rice.

[0045] Furthermore, in this first embodiment, the stress control unit 65 determines information on the stress to be applied to the rice plant based on at least one of the following: whether the estimated amount of leaf compounds is below threshold α, and whether the estimated grain protein content is below threshold β, and controls the irrigation device 30 accordingly. This makes it possible to determine information on the stress to be applied to the rice plant using a simple method based on thresholds.

[0046] Furthermore, in this first embodiment, the machine learning model for estimating the amount of leaf compounds is a model that has been trained using a combination of first spectral data obtained from cultivated rice and measured values ​​of leaf compound amounts as training data. In addition, the machine learning model for estimating the grain protein content is a model that has been trained using a combination of second spectral data obtained from cultivated rice and measured values ​​of grain protein content as training data. As a result, the amount of leaf compounds and grain protein content can be estimated with high accuracy.

[0047] In the first embodiment described above, the case in which the stress intensity is increased (S22) when the judgment in step S20 is rejected was explained, but it is not limited to this. For example, the stress intensity may be increased when the grain protein content is below the threshold β and has not changed much compared to the previous day, or even if the grain protein content is below the threshold β, if it has changed by a predetermined amount compared to the previous day, the stress intensity may be maintained without being increased. In addition, the stress intensity (nutrient solution EC) may be adjusted using other logic.

[0048] Note that the processing shown in Figure 6 of the first embodiment described above is just one example. For example, the stress control unit 65 may determine how to adjust the stress (nutrient solution EC) applied to the rice based on the estimated amount of leaf compounds and the seed protein content, without using thresholds α and β.

[0049] In the first embodiment described above, the case in which the leaf compound amount estimation unit 63 and the seed protein content estimation unit 64 estimate the leaf compound amount and seed protein content using a machine learning model was described, but the invention is not limited to this. For example, formulas showing the relationship between the reflectance of light of a specific wavelength and the amount of leaf compounds (a multivariate analysis model) and formulas showing the relationship between the reflectance of light of a specific wavelength and the seed protein content (a multivariate analysis model) may be prepared in advance, and the leaf compound amount and seed protein content may be estimated by inputting the reflectance of light of a specific wavelength obtained from spectral data into these formulas.

[0050] (Variation 1) In the first embodiment described above, the case in which the nutrient solution EC is adjusted when applying stress to rice was explained, but this is not the only example. Figure 8 is a schematic diagram showing the configuration of the cultivation support system 200 according to Modification 1. As shown in Figure 8, the cultivation support system 200 of Modification 1 is equipped with a UV (ultraviolet) irradiation device 40 as a device for applying stress. The irrigation device 30 is used for irrigating rice for cultivation, as in the first embodiment, but is not used to adjust the stress applied to the rice.

[0051] It is known that irradiating crops such as rice with UV (ultraviolet) light increases the content of grain protein and anthocyanins (see, for example, "Hidema J et al (2005) Changes in grain size and grain storage protein of rice (Oryza Sativa L.) in response to elevated UV-B radiation under outdoor conditions. J Radiat Res 46(2), 143-9. doi: 10.1269 / jrr.46.143.", "Zhou B et al (2016) Exploring miRNAs involved in blue / UV-A light response in Brassica rapa reveals special regulatory mode during seedling development. BMC Plant Biol 16(1), 111. doi: 10.1186 / s12870-016-0799-z").

[0052] Therefore, in this modified example 1, the UV intensity (W / m²) of the UV irradiation device 40 is 2 By adjusting this, the stress imposed on the rice plants can be controlled.

[0053] The configuration and functions of the control device 10 in this modified example 1 are the same as in Figures 2 and 3, and the processing of the control device 10 is the same as in Figure 6. However, it differs from the above embodiment in that the stress intensity adjustment in steps S22 and S26 of Figure 6 is performed using the UV irradiation device 40.

[0054] For example, in this modified example 1, assuming that the process shown in Figure 6 is started, during a predetermined period d1 after heading, the UV intensity is u(W / m) as shown in Figure 9(a). 2) is assumed to be maintained at ). Then, in step S22, if the stress control unit 65 increases the stress intensity, it instructs the UV irradiation device 40 to increase the UV intensity by, for example, Δu1, as shown in Figure 9(b) at day d2 after heading. By increasing the UV intensity in this way, an increase in the grain protein content can be expected. Note that Δu1 may be a predetermined value or a value corresponding to the estimated grain protein content (for example, a value corresponding to the difference between the threshold β and the estimated grain protein content). Also, in step S26, if the stress control unit 65 decreases the stress intensity, it instructs the UV irradiation device 40 to stop UV irradiation, for example, as shown in Figure 9(c) at day d2 after heading. Note that in step S26, the stress control unit 65 may also weaken the UV intensity, and the degree of weakening may be predetermined or a degree corresponding to the estimated amount of leaf compounds (for example, a value corresponding to the difference between the threshold α and the estimated amount of leaf compounds).

[0055] In this modified example 1, the period during which the treatment shown in Figure 6 is performed may differ from the period during which the treatment shown in Figure 6 is performed in the first embodiment (from heading to harvest). For example, if the effects of UV irradiation on the grain protein content continue until just before harvest, the treatment shown in Figure 6 may be performed until just before harvest in this modified example 1.

[0056] (Modification 2) In the first embodiment described above, the nutrient solution EC was adjusted when stressing the rice plants, and in the modified example 1 described a case where the UV intensity was adjusted when stressing the rice plants. However, the invention is not limited to these cases. For example, both the nutrient solution EC and the UV intensity may be adjusted. Figure 10 is a schematic diagram showing the configuration of the cultivation support system 300 according to modified example 2. As shown in Figure 10, the cultivation support system 300 of this modified example 2 is equipped with an irrigation device 30 and a UV irradiation device 40 as devices for applying stress. The stress control unit 65 adjusts the stress applied to the rice plants by combining the adjustment of the nutrient solution EC by the irrigation device 30 and the adjustment of the UV intensity by the UV irradiation device 40. This makes it possible to adjust the stress more appropriately according to the cultivation conditions, etc.

[0057] In the first embodiment and variations 1 and 2 described above, the target crop is rice, and the method describes adjusting the protein content as a useful substance contained in the rice grain. However, it is not limited to this. For example, the target crop may be grains such as wheat and barley, or legumes such as soybeans and adzuki beans. The useful substance may also be carbohydrates or sugars. Furthermore, in the first embodiment described above, genetically modified crops were used as an example, but it is not limited to these, and non-genetically modified crops may also be used. Moreover, if the target crop is a crop that bears fruit, the first embodiment and variations 1 and 2 may be applied to adjust the content of useful substances contained in the fruit.

[0058] In the first embodiment and modifications 1 and 2 described above, the control device 10 outputs information for stress adjustment to the irrigation device 30 and the UV irradiation device 40, but it is not limited to this. For example, the control device 10 may notify the operator by displaying (outputting) information for stress adjustment on a display screen. In this case, the operator may manually adjust the nutrient solution EC of the irrigation device 30 and the UV intensity of the UV irradiation device 40 based on the outputted information.

[0059] In the first embodiment and modifications 1 and 2 described above, the first spectral data of the leaf portion and the second spectral data of the fruit portion were obtained using the first spectrometer 21 and the second spectrometer 22, but the invention is not limited to this. For example, the range for obtaining the first spectral data and the range for obtaining the second spectral data may be separated in a single spectrometer, and the first and second spectral data may be obtained from each range.

[0060] In the first embodiment and modifications 1 and 2 described above, the amount of leaf pigments (anthocyanins and chlorophyll) was given as an example of the amount of intraleaf compounds estimated based on the first spectral data of the leaf portion, but it is not limited to this. The amount of intraleaf compounds other than leaf pigments may also be estimated from the first spectral data of the leaf portion. Intraleaf compounds other than leaf pigments can be compounds whose content changes due to stress, such as sugars and flavonoids (including anthocyanins) in the leaves.

[0061] 《Second Embodiment》 In the first embodiment and modifications 1 and 2 described above, a case was described in which first spectral data of the leaf portion and second spectral data of the grain portion are acquired, and the stress (nutrient solution EC and UV) applied to the rice is adjusted based on these. In contrast, in this second embodiment, the stress (nutrient solution EC and UV) applied to the rice is adjusted based on the first spectral data of the leaf portion.

[0062] Furthermore, the configuration of the cultivation support system 100 in this second embodiment is assumed to be the same as the configuration in Figure 1, but with the second spectrometer 22 removed.

[0063] (Regarding the processing of the control device 10) Next, the processing of the control device 10 in the second embodiment will be described in detail, following the flowchart in Figure 11. As a prerequisite for the start of the processing in Figure 11, the rice plants have gone through the tillering stage after planting and reached the heading stage, and during a predetermined period d1 after heading, the nutrient solution EC is gradually increased and reaches a predetermined value e [mS / cm] (see Figure 7(a)).

[0064] When the process shown in Figure 11 begins, in step S10, the first spectral data acquisition unit 61 acquires first spectral data from the first spectrometer 21. The first spectral data acquisition unit 61 inputs the acquired first spectral data to the leaf compound amount estimation unit 63. Then, in step S12, the leaf compound amount estimation unit 63 estimates the amount of leaf compounds based on the first spectral data input from the first spectral data acquisition unit 61 (similar to step S12 in Figure 6).

[0065] Next, in step S14, the stress control unit 65 determines whether the amount of leaf compounds (estimated value) is α or less. In other words, it determines whether the rice plant is in a state of high stress. If the determination in step S14 is affirmed, that is, if the plant is not in a state of high stress, the process proceeds to step S21.

[0066] When the process moves to step S21, the stress control unit 65 determines whether or not it is within a predetermined period from the start of processing in Figure 11. Here, the predetermined period refers to the period (initial period) during which, due to the characteristics of rice growth, the grain protein content does not reach the target even if the stress intensity is set high. If the determination in step S21 is affirmative, the process moves to step S24, and the stress control unit 65 waits for one day. During this waiting period, the stress intensity is maintained.

[0067] On the other hand, if the judgment in step S21 is rejected, it means that the rice grain protein content remains low even after the predetermined period. In this case, the stress control unit 65 proceeds to step S22 and instructs the irrigation device 30 to increase the stress intensity. As a result, the irrigation device 30 increases the nutrient solution EC by a predetermined amount (for example, Δe1 in Figure 7(b)). By increasing the nutrient solution EC, an increase in grain protein content can be expected. After step S22, in step S24, the stress control unit 65 waits for one day. After waiting for one day in step S24, the process returns to step S10.

[0068] In this first embodiment, the process shown in Figure 6 is repeated on a daily basis, so a one-day wait is performed in step S24 and step S28, which will be described later. However, the waiting time is arbitrary and may be longer or shorter than one day.

[0069] On the other hand, if the judgment in step S14 is rejected, that is, if the rice is in a state of high stress, the process proceeds to step S26, where the stress control unit 65 instructs the irrigation device 30 to reduce the stress intensity. As a result, the irrigation device 30 reduces the nutrient solution EC by a predetermined amount (for example, Δe2 in Figure 7(c)). By reducing the nutrient solution EC, the stress on the rice can be reduced. After step S26, the stress control unit 65 waits for one day in step S28, and after one day has elapsed, the process returns to step S10.

[0070] The above process is then repeated. For example, once the maturity stage is passed (before harvest time), the process shown in Figure 11 is forcibly terminated.

[0071] As mentioned above, the effects of stress on rice plants appear relatively quickly in the amount of compounds within the leaves. Therefore, by performing the process shown in Figure 11, as in this second embodiment, the stress intensity can be identified from the amount of compounds within the leaves in the initial stage, and the stress can be controlled (nutrient solution EC can be adjusted). Furthermore, even after that (when the amount of compounds within the leaves exceeds α), the stress can be controlled (nutrient solution EC can be adjusted) based on the amount of compounds within the leaves. This allows for appropriate control of stress (nutrient solution EC), enabling rice plants to produce more grain protein.

[0072] Figure 12(a) is a graph showing the grain protein content at harvest when EC control is not performed (comparative example) and when EC control is performed as shown in Figure 11 (in this second embodiment). As shown in Figure 12(a), it can be seen that the grain protein content increases when EC control is performed compared to when EC control is not performed. Since the antigen protein content is proportional to the grain protein content, cultivating rice into which vaccine antigen protein has been introduced by genetic modification while performing EC control as in this second embodiment can increase the target antigen protein content.

[0073] As described in detail above, according to this second embodiment, the first spectral data acquisition unit 61 acquires first spectral data of the rice leaf portion (S10). The leaf compound amount estimation unit 63 inputs the acquired first spectral data to a machine learning model that shows the relationship between the first spectral data and the amount of leaf compounds, and estimates the amount of leaf compounds (S12). Then, the stress control unit 65 determines the stress information to be applied to the rice based on the estimated amount of leaf compounds and outputs it to the irrigation device 30 (S22, S26). In this second embodiment, by monitoring the amount of leaf compounds that respond to stress and adjusting the nutrient solution EC via the irrigation device 30 based on this, the amount of the target substance, grain protein, can be appropriately adjusted. Furthermore, when cultivating rice into which vaccine antigen protein has been introduced by genetic modification, it is necessary to increase the production of vaccine antigen protein, but as described above, by increasing the grain protein content of the rice, it becomes possible to produce more vaccine antigen protein. Furthermore, in this second embodiment, since it is not necessary to use the relatively expensive second spectrometer 22, the amount of rice grain protein can be appropriately adjusted at a low cost.

[0074] In the second embodiment described above, the case in which the nutrient solution EC is adjusted when stressing the rice plants was explained, but the invention is not limited to this, and as with the modification 1 of the first embodiment, the UV intensity (W / m²) of the UV irradiation device 40 may be adjusted. 2The stress imposed on the rice plants may be adjusted by adjusting the following:

[0075] Figure 12(b) is a graph showing the grain protein content at harvest when UV irradiation control is not performed (comparative example) and when UV irradiation control is performed. As shown in Figure 12(b), it can be seen that the rice grain protein content increases when UV irradiation control is performed compared to when UV irradiation control is not performed. As a result, the production of vaccine antigen protein can be increased, similar to the second embodiment described above.

[0076] The above processing functions can be implemented by a computer. In this case, a program describing the processing content of the functions that the processing unit should have is provided. By executing this program on a computer, the above processing functions are implemented on the computer. The program describing the processing content can be recorded on a computer-readable recording medium (excluding carrier waves).

[0077] When distributing a program, it may be sold in the form of a portable recording medium such as a DVD (Digital Versatile Disc) or CD-ROM (Compact Disc Read Only Memory) on which the program is recorded. Alternatively, the program can be stored in the storage device of a server computer and transferred from the server computer to other computers via a network.

[0078] A computer executing a program stores the program, for example, on a portable storage medium or transferred from a server computer, in its own memory. The computer then reads the program from its memory and executes the processing according to the program. Alternatively, the computer can directly read the program from the portable storage medium and execute the processing according to that program. Furthermore, the computer can sequentially execute the processing according to the program received each time it is transferred from a server computer.

[0079] The embodiments described above are preferred examples of the present invention. However, the invention is not limited thereto, and various modifications are possible without departing from the spirit of the invention. [Explanation of Symbols]

[0080] 10 Control device (crop cultivation support device) 21 1st spectrometer 22 Second spectrometer 30 Irrigation equipment 40 UV irradiation device 61. First spectral data acquisition unit (part of the acquisition unit) 62 Second spectral data acquisition unit (part of the acquisition unit) 63. Leaf compound quantity estimation unit (first estimation unit) 64. Seed protein content estimation unit (second estimation unit) 65 Stress Control Unit (Processing Unit) 70. Learning Model Generation Unit 100 Cultivation Support Systems

Claims

1. Obtain first spectral data from the leaf portion of the crop and second spectral data from the fruit or grain portion. The acquired first spectral data is input into a first model that shows the relationship between the first spectral data and the amount of a specific compound contained in the leaves, and the amount of the specific compound corresponding to the acquired first spectral data is estimated. The acquired second spectral data is input into a second model that shows the relationship between the second spectral data and the amount of a specific useful substance contained in the fruit or grain, and the amount of the specific useful substance corresponding to the acquired second spectral data is estimated. A method for assisting crop cultivation, characterized in that a computer performs a process to determine and output information on the stress to be applied to the crop based on the estimated amount of the specific compound and the estimated amount of the specific useful substance.

2. In the process of determining and outputting the aforementioned stress information, The crop cultivation assistance method according to claim 1, characterized in that it determines and outputs whether to increase or decrease the stress to be applied to the crop based on at least one of whether the estimated amount of the specific compound is below a first threshold and whether the estimated amount of the specific useful substance is below a second threshold.

3. The crop cultivation support method according to claim 1, characterized in that the stress to be applied to the crop is at least one of high EC irrigation and UV irradiation.

4. The first model is a model that has been trained using a combination of the first spectral data obtained from a crop under cultivation and the measured amount of the specific compound as training data. The crop cultivation assistance method according to claim 1, characterized in that the second model is a model that has been trained using a combination of the second spectral data obtained from the crop being cultivated and the measured amount of the specific useful substance as training data.

5. The aforementioned crop is rice. The aforementioned specific compound is anthocyanin or chlorophyll. The crop cultivation support method according to claim 1, characterized in that the aforementioned specific useful substance is a protein.

6. Obtain first spectral data from the leaf portion of the crop and second spectral data from the fruit or grain portion. The acquired first spectral data is input into a first model that shows the relationship between the first spectral data and the amount of a specific compound contained in the leaves, and the amount of the specific compound corresponding to the acquired first spectral data is estimated. The acquired second spectral data is input into a second model that shows the relationship between the second spectral data and the amount of a specific useful substance contained in the fruit or grain, and the amount of the specific useful substance corresponding to the acquired second spectral data is estimated. A crop cultivation support program characterized by causing a computer to perform a process that determines and outputs information on the stress to be applied to the crop based on the estimated amount of the specific compound and the estimated amount of the specific useful substance.

7. An acquisition unit that acquires first spectral data of the leaf portion of a crop and second spectral data of the fruit or grain portion, A first estimation unit inputs the acquired first spectral data into a first model that shows the relationship between first spectral data and the amount of a specific compound contained in the leaves, and estimates the amount of the specific compound corresponding to the acquired first spectral data. A second estimation unit inputs the acquired second spectral data into a second model that shows the relationship between the second spectral data and the amount of a specific useful substance contained in the fruit or grain, and estimates the amount of the specific useful substance corresponding to the acquired second spectral data. A crop cultivation support device comprising: a processing unit that determines and outputs information on the stress to be applied to the crop based on the estimated amount of the specific compound and the estimated amount of the specific useful substance.

8. First spectral data of the leaf portion of the crop was obtained. The acquired first spectral data is input into a first model that shows the relationship between the first spectral data and the amount of a specific compound contained in the leaves, and the amount of the specific compound corresponding to the acquired first spectral data is estimated. A method for assisting crop cultivation, characterized in that a computer performs a process to determine and output information on the stress to be applied to the crop based on the estimated amount of the specific compound.

9. First spectral data of the leaf portion of the crop was obtained. The acquired first spectral data is input into a first model that shows the relationship between the first spectral data and the amount of a specific compound contained in the leaves, and the amount of the specific compound corresponding to the acquired first spectral data is estimated. A crop cultivation support program characterized by causing a computer to perform a process that determines and outputs information on the stress to be applied to the crop based on the estimated amount of the specific compound.

10. An acquisition unit that acquires first spectral data of the leaf portion of a crop, A first estimation unit inputs the acquired first spectral data into a first model that shows the relationship between first spectral data and the amount of a specific compound contained in the leaves, and estimates the amount of the specific compound corresponding to the acquired first spectral data. A crop cultivation support device comprising: a processing unit that determines and outputs information on the stress to be applied to the crop based on the estimated amount of the specific compound.

Images