Fruit vegetable information prediction method and fruit vegetable information prediction program

Abstract

To accurately predict the size or standards of fruit vegetables at the time of harvesting.SOLUTION: Provided is a fruit vegetable information prediction method for predicting the size or standards of fruit vegetables at the time of harvesting, wherein a computer executes processing of predicting the size or standards of fruit vegetables to be predicted at the time of harvesting by obtaining a plurality of pieces of learning data including the size of the fruit vegetables measured in the course of cultivation, an accumulated temperature from flowering at a time point when the size is measured, and the size or standards at the time of harvesting of the fruit vegetables, generating a prediction model that predicts the size or standards of the fruit vegetables at the time of harvesting using the plurality of pieces of learning data, and inputting input data including the size of the fruit vegetables to be predicted measured during cultivation of the fruit vegetables to be predicted, and the accumulated temperature from flowering at the time point when the size is measured, into the prediction model.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present invention relates to a fruit and vegetable information prediction method and a fruit and vegetable information prediction program. [Background technology]

[0002] Fruit vegetables such as tomatoes are classified into shipping standards after harvest, and a unit price is set for each shipping standard. For this reason, if the yield for each shipping standard can be predicted, for example, two weeks before harvest, producers can negotiate with the shipping destination regarding the unit price, etc., based on the prediction results. In addition, if the yield for each shipping standard can be predicted in the early stages of fruit growth, it is possible to determine the settings for environmental control and cultivation management policies such as fruit thinning based on the prediction results. This makes it possible to increase the proportion of high-priced grades and reduce non-standard products, thereby increasing profitability.

[0003] Conventionally, there is known a technique for estimating the condition of a fruit based on biological information and environmental information of the crop, and estimating the yield of the crop based on the estimated condition of the fruit (see, for example, Patent Document 1, etc.). Also, there is known a technique for predicting the specifications of a harvested crop by inputting environmental information up to the harvest date into a prediction model that predicts the specifications of the harvested crop from environmental information (see, for example, Patent Document 2, etc.). Also, there is known a technique for predicting the size of a fruit based on an integrated value of the leaf area (see, for example, Patent Document 3, etc.). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2021-26771 [Patent Document 2] Patent Publication No. 2023-3705 [Patent Document 3] JP 2019-187259 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, the above-mentioned Patent Document 1 only estimates the yield of the crop, and is not capable of estimating the size (including dimensions and weight) or specifications of the crop. In addition, the above-mentioned Patent Documents 2 and 3 are technologies for predicting crop specifications and fruit size, but there is room for improvement in terms of prediction accuracy when it comes to predicting the size and specifications of fruit vegetables.

[0006] The present invention aims to provide a fruit and vegetable information prediction method and a fruit and vegetable information prediction program that are capable of accurately predicting the size and specifications of fruit and vegetable at the time of harvest. [Means for solving the problem]

[0007] According to a first aspect, the fruit and vegetable information prediction method is a fruit and vegetable information prediction method for estimating the size or standard of a fruit or vegetable at the time of harvest, which includes obtaining a plurality of learning data including the size of the fruit or vegetable measured during cultivation, the accumulated temperature from after flowering at the time the size was measured, and the size or standard of the fruit or vegetable when it is harvested, generating a prediction model for predicting the size or standard of the fruit or vegetable at the time of harvest using the plurality of learning data, and predicting the size or standard of the fruit or vegetable at the time of harvest by inputting input data including the size of the fruit or vegetable to be predicted measured while the fruit or vegetable to be predicted is being cultivated, and the accumulated temperature from after flowering at the time the size was measured into the prediction model, in which the computer performs processing to predict the size or standard of the fruit or vegetable to be predicted at the time of harvest.

[0008] According to a second aspect, the fruit and vegetable information prediction method is a fruit and vegetable information prediction method for estimating the size or standard of a fruit or vegetable at the time of harvest, which includes acquiring a plurality of learning data including the maximum value of the rate of enlargement of the size of the fruit or vegetable during cultivation, the accumulated temperature from after flowering at the time when the maximum value was obtained and / or the size of the fruit or vegetable at the time when the maximum value was obtained, and the size or standard of the fruit or vegetable when it is harvested, generating a prediction model for predicting the size or standard of the fruit or vegetable at harvest using the plurality of learning data, and predicting the size or standard of the fruit or vegetable at harvest by inputting input data including the maximum value of the rate of enlargement of the size of the fruit or vegetable to be predicted while the fruit or vegetable to be predicted is being cultivated, the accumulated temperature from after flowering at the time when the maximum value was obtained and / or the size of the fruit or vegetable to be predicted at the time when the maximum value was obtained into the prediction model, in which a computer executes a process to predict the size or standard of the fruit or vegetable to be predicted at harvest. Effect of the Invention

[0009] The fruit and vegetable information prediction method and fruit and vegetable information prediction program of the present invention have the effect of accurately predicting the size and specifications of fruit and vegetables at the time of harvest. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram showing a configuration of an agricultural system according to a first embodiment. [Diagram 2] FIG. 2(a) is a diagram showing the hardware configuration of a server, and FIG. 2(b) is a diagram showing the hardware configuration of a user terminal. [Diagram 3] FIG. 3(a) is a functional block diagram of a user terminal, and FIG. 3(b) is a diagram for explaining a method for measuring the size of a fruit using an image. [Figure 4] FIG. 4 is a functional block diagram of the server. [Diagram 5] FIG. 5 is a diagram showing an example of a data structure of the fruit information DB. [Figure 6]FIG. 6(a) is a diagram showing an example of a data structure of a learning data DB, and FIG. 6(b) is a diagram for explaining a prediction model according to the first embodiment. [Figure 7] FIG. 7 is a graph showing the change in fruit fresh weight versus the number of days after flowering for each of the different genotypes of tomatoes. [Figure 8] FIG. 8(a) is a diagram showing an example of input data, FIG. 8(b) is a diagram showing an example of predicted data, and FIG. 8(c) is a diagram showing an example of a counting result. [Figure 9] FIG. 9(a) is a flowchart showing the processing performed by the server during learning, and FIG. 9(b) is a flowchart showing the processing performed by the server during prediction. [Figure 10] Figure 10(a) is a graph showing the relationship between the results of predicting the vertical diameter x horizontal diameter at harvest using Voting Regression using input data for accumulated temperatures after flowering up to 800°C and the actual vertical diameter x horizontal diameter at harvest, and Figure 10(b) is a graph showing the average fruit diameter error (%) when using the random forest method, Lasso regression, and Voting Regression. [Figure 11] Figure 11(a) is a graph showing the relationship between the vertical diameter x horizontal diameter (cm2) of a fruit and fruit weight, Figure 11(b) is a table showing the relationship between fruit weight and shipping standards, and Figure 11(c) is a diagram showing an example of the harvest yield (number of fruits) tabulated for each shipping standard. [Figure 12] FIG. 12 is a diagram showing an example of an output when the tabulation result of FIG. 11(C) is output. [Figure 13] FIG. 13(a) is a diagram showing a learning data DB when fruit and vegetable weight is used as learning data, and FIG. 13(b) is a diagram showing a prediction model generated using the learning data of FIG. 13(a). [Figure 14] FIG. 14(a) is a diagram for explaining a prediction model of variant example 1 of the first embodiment, and FIG. 14(b) is a diagram for explaining a prediction model when variant example 1 is applied to the example of FIG. 13(b). [Figure 15]FIG. 15 is a diagram illustrating an example of a data structure of the fruit information DB in the first modification of the first embodiment. [Figure 16] FIG. 16 is a diagram illustrating an example of a data structure of the learning data DB in the first modification of the first embodiment. [Figure 17] FIG. 17(a) is a diagram for explaining a prediction model of the modified example 2 of the first embodiment, FIG. 17(b) is a diagram showing an example of a data structure of a learning data DB in the modified example 2 of the first embodiment, and FIG. 17(c) is a diagram showing a modified example of FIG. 13(b). [Figure 18] FIG. 18(a) is a diagram showing an example of a data structure of a learning data DB according to the second embodiment, and FIG. 18(b) is a diagram for explaining a prediction model according to the second embodiment. [Figure 19] FIG. 19 is a graph showing the change in fruit enlargement rate relative to the number of days after flowering for each of the different tomato genotypes. [Figure 20] 20(a) and 20(b) are diagrams for explaining a prediction model according to a modification of the second embodiment. [Figure 21] FIG. 21(a) is a diagram showing a learning data DB when fruit and vegetable weight is used as learning data in the second embodiment, and FIG. 21(b) is a diagram showing a prediction model generated using the learning data of FIG. 21(a). [Figure 22] 22(a) and 22(b) are diagrams showing a modification of the model in FIG. 21(b). [Figure 23] FIG. 23(a) is a diagram showing an example of a data structure of a learning data DB according to the third embodiment, and FIG. 23(b) is a diagram for explaining a prediction model according to the third embodiment. [Figure 24] 24(a) and 24(b) are diagrams for explaining a modified example of the third embodiment. [Diagram 25] 25(a) and 25(b) are diagrams for explaining the fourth embodiment. [Figure 26] 26(a) and 26(b) are diagrams (part 1) for explaining a modified example of the fourth embodiment. [Figure 27] 27(a) and 27(b) are diagrams (part 2) for explaining a modified example of the fourth embodiment. [Figure 28] 28(a) and 28(b) are diagrams (part 3) for explaining a modified example of the fourth embodiment. [Figure 29] FIG. 29 is a diagram (part 4) for explaining a modified example of the fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] First Embodiment A first embodiment will be described in detail below with reference to Figs. 1 to 12. Fig. 1 shows a schematic configuration of an agricultural system 100 according to the first embodiment. The agricultural system 100 of this embodiment is a system that predicts and outputs the size (dimension in this first embodiment) of each fruit vegetable (fruit) at the time of harvest and the harvest volume (or shipping volume) for each size in a facility (e.g., a farm field, a greenhouse, etc.) where fruit vegetables such as tomatoes, peppers, and cucumbers are cultivated. In this first embodiment, a case will be described where the fruit vegetable cultivated in the facility is tomatoes.

[0012] 1, the agricultural system 100 includes a server 10 and a user terminal 70. The server 10 and the user terminal 70 are connected via a network 80 such as the Internet.

[0013] The server 10 predicts the size of each fruit at the time of harvesting based on the information input from the user terminal 70. The server 10 also counts the number of fruits that can be harvested (or shipped) for each size, and outputs the counting result to the user terminal 70.

[0014] FIG. 2(a) shows a schematic diagram of the hardware configuration of the server 10. As shown in FIG. 2(a), the server 10 includes a central processing unit (CPU) 90, a read only memory (ROM) 92, a random access memory (RAM) 94, a storage (such as a hard disk drive (HDD) or a solid state drive (SSD)) 96, a network interface 97, and a drive for a portable storage medium 99. These components of the server 10 are connected to a bus 98. In the server 10, the CPU 90 executes a program (including a fruit and vegetable information prediction program) stored in the ROM 92 or the HDD 96, or a program read by the drive for a portable storage medium 99 from the portable storage medium 91, thereby realizing the functions of the components shown in FIG. 4. The functions of the components shown in FIG. 4 may be realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). Details of the functions of FIG. 4 will be described later.

[0015] Returning to FIG. 1, the user terminal 70 is an information processing device such as a PC (Personal Computer) or a smartphone used by a producer. FIG. 2(b) shows a schematic diagram of the hardware configuration of the user terminal 70. As shown in FIG. 2(b), the user terminal 70 includes a CPU 190, a ROM 192, a RAM 194, a storage 196, a network interface 197, a display unit 193, an input unit 195, and the like. These components of the user terminal 70 are connected to a bus 198. The display unit 193 includes a liquid crystal display or the like, and the input unit 195 includes a keyboard, a mouse, a touch panel, and the like.

[0016] The user terminal 70 is capable of communicating with a temperature sensor 72 and a camera 74 installed in the facility. The temperature sensor 72 measures the temperature in the facility and transmits the measurement results to the user terminal 70 as appropriate. The camera 74 photographs the tomato plants grown in the facility at a predetermined time interval (for example, every predetermined number of days) and transmits the photographed images to the user terminal 70. The camera 74 is provided with a position measurement sensor such as a GNSS (Global Navigation Satellite System) sensor, and the photographed images are accompanied by information on the photographing position. Furthermore, when it is time to photograph, the camera 74 moves along a track laid in the facility and photographs images at a plurality of predetermined locations in the facility. However, this is not limited to the above, and the producer may operate the camera 74 to photograph images at a plurality of predetermined locations in the facility.

[0017] (Functions of the user terminal 70) Fig. 3(a) shows a functional block diagram of the user terminal 70. As shown in Fig. 3(a), in the user terminal 70, the CPU 190 executes a program to realize functions as a temperature information acquisition unit 170, an image acquisition unit 172, a fruit identification unit 174, a size measurement unit 176, a harvest size acquisition unit 178, an information collection unit 180, a transmission unit 182, a reception unit 184, and a display processing unit 186. Fig. 3(a) also shows a fruit information DB 188 stored in the storage 196 or the like.

[0018] The temperature information acquiring unit 170 acquires information on the temperature inside the facility transmitted from the temperature sensor 72, and transmits it to the information collecting unit 180. The temperature information acquiring unit 170 may also acquire temperature information manually input by the producer.

[0019] The image acquiring unit 172 transmits the image of the tomato stalk transmitted from the camera 74 to the fruit identifying unit 174. Fig. 3(b) shows an example of an image acquired by the image acquiring unit 172. As shown in Fig. 3(b), it is assumed that in the image, a reference object serving as a dimensional reference (such as a piece of paper on which a two-dimensional code recording dimensional information is printed) is photographed together with the tomato stalk.

[0020] The fruit identification unit 174 extracts bloomed flowers from the image, assigns a fruit ID to the extracted flowers, and identifies the blooming date. The fruit identification unit 174 also extracts fruits from the image, and assigns fruit identification information (stump ID, bunch ID, fruit ID) to each fruit. At this time, the fruit identification unit 174 assigns the identification information taking into consideration the location information at which the image was captured. For example, if a flower extracted in the current image was not extracted in an image captured previously at the same location, the fruit identification unit 174 assigns new identification information to the flower. If a flower (or fruit) corresponding to the flower (or fruit) extracted in the current image was also present in the image captured previously, the fruit identification unit 174 assigns the identification information already assigned to the flower (or fruit) as is.

[0021] The fruit identification unit 174 transmits the image together with identification information of the flowers and fruits present in the image to the size measurement unit 176. In addition, the fruit identification unit 174 transmits information on the flowering date of each flower (fruit) to the information collection unit 180.

[0022] The size measuring unit 176 measures the size of each fruit extracted from the image. In this embodiment, as an example, the vertical diameter (cm) and horizontal diameter (cm) of the fruit are measured (see the double-headed arrow in FIG. 3(b)). At this time, the size measuring unit 176 measures the size of the fruit based on the size of a reference object captured in the image. Note that the reference objects are provided at multiple locations on the tomato stalk, and when measuring the size of the fruit, a reference object located near the fruit is used. The size measuring unit 176 links the measurement results of each fruit with the identification information and transmits them to the information collecting unit 180.

[0023] The harvest size acquisition unit 178 acquires the size of each fruit at harvest time input by the producer via the input unit 195, and transmits it to the information collection unit 180. The producer may measure the size (vertical diameter and horizontal diameter) of the harvested fruit with a vernier caliper, ruler, or the like, and manually input the measurement results into the user terminal 70 by linking them with the identification information.

[0024] The information collecting unit 180 compiles the information received from each unit and stores it in the fruit information DB 188. Here, the fruit information DB 188 has a data structure as shown in FIG. 5. Specifically, as shown in FIG. 5, the fruit information DB 188 stores the identification information of each fruit ("Plant ID", "Fruit Cluster ID", "Fruit ID"), the "flowering date" of each fruit, the "accumulated temperature after flowering (°C)" at the time each fruit was photographed and the size at that time ("Vertical diameter (cm)", "Horizontal diameter (cm)", "Vertical diameter x horizontal diameter (cm)"), and the like. 2 ) for each fruit, and the "length x width at harvest (cm 2 For example, for the fruit with fruit ID = "A1-1-1" in Figure 5, the flowering date is September 10, 2021, and the size when the accumulated temperature after flowering is 150°C is vertical diameter = 1.2 cm, horizontal diameter = 1.2 cm, vertical diameter x horizontal diameter = 1.44 cm. 2 When the accumulated temperature after flowering is 220℃, the size is vertical diameter = 1.8cm, horizontal diameter = 1.8cm, vertical diameter x horizontal diameter = 3.24cm. 2 ...and the length x width at harvest is 44.88 cm. 2 The data in Figure 5 is managed collectively for each facility.

[0025] The transmission unit 182 acquires data that can be learning data from the fruit information DB 188, and transmits it to the server 10. The data that can be learning data is the vertical diameter × horizontal diameter (cm) at harvest time from the flowering date in FIG. 2 ) is data on fruits for which all information up to the point of collection has been collected.

[0026] The transmission unit 182 also acquires data (input data used to predict the size of the fruit at harvest) obtained in the facility to be predicted from the fruit information DB 188, and transmits it to the server 10. The input data is data (accumulated temperature after blooming, vertical diameter, horizontal diameter, vertical diameter x horizontal diameter) obtained at a predetermined time point from when the fruit starts to turn color and the dimensions of the fruit can be measured in the facility to be predicted, until a predetermined number of days (for example, two weeks) before the scheduled shipping date. More specifically, the input data is data (accumulated temperature after blooming, vertical diameter, horizontal diameter, vertical diameter x horizontal diameter) obtained at a certain time point in the second, third, and fourth periods of the first, second, third, fourth, and fifth periods that are obtained by dividing the period from blooming to the scheduled shipping date in the facility to be predicted based on the accumulated temperature after blooming.

[0027] The receiving unit 184 receives information output from the server 10. The display processing unit 186 displays on the display unit 193 the information received by the receiving unit 184.

[0028] (About Server 10 functions) Fig. 4 shows a functional block diagram of the server 10. In the server 10, a CPU 90 executes a program to realize functions as a learning data collection unit 30, a learning data processing unit 31, a prediction model generation unit 32, an information acquisition unit 34, a prediction unit 36, and an output unit 38, as shown in Fig. 4. Note that Fig. 4 also shows a learning data DB 50 and a prediction model storage unit 52 stored in a storage 96 or the like.

[0029] The learning data collection unit 30 collects learning data (past data) transmitted from the transmission unit 182 of the user terminal 70 , and transmits it to the learning data processing unit 31 .

[0030] The learning data processing unit 31 processes (standardizes) the learning data (past data in FIG. 5) received from the learning data collection unit 30, and stores the data in a learning data DB 50 as shown in FIG. 6(a). Specifically, the learning data processing unit 31 stores the accumulated temperature after flowering and the vertical diameter × horizontal diameter (cm 2Then, the learning data processing unit 31 creates a simple regression curve or a regression curve showing the relationship between the vertical diameter x horizontal diameter (cm) when the post-flowering accumulated temperature is 300°C, 400°C, 500°C, 600°C, 700°C, and 800°C from the created curve. 2 ) and stores it in the learning data DB 50 in FIG. 6(a). In addition, the learning data processing unit 31 acquires the vertical diameter × horizontal diameter (cm 2 ) is stored as is in the learning data DB50.

[0031] The prediction model generating unit 32 generates a prediction model using the learning data stored in the learning data DB 50. Specifically, the prediction model generating unit 32 uses a machine learning algorithm (a regression model or classification model by machine learning) to calculate the post-flowering accumulated temperature in FIG. 6(a) and the vertical diameter × horizontal diameter (cm) corresponding to the post-flowering accumulated temperature. 2 ) combination was used as an explanatory variable, and the vertical diameter × horizontal diameter (cm 2 ) as the objective variable. As the machine learning algorithm, a random forest method, Lasso regression, Voting Regression, which is a type of ensemble learning, or the like can be used. The prediction model generation unit 32 generates a prediction model (FIG. 6(b)) for each variety using the learning data for each variety, and stores the generated prediction model in the prediction model storage unit 52.

[0032] Figure 7 is a graph showing the change in fresh fruit weight (g) versus the number of days after flowering (days) for each of the different genotypes of tomatoes. This graph is shown as Fig. 2A in 'Identification of growth processes involved in QTLs for tomato fruit size and composition' Nadia Bertin, Journal of Experimental Botany, Volume 60, Issue 1, January 2009, Pages 237-248. As shown in Figure 7, it is highly likely that the way in which fresh fruit weight changes after flowering will affect the fresh fruit weight at harvest. For this reason, it is highly likely that the way in which fresh fruit weight changes will also affect the size of the fruit at harvest (vertical diameter × horizontal diameter at harvest). Therefore, as shown in Figure 6(b), the relationship between the accumulated temperature after flowering and the vertical diameter × horizontal diameter (cm) corresponding to the accumulated temperature after flowering is shown. 2 ) and each combination of are used as explanatory variables, and the vertical diameter × horizontal diameter at harvest (cm 2 If we use a prediction model with the objective variable of ) the vertical diameter × horizontal diameter (cm 2 ) can be predicted with high accuracy.

[0033] The information acquisition unit 34 acquires information (input data to be input to the prediction model) required to predict the size of fruits at harvest time from the user terminal 70. The input data is assumed to be, for example, data as shown in Fig. 8(a). That is, the input data includes information on the post-flowering accumulated temperature and the vertical diameter x horizontal diameter of each fruit (prediction target fruit) being cultivated in a certain facility (prediction target facility) at the time of the post-flowering accumulated temperature.

[0034] The prediction unit 36 ​​inputs the information received from the information acquisition unit 34 (a combination of the accumulated temperature after flowering of each fruit and the value of the vertical diameter x horizontal diameter) into a prediction model (FIG. 6(b)) to predict the size (vertical diameter x horizontal diameter) of each fruit harvested at the facility being predicted (see the predicted data in FIG. 8(b)). The prediction unit 36 ​​also organizes the data in FIG. 8(b) by size range and tabulates the harvest volume (number of fruits) for each size range.

[0035] The output unit 38 transmits the tabulated result of the prediction unit 36 ​​to the user terminal 70.

[0036] (flowchart) Next, the processing of the server 10 will be described in detail with reference to the flowcharts of FIGS. 9(a) and 9(b).

[0037] (Learning process) 9(a) is a flowchart showing a learning process performed by the server 10. The process in FIG. 9(a) is executed when learning data is received from the user terminal 70.

[0038] 9(a) is started, first, in step S10, the learning data collection unit 30 collects learning data (past data) transmitted from the transmission unit 182 of the user terminal 70, and transmits the learning data to the learning data processing unit 31. The learning data collection unit 30 collects, for example, data as shown in FIG. 5. It is assumed that the learning data is transmitted from a large number of user terminals 70.

[0039] Next, in step S12, the learning data processing unit 31 processes (normalizes) the learning data (FIG. 5) received from the learning data collection unit 30, and stores the data in the learning data DB 50 (FIG. 6(a)).

[0040] Next, in step S14, the prediction model generation unit 32 generates a prediction model ( FIG. 6(b) ) by machine learning using the learning data stored in the learning data DB 50. The prediction model generation unit 32 stores the generated prediction model ( FIG. 6(b) ) in the prediction model storage unit 52.

[0041] With the above, the entire process of FIG. 9(a) is completed.

[0042] (Prediction processing) Fig. 9(b) is a flowchart showing the process at the time of prediction by the server 10. The process of Fig. 9(b) is executed when input data (Fig. 8(a)) is received from the user terminal 70.

[0043] 9(b) is started, first, in step S20, the information acquisition unit 34 acquires input data from the user terminal 70. Note that each piece of input data is linked with information about the variety of tomato that the data pertains to.

[0044] Next, in step S22, the prediction unit 36 ​​inputs the input data (FIG. 8(a)) received from the information acquisition unit 34 into the prediction model (FIG. 6(b)) to predict the size (vertical diameter x horizontal diameter) of each fruit at harvest time (see the prediction data in FIG. 8(b)). Note that the prediction unit 36 ​​reads out the prediction model corresponding to the variety linked to the input data from the prediction model storage unit 52.

[0045] Next, in step S24, the prediction unit 36 ​​tally up the harvest yield (number of fruits) for each size range. The output unit 38 transmits (outputs) the tally up result (FIG. 8(c)) by the prediction unit 36 ​​to the user terminal 70.

[0046] Figure 10(a) is a graph showing the relationship between the results of predicting the vertical diameter x horizontal diameter at harvest time by Voting Regression using input data for accumulated temperatures after flowering up to 800°C, and the actual vertical diameter x horizontal diameter at harvest time (actual measured value). The approximate formula was the formula (y = 0.8848x + 5.5294) shown in Figure 10(a). The coefficient of determination R 2 The value was 0.8827, which indicates that the vertical diameter x horizontal diameter at harvest could be predicted with high accuracy.

[0047] Figure 10(b) is a graph showing the average error (%) of fruit size when using the random forest method, lasso regression, and voting regression. In each graph, the black bar graph shows the average error of fruit size when using input data with an accumulated temperature after flowering up to 500°C, and the white bar graph shows the average error of fruit size when using input data with an accumulated temperature after flowering up to 800°C. From Figure 10(b), it can be seen that the prediction accuracy is higher when using input data with an accumulated temperature after flowering up to 800°C than when using input data with an accumulated temperature after flowering up to 500°C. It can also be seen that among the random forest method, lasso regression, and voting regression, the prediction accuracy is highest when using voting regression.

[0048] For example, when prediction is made using input data obtained at an early stage after flowering (data on accumulated temperature after flowering of 300℃ to 500℃), the yield for each size can be predicted at the early stage of fruit growth. In this case, producers can determine the policy of cultivation management such as setting environmental control and thinning of fruit based on the prediction results. This increases the ratio of high-priced grades and reduces non-standard products, thereby improving profitability. On the other hand, when prediction is made using input data obtained at a stage when the accumulated temperature after flowering is about 800℃ (about two weeks before harvest) (data on accumulated temperature after flowering of 300℃ to 800℃), the yield for each size can be predicted with high accuracy. In this case, producers can negotiate with shipping destinations about unit prices, etc. based on the prediction results.

[0049] The length x width of the fruit (cm 2 There is a relationship between the length x width (cm) of each fruit at harvest and the fruit weight as shown in FIG. 11(a). 2) and the relational expression in FIG. 11(a), the fruit weight (g) at the time of harvesting of each fruit may be calculated. The prediction unit 36 ​​may then specify the shipping standard of each fruit based on a table as shown in FIG. 11(b) showing the relationship between fruit weight and shipping standard, and may tally up the harvest volume (number of fruits) for each shipping standard as shown in FIG. 11(c). In this case, the output unit 38 may output the data in FIG. 11(c) to the user terminal 70, or alternatively or in addition to this, may generate a graph as shown in FIG. 12 based on the data in FIG. 11(c) and output it to the user terminal 70. The predicted shipping standard ratio on the vertical axis in FIG. 12 means the ratio of the number of fruits of each shipping standard to the total number of fruits (the total number of fruits of 3S to 2L). Depending on the type of fruit or vegetable, the shipping standard may be determined not by the fruit weight but by the dimensions (for example, vertical diameter x horizontal diameter). In this case, the shipping standard may be specified from a table showing the relationship between the dimensions and shipping standard without using the relational expression as shown in FIG. 11(a).

[0050] As described above in detail, according to the first embodiment, the learning data collection unit 30 collects learning data including a combination of the accumulated temperature after blooming and the size (vertical diameter x horizontal diameter) of the fruit at that time for each fruit, and data on the vertical diameter x horizontal diameter at harvest. The prediction model generation unit 32 generates a prediction model for predicting the vertical diameter x horizontal diameter at harvest based on the learning data. The prediction unit 36 ​​predicts the vertical diameter x horizontal diameter at harvest of the fruit being cultivated by inputting the accumulated temperature after blooming of the fruit being cultivated and the size (vertical diameter x horizontal diameter) of the fruit at that time into the prediction model. Since the vertical diameter x horizontal diameter at harvest of the fruit is considered to be affected by the way in which the size of the fruit changes during cultivation (FIG. 7), the above-mentioned prediction model can be used to accurately predict the vertical diameter x horizontal diameter at harvest of the fruit being cultivated.

[0051] Furthermore, in the first embodiment, the vertical diameter x horizontal diameter at harvest can be predicted in the early stage after flowering (after the fruit size can be measured) up to a predetermined number of days (for example, two weeks) before the shipping date. Therefore, if the vertical diameter x horizontal diameter at harvest is predicted based on input data obtained up to an early stage after flowering, the predicted result can be used for environmental control, etc. Furthermore, if the vertical diameter x horizontal diameter at harvest is predicted based on input data obtained up to a predetermined number of days before the shipping date, the predicted result can be used for negotiations with the shipping destination, etc.

[0052] In the above first embodiment, the input data of Fig. 8(a) is directly input to the prediction model, but the present invention is not limited to this. For example, the input data may be processed (normalized) in the same manner as the learning data, and the processed input data may be input to the prediction model.

[0053] As mentioned above, the length x width of the fruit (cm 2 ) and fruit weight has a relationship as shown in FIG. 11(a). Therefore, the learning data processing unit 31 may further process the learning data as shown in FIG. 6(a) obtained by processing (standardizing) the learning data received from the learning data collection unit 30 (past data in FIG. 5), into learning data as shown in FIG. 13(a). That is, based on the relationship in FIG. 11(a), the vertical diameter × horizontal diameter (cm 2 The weight of the fruit or vegetable (g) may be estimated (converted) from the data.

[0054] In this case, the prediction model generation unit 32 generates a prediction model using the learning data stored in the learning data DB 50 in Fig. 13(a). Specifically, the prediction model generation unit 32 uses a machine learning algorithm (a regression model or classification model by machine learning) to calculate a vertical diameter x horizontal diameter (cm) at harvest time using each combination of the post-flowering accumulated temperature in Fig. 13(a) and the fruit vegetable weight (g) corresponding to the post-flowering accumulated temperature as explanatory variables. 2 ) as the objective variable to generate a prediction model (see FIG. 13(b)).

[0055] The prediction unit 36 ​​then inputs the information received from the information acquisition unit 34 (a combination of the accumulated temperature after flowering of each fruit and the value of the fruit and vegetable weight) into the prediction model (FIG. 13(b)) to predict the size (vertical diameter x horizontal diameter) at the time of harvest of each fruit to be harvested in the facility being predicted. When an actual measurement value of the fruit and vegetable weight is input from the user terminal 70 to the information acquisition unit 34, the prediction unit 36 ​​may input the actual measurement value of the fruit and vegetable weight into the prediction model. When an actual measurement value of the fruit and vegetable dimensions (vertical diameter x horizontal diameter) is input from the user terminal 70 to the information acquisition unit 34, the prediction unit 36 ​​may input an estimated value of the fruit and vegetable weight estimated from the value of the vertical diameter x horizontal diameter into the prediction model.

[0056] In the above, the learning data processing unit 31 calculates the vertical diameter x horizontal diameter (cm) for each accumulated temperature after flowering in FIG. 6(a) based on the relationship in FIG. 11(a). 2 ), but the present invention is not limited to this. If the fruit information DB 188 in Fig. 5 contains data on the fruit and vegetable weight for each accumulated temperature after flowering, the data on the fruit and vegetable weight for each accumulated temperature after flowering can be obtained from the fruit information DB 188 in Fig. 5.

[0057] (Variation 1) In the first embodiment, as shown in FIG. 6(b), the input data for the prediction model is the accumulated temperature after flowering (°C) and the vertical diameter × horizontal diameter (cm 2 ) and the prediction result (prediction data) by the prediction model is the vertical diameter × horizontal diameter at harvest time. However, this is not limited to this, and as shown in FIG. 14(a), the input data of the prediction model may be the accumulated temperature after flowering (°C) and the vertical diameter × horizontal diameter at that time (cm 2 ) and the prediction result (prediction data) by the prediction model may be the fruit weight (g) at harvest time.

[0058] In this case, data (learning data) as shown in Fig. 15 is stored in the fruit information DB 188. Moreover, the prediction model generation unit 32 may use data as shown in Fig. 16 obtained by processing the learning data for generating a prediction model.

[0059] Even in this way, it is possible to accurately predict the weight of each fruit at harvest. Therefore, the output unit 38 can output the harvest volume (number of fruits) for each fruit weight range. In addition, the output unit 38 can convert the fruit weight into shipping standards using the table in Fig. 11(b), and output to the user terminal 70 the harvest volume (number of fruits) for each shipping standard as shown in Fig. 11(c) and information on the predicted shipping standard ratio for each shipping standard as shown in Fig. 12.

[0060] The example of Fig. 13(b) can also be modified in the same manner as Modification 1. That is, as shown in Fig. 14(b), the input data of the prediction model may be a combination of the accumulated temperature after flowering (°C) and the weight of the fruit or vegetable at that time, and the prediction result (prediction data) by the prediction model may be the fruit weight at harvest (g).

[0061] (Variation 2) In the above modification 1, the case where the prediction result (prediction data) by the prediction model is the fruit weight (g) at harvest time has been described, but this is not limiting, and the prediction result (prediction data) by the prediction model may be shipping specifications, as shown in FIG. 17(a). In this case, the prediction model generation unit 32 may use data such as that shown in FIG. 17(b) to generate the prediction model. In this case, the output unit 38 can output the information in FIG. 11(c) or FIG. 12 to the user terminal 70 without using the relational equation in FIG. 11(a) or the table in FIG. 11(b).

[0062] The example of Fig. 13(b) can also be modified in the same manner as Modification 2. That is, as shown in Fig. 17(c), the input data of the prediction model may be a combination of the accumulated temperature after flowering (°C) and the weight of the fruit or vegetable at that time, and the prediction result (prediction data) by the prediction model may be a harvest standard.

[0063] Second Embodiment The second embodiment will be described below with reference to Fig. 18(a) to Fig. 19. The second embodiment is characterized in that the learning data processing unit 31 processes the past data in Fig. 5, so that the learning data DB 50 stores learning data as shown in Fig. 18(a).

[0064] As can be seen by comparing the learning data in FIG. 18(a) with FIG. 6 (first embodiment), the maximum fruit enlargement speed (cm 2 The maximum fruit growth rate (cm / ℃) is shown below. 2 / ℃) will be explained.

[0065] The learning data processing unit 31 first calculates the fruit enlargement rate (a value indicating how much the vertical diameter x horizontal diameter increases) during the period when the accumulated temperature after flowering increases from 300°C to 400°C based on the following formula (1). Fruit growth rate = (20.7-11.6) / (400-300)=0.091(cm 2 / ℃) …(1)

[0066] The learning data processing unit 31 also calculates the fruit enlargement rate during the period when the post-flowering accumulated temperature is between 400°C and 500°C. Fruit growth rate = (31.4-20.7) / (500-400)=0.107(cm 2 / ℃) …(2)

[0067] In addition, the learning data processing unit 31 calculates the fruit enlargement rate when the accumulated temperature after flowering is from 500°C to 600°C, the fruit enlargement rate when the accumulated temperature after flowering is from 600°C to 700°C, and the fruit enlargement rate when the accumulated temperature after flowering is from 700°C to 800°C.

[0068] Then, the learning data processing unit 31 determines the maximum value of the calculated fruit thickening speeds as the maximum fruit thickening speed. In the case of FIG. 18(a), the fruit thickening speed (0.107 (cm)) during the period when the accumulated temperature after flowering is from 400° C. to 500° C. 2 / ℃) is the maximum fruit enlargement rate.

[0069] In the second embodiment, the prediction model generation unit 32 uses a machine learning algorithm to calculate the post-flowering accumulated temperature in FIG. 18(a) and the vertical diameter × horizontal diameter (cm 2 ) and the maximum fruit growth rate (cm 2 / ℃) and are the explanatory variables, and the vertical diameter × horizontal diameter at harvest (cm 2 ) as the objective variable to generate a prediction model (see FIG. 18(b)).

[0070] Figure 19 is a graph showing the change in fruit enlargement rate versus the number of days after flowering for tomatoes of different genotypes. This graph is shown as Fig. 2B in 'Identification of growth processes involved in QTLs for tomato fruit size and composition' Nadia Bertin, Journal of Experimental Botany, Volume 60, Issue 1, January 2009, Pages 237-248. The types of lines in Figure 19 represent the same genotypes as the types of lines in Figure 7. Figures 19 and 7 show that the higher the maximum value of the fruit enlargement rate (maximum fruit enlargement rate), the higher the possibility of the fruit fresh weight at harvest. Therefore, as mentioned above, the maximum fruit enlargement rate (cm 2 If a prediction model is created using the explanatory variables, the vertical diameter × horizontal diameter (cm 2 ) can be predicted with greater accuracy.

[0071] In the second embodiment, the information acquisition unit 34 acquires the same data as that shown in FIG. 8(a) as input data. Then, the prediction unit 36 ​​predicts the post-flowering accumulated temperature and the vertical diameter × horizontal diameter (cm 2 ) and the maximum fruit growth rate obtained from the input data using the same formulas as (1) and (2) above, are input into the prediction model. 2 ) can be predicted with high accuracy.

[0072] As described above, according to the second embodiment, the learning data includes the maximum fruit thickening speed, and the prediction model generation unit 32 generates a prediction model for predicting the vertical diameter x horizontal diameter at harvest based on the learning data including the maximum fruit thickening speed. The prediction unit 36 ​​predicts the vertical diameter x horizontal diameter at harvest by inputting a combination of the accumulated temperature after flowering of the fruit under cultivation and the size of the fruit at that time (vertical diameter x horizontal diameter), and the maximum fruit thickening speed into the prediction model. Since the vertical diameter x horizontal diameter at harvest of the fruit is considered to be affected by the maximum fruit thickening speed (FIGS. 19 and 7), the above-mentioned prediction model can be used to accurately predict the vertical diameter x horizontal diameter at harvest of the fruit under cultivation.

[0073] In addition, as shown in Figures 20(a) and 20(b), the prediction model of this second embodiment may, like the variant 1 and variant 2 of the first embodiment described above, generate prediction results (prediction data) in the form of fruit weight at harvest (g) or shipping specifications.

[0074] In the second embodiment, as in the first embodiment, the vertical diameter x horizontal diameter (cm 2 21(a), the fruit and vegetable weight (g) may be used instead of the cumulative temperature after flowering in FIG. 21(a). In this case, the prediction model generation unit 32 uses a machine learning algorithm to calculate the vertical diameter × horizontal diameter (cm) at harvest time using a combination of the post-flowering accumulated temperature in FIG. 21(a), the fruit and vegetable weight (g) corresponding to the post-flowering accumulated temperature, and the maximum fruit thickening rate as explanatory variables. 2 ) as the objective variable. The prediction unit 36 ​​then inputs the information received from the information acquisition unit 34 (the combination of the post-flowering accumulated temperature and the fruit weight of each fruit, and the maximum fruit thickening rate) into the prediction model (FIG. 21(b)) to predict the size (vertical diameter x horizontal diameter) of each fruit harvested in the facility being predicted. Note that the maximum fruit thickening rate in FIG. 21(a) is calculated from the same equations as the above equations (1) and (2) using the fruit weight (g) corresponding to each post-flowering accumulated temperature, and is therefore expressed in units of g / °C. In this way, the size of each fruit at harvest can be predicted, as in the second embodiment.

[0075] In addition, with regard to the prediction model of Figure 21(b), the prediction results (prediction data) by the prediction model may be fruit weight (g) at harvest time or shipping standards, as shown in Figures 22(a) and 22(b).

[0076] Third embodiment The third embodiment will be described below with reference to FIGS. 23(a) and 23(b).

[0077] In the third embodiment, the learning data processing unit 31 processes the past data of FIG. 5 in the same manner as in the second embodiment, so that the learning data DB 50 stores learning data as shown in FIG. 23(a). Then, the prediction model generation unit 32 uses a machine learning algorithm to calculate the maximum fruit enlargement speed (cm 2 / ℃), the accumulated temperature after flowering when the maximum fruit enlargement rate was reached (℃), and the vertical diameter × horizontal diameter of the fruit when the maximum fruit enlargement rate was reached (cm 2 ) and are the explanatory variables, and the vertical diameter x horizontal diameter at harvest (cm 2 A prediction model (see FIG. 23(b)) is generated using the objective variable, as shown in FIG. 23(a). In the case of the fruit with the fruit ID "A1-1-1" in FIG. 23(a), the maximum fruit enlargement speed (cm 2 / ℃) is 0.107(cm 2 / ℃), the accumulated temperature after flowering when the fruit reached the maximum enlargement rate (℃) was 500 (℃), and the vertical diameter × horizontal diameter (cm 2 ) is 31.4(cm 2 ), so we use this information as explanatory variables.

[0078] Here, from Figure 19 and Figure 7, it can be seen that the fresh weight of the fruit at harvest is likely to differ depending on the time when the fruit enlargement rate reaches its maximum value (maximum fruit enlargement rate). Therefore, as described above, if a prediction model is used in which the accumulated temperature after flowering when the fruit enlargement rate reaches its maximum value is used as an explanatory variable, the vertical diameter x horizontal diameter (cm 2 ) can be predicted with greater accuracy.

[0079] In the third embodiment, the information acquisition unit 34 acquires data similar to that shown in FIG. 8(a) as input data. The prediction unit 36 ​​processes the input data in the same manner as the learning data DB 50 to predict the maximum fruit thickening rate, the accumulated temperature after flowering when the maximum fruit thickening rate is reached, and the vertical diameter × horizontal diameter (cm) of the fruit when the maximum fruit thickening rate is reached. 2 ) is calculated and input into the prediction model. This allows the vertical diameter x horizontal diameter (cm 2 ) to predict.

[0080] According to the third embodiment, the learning data includes the maximum fruit enlargement rate, the accumulated temperature after flowering when the maximum fruit enlargement rate is reached, and the vertical diameter × horizontal diameter (cm 2 ), and the prediction model generation unit 32 calculates the maximum fruit enlargement rate, the accumulated temperature after flowering and the vertical diameter × horizontal diameter (cm 2 ), and a prediction model is generated to predict the vertical diameter x horizontal diameter at harvest based on the learning data including the maximum fruit enlargement rate of the fruit being cultivated, the accumulated temperature after flowering and the vertical diameter x horizontal diameter (cm 2 ) into the prediction model to predict the vertical x horizontal diameter at harvest. This allows the maximum fruit enlargement rate and the accumulated temperature after flowering when the maximum fruit enlargement rate is reached to be taken into account when predicting the vertical x horizontal diameter at harvest of the fruit under cultivation, improving the prediction accuracy.

[0081] In the above third embodiment, the explanatory variables of the learning data and the input data include the post-anthesis accumulated temperature when the maximum fruit thickening speed is reached and the fruit dimensions (vertical diameter x horizontal diameter) when the maximum fruit thickening speed is reached, but this is not limited to the above. The explanatory variables of the learning data and the input data only need to include at least one of the post-anthesis accumulated temperature when the maximum fruit thickening speed is reached and the fruit dimensions (vertical diameter x horizontal diameter) when the maximum fruit thickening speed is reached.

[0082] In the third embodiment, as shown in FIG. 23(b), the vertical diameter x horizontal diameter (cm) of the fruit when it reaches the maximum fruit growth rate is calculated. 2) was used as the learning data and explanatory variables, but instead, the fruit and vegetable weight (g) at the time when the maximum fruit enlargement rate was reached, as shown in Figure 24(a), may be used as the learning data and explanatory variables (see Figure 24(b)).

[0083] In the third embodiment, as in the first and second embodiments, the prediction result (prediction data) by the prediction model may be the fruit weight (g) at harvest time or the shipping standard.

[0084] Fourth embodiment In the second embodiment, the learning data is, as shown in FIG. 18(a), a vertical diameter × horizontal diameter (cm 2 ) and maximum fruit growth rate (cm 2 The prediction model includes the cumulative temperature after flowering and the vertical diameter x horizontal diameter (cm 2 ) and the maximum fruit growth rate (cm 2 / ℃) and are the explanatory variables, and the vertical diameter × horizontal diameter at harvest (cm 2 ) as the objective variable.

[0085] In contrast to this, in the fourth embodiment, the average temperature (°C) in the cultivation environment from after flowering to the time of prediction is added as learning data as shown in Fig. 25(a), and the average temperature (°C) in the cultivation environment from after flowering to the time of prediction is added to the explanatory variables of the prediction model as shown in Fig. 25(b). Note that the "time of prediction" is determined in advance, such as when the accumulated temperature after flowering is 500°C or 800°C.

[0086] Here, if the average temperature (℃) in the cultivation environment from after flowering until the prediction time is high, the number of days (number of days from flowering) required to reach the accumulated temperature (e.g., 1400℃) indicating that the harvest time has arrived will be shorter. In this case, it is known that the size (dimensions and weight) of the fruit and vegetables that can be harvested will be smaller. Also, if the average temperature (℃) in the cultivation environment from after flowering until the prediction time is low, the number of days required to reach the accumulated temperature (e.g., 1400℃) indicating that the harvest time has arrived will be longer. In this case, it is known that the size (dimensions and weight) of the fruit and vegetables that can be harvested will be larger. As such, since there is a correlation between the average temperature (℃) in the cultivation environment from after flowering until the prediction time and the size of the fruit and vegetables at harvest, by using a prediction model such as that in Figure 25(b), it is possible to predict the vertical diameter x horizontal diameter (cm) at harvest more accurately than with the prediction model in Figure 18(b). 2 ) can be predicted with high accuracy.

[0087] As described above, according to the fourth embodiment, the learning data includes the average temperature (°C) in the cultivation environment from after flowering to the prediction time, and the prediction model generation unit 32 generates a prediction model for predicting the vertical diameter x horizontal diameter at harvest based on the learning data including the average temperature (°C) in the cultivation environment from after flowering to the prediction time. The prediction unit 36 ​​predicts the vertical diameter x horizontal diameter at harvest by inputting a combination of the accumulated temperature after flowering of the fruit under cultivation and the size of the fruit at that time (vertical diameter x horizontal diameter), the maximum fruit enlargement rate, and the average temperature (°C) in the cultivation environment from after flowering to the prediction time into the prediction model. Since the vertical diameter x horizontal diameter at harvest of the fruit is considered to be correlated with the average temperature (°C) in the cultivation environment from after flowering to the prediction time, the above-mentioned prediction model can be used to accurately predict the vertical diameter x horizontal diameter at harvest of the fruit under cultivation.

[0088] In the above fourth embodiment, the average temperature (°C) in the cultivation environment from after flowering to the time of prediction is added to the learning data and explanatory variables of the second embodiment, but the present invention is not limited to this. For example, the average temperature (°C) in the cultivation environment from after flowering to the time of prediction may be added to the learning data and explanatory variables of the first embodiment (FIG. 6(b)) and its modified examples (FIGS. 13(b), 14(a), 14(b), 17(a), and 17(c)) (see FIG. 26(a) and FIG. 26(b)). In FIG. 26(a) and FIG. 26(b), the objective variable may be any one of the vertical diameter x horizontal diameter at harvest, the fruit weight at harvest, and the shipping standard. In addition, the average temperature (°C) in the cultivation environment from after flowering to the prediction time may be added to the learning data and explanatory variables of the modified example of the second embodiment (Figures 20(a), 20(b), 21(b), 22(a), and 22(b)) (see Figures 27(a) and 27(b)). In Figure 27(a), the objective variable can be either the fruit weight at harvest or the shipping standard. In Figure 27(b), the objective variable can be either the vertical diameter x horizontal diameter at harvest, the fruit weight at harvest, or the shipping standard.

[0089] Furthermore, in the fourth embodiment, the average temperature (°C) in the cultivation environment from after flowering to the time of prediction may be added to the learning data and explanatory variables in the third embodiment. That is, the prediction model generation unit 32 uses a machine learning algorithm to calculate the maximum fruit enlargement speed (cm 2 / ℃), the accumulated temperature after flowering when the maximum fruit enlargement rate was reached (℃), and the vertical diameter × horizontal diameter of the fruit when the maximum fruit enlargement rate was reached (cm 2 ) and the average temperature (℃) in the cultivation environment from flowering to the prediction time were used as explanatory variables, and the vertical diameter × horizontal diameter (cm 2 A prediction model (see FIG. 28(b)) may be generated using the target variable, as shown by the dashed line in FIG. 28(b). Note that the target variable of the prediction model may be either the harvest fruit weight or the shipping specification.

[0090] In addition, as shown in Figure 29, the vertical diameter × horizontal diameter (cm) of the fruit when it reaches the maximum fruit growth rate in Figure 28(b)2 ), the weight (g) of the fruit or vegetable when it reaches the maximum fruit growth rate may be used as learning data or an explanatory variable. As shown by the dashed line in FIG. 29(b), the objective variable of the prediction model in FIG. 29(b) may be either the fruit weight at harvest or the shipping standard.

[0091] In the above first to fourth embodiments and modified examples, the prediction model is a prediction model (artificial intelligence model) that uses a machine learning algorithm, but the present invention is not limited to this. The prediction model may be, for example, a mathematical model.

[0092] In the above first to fourth embodiments and modified examples, the case where "vertical diameter x horizontal diameter" is used as the dimension of each fruit to be input into the prediction model has been described, but this is not limiting, and other dimensions (vertical diameter only, horizontal diameter only, area, perimeter length, etc.) can also be used as the dimension of each fruit. Furthermore, when there is a relationship as shown in Fig. 11(a) between a dimension other than "vertical diameter x horizontal diameter" (such as horizontal diameter) and fruit weight, the fruit weight (g) can be estimated (converted) from the dimension other than "vertical diameter x horizontal diameter" (such as horizontal diameter) based on the relationship, and used as learning data or explanatory variables.

[0093] In the above first to fourth embodiments and modified examples, the case has been described in which the server 10 has the function of the learning data processing unit 31, but the function of the learning data processing unit 31 may be included in the user terminal 70. In this case, the learning data (past data) is processed in the user terminal 70, and the processed data is transmitted to the server 10.

[0094] The above processing functions can be realized by a computer. In that case, a program is provided that describes the processing contents of the functions that the processing device should have. The above processing functions are realized on the computer by executing the program on the computer. The program that describes the processing contents can be recorded on a computer-readable storage medium (excluding carrier waves).

[0095] When a program is distributed, it is sold in the form of a portable storage medium on which the program is recorded, such as a DVD (Digital Versatile Disc), a CD-ROM (Compact Disc Read Only Memory), etc. The program can also be stored in a storage device of a server computer and transferred from the server computer to other computers via a network.

[0096] A computer that executes a program stores, for example, a program recorded on a portable storage medium or a program transferred from a server computer in its own storage device. The computer then reads the program from its own storage device and executes processing in accordance with the program. The computer can also read the program directly from a portable storage medium and execute processing in accordance with that program. The computer can also execute processing in accordance with the received program each time a program is transferred from the server computer.

[0097] The above-described embodiment is a preferred example of the present invention, but the present invention is not limited to this embodiment and can be modified in various ways without departing from the scope of the present invention. [Explanation of symbols]

[0098] 10 Server 30 Learning Data Collection Unit 31 Learning Data Processing Department 32 Prediction model generation unit 34 Information Acquisition Department 36 Prediction Department 38 Output section 70 User terminal 72 Temperature Sensor 74 Camera 100 Agriculture System

Claims

1. A method for predicting fruit and vegetable information for estimating the size or specifications of fruit and vegetable at the time of harvest, comprising: Acquire a plurality of pieces of learning data including the size of the fruit vegetable measured during cultivation, the accumulated temperature from flowering at the time when the size was measured, and the size or specifications of the fruit vegetable when harvested; A prediction model is generated using a plurality of the learning data to predict the size or specifications of the fruit or vegetable at the time of harvest; inputting input data including the size of the fruit vegetable to be predicted measured during cultivation of the fruit vegetable to be predicted and the accumulated temperature from flowering at the time when the size was measured into the prediction model, thereby predicting the size or specifications of the fruit vegetable to be predicted at the time of harvest; A method for predicting fruit and vegetable information, characterized in that processing is executed by a computer.

2. Each of the learning data includes the size of the fruit vegetable measured at a plurality of time points and the accumulated temperature from flowering at each time point; In the process of generating the prediction model, a maximum value of the thickening rate of the fruit vegetable obtained from each of the learning data is used to generate the prediction model; 2. The method for predicting information on fruit and vegetable according to claim 1, wherein the input data includes a maximum value of a growth rate of the fruit and vegetable during cultivation of the fruit and vegetable to be predicted.

3. The learning data includes an average temperature of the cultivation environment from after the fruit vegetable blooms to a predetermined time, 3. The fruit and vegetable information prediction method according to claim 1, wherein the input data includes an average temperature of a cultivation environment from after flowering of the fruit and vegetable to be predicted to the specified time.

4. The fruit and vegetable information prediction method described in claim 1, characterized in that the input data includes data measured during the period from when the size of the fruit or vegetable to be predicted becomes measurable to a specified number of days before the scheduled shipping date of the fruit or vegetable to be predicted.

5. The fruit and vegetable information prediction method described in claim 1, characterized in that the input data includes data measured during the second, third, and fourth periods, out of five equal periods divided into a first period, a second period, a third period, a fourth period, and a fifth period from the flowering of the fruit and vegetable to be predicted to the scheduled shipping date.

6. The fruit and vegetable information prediction method according to claim 5, wherein the accumulated temperatures in the first to fifth periods are the same or approximately the same.

7. A method for predicting fruit and vegetable information for estimating the size or specifications of fruit and vegetable at the time of harvest, comprising: Obtaining a plurality of pieces of learning data including a maximum value of a speed at which the size of the fruit vegetable is enlarged during cultivation, an accumulated temperature from flowering at the time when the maximum value was obtained and / or a size of the fruit vegetable at the time when the maximum value was obtained, and a size or a standard of the fruit vegetable when harvested; A prediction model is generated using a plurality of the learning data to predict the size or specifications of the fruit or vegetable at the time of harvest; input data including a maximum value of the enlargement rate of the size of the fruit vegetable to be predicted during cultivation of the fruit vegetable to be predicted, and an accumulated temperature from flowering at the time when the maximum value was obtained and / or the size of the fruit vegetable to be predicted at the time when the maximum value was obtained, into the prediction model, thereby predicting the size or specifications of the fruit vegetable to be predicted at the time of harvest; A method for predicting fruit and vegetable information, characterized in that processing is executed by a computer.

8. The learning data includes an average temperature of the cultivation environment from after the fruit vegetable blooms to a predetermined time, 8. The fruit and vegetable information prediction method according to claim 7, wherein the input data includes an average temperature of the cultivation environment from after the fruit and vegetable to be predicted blooms to the specified time.

9. The fruit and vegetable information prediction method according to claim 1 or 7, characterized in that the prediction model is a regression model or a classification model based on machine learning.

10. The method for predicting fruit and vegetable information according to claim 1 or 7, wherein the standard is based on weight.

11. A fruit and vegetable information prediction program that estimates the size or specifications of fruit and vegetables at the time of harvest, Acquire a plurality of pieces of learning data including the size of the fruit vegetable measured during cultivation, the accumulated temperature from flowering at the time when the size was measured, and the size or specifications of the fruit vegetable when harvested; A prediction model is generated using a plurality of the learning data to predict the size or specifications of the fruit or vegetable at the time of harvest; inputting input data including the size of the fruit vegetable to be predicted measured during cultivation of the fruit vegetable to be predicted and the accumulated temperature from flowering at the time when the size was measured into the prediction model, thereby predicting the size or specifications of the fruit vegetable to be predicted at the time of harvest; A fruit and vegetable information prediction program that causes a computer to execute processing.

12. A fruit and vegetable information prediction program that estimates the size or specifications of fruit and vegetables at the time of harvest, Obtaining a plurality of pieces of learning data including a maximum value of a speed at which the size of the fruit vegetable is enlarged during cultivation, an accumulated temperature from flowering at the time when the maximum value was obtained and / or a size of the fruit vegetable at the time when the maximum value was obtained, and a size or a standard of the fruit vegetable when harvested; A prediction model is generated using a plurality of the learning data to predict the size or specifications of the fruit or vegetable at the time of harvest; input data including a maximum value of the enlargement rate of the size of the fruit vegetable to be predicted during cultivation of the fruit vegetable to be predicted, and an accumulated temperature from flowering at the time when the maximum value was obtained and / or the size of the fruit vegetable to be predicted at the time when the maximum value was obtained, into the prediction model, thereby predicting the size or specifications of the fruit vegetable to be predicted at the time of harvest; A fruit and vegetable information prediction program that causes a computer to execute processing.

13. The learning data includes an average temperature of the cultivation environment from after the fruit vegetable blooms to a predetermined time, 13. The fruit and vegetable information prediction program according to claim 11 or 12, wherein the input data includes an average temperature of a cultivation environment from after flowering of the fruit and vegetable to be predicted to the specified time.

Images