Plant cultivation management device, plant cultivation management method, and plant cultivation management program

The plant cultivation management system uses machine learning to stabilize strawberry yields by optimizing environmental conditions, addressing the instability in year-round strawberry production and improving yield consistency and profitability.

WO2024232132A1PCT designated stage expired Publication Date: 2024-11-14MITSUBISHI HEAVY IND LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
PCT/JP2024/003887
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-10
Filing Date
2024-02-06
Publication Date
2024-11-14

AI Technical Summary

Technical Problem

Year-round strawberry cultivation faces instability in production levels due to variations in environmental conditions, leading to inconsistent yields and challenges in securing customer demand and maintaining producer profits.

Method used

A plant cultivation management system utilizing machine learning to analyze environmental data and fruit production data, generating learning models that predict and adjust cultivation conditions such as flower thinning, temperature, sunlight, and red light irradiation to stabilize strawberry yields throughout the year.

Benefits of technology

The system enables stable and consistent strawberry production by optimizing cultivation management, enhancing yield predictability and profitability by adjusting environmental conditions based on machine-learned models, ensuring higher yields during peak periods and reducing fluctuations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2024003887_14112024_PF_FP_ABST
    Figure JP2024003887_14112024_PF_FP_ABST
Patent Text Reader

Abstract

This plant cultivation management device comprises: an acquisition unit that acquires an estimation result of inverse estimation when desired fruit data is input into a learning model having performed machine learning in which used as an explanatory variable is environment data indicating the cultivation state of a plant when cultivating the plant and harvesting a fruit, and used as a target variable is fruit data relating to the fruit to be harvested; and a presentation unit that presents environment data corresponding to the desired fruit data, on the basis of the estimation result.
Need to check novelty before this filing date? Find Prior Art

Description

Plant cultivation management device, plant cultivation management method, and plant cultivation management program

[0001] The present disclosure relates to a plant cultivation management device, a plant cultivation management method, and a plant cultivation management program.

[0002] There are known strawberry production techniques that enable continuous fruit production through year-round cultivation. When producing strawberries through year-round cultivation, if proper cultivation management is not performed in terms of flower thinning, temperature, sunshine hours, etc., variations in production volume occur, for example, depending on the production month. To ensure efficient shipping, there is known a technique for predicting plant growth based on the plant's cultivation status (see, for example, Patent Document 1).

[0003] Japanese Patent Application Laid-Open No. 2019-170217

[0004] The technology in Patent Document 1 makes it possible to predict plant growth, but it is difficult to predict year-round production fluctuations. Unstable production leads to fluctuations in shipping volume, making it difficult for producers to secure sales outlets and for consumers to secure sources of supply. This can have a significant impact on producers' profits. Strawberries, which are grown year-round as described above, are required to ensure a stable production volume, for example, from month to month. Therefore, technology for appropriate cultivation management is required.

[0005] The present disclosure has been made in consideration of the above, and aims to provide a plant cultivation management device, a plant cultivation management method, and a plant cultivation management program that are capable of performing appropriate cultivation management when cultivating plants and harvesting their fruits.

[0006] The plant cultivation management device of the present disclosure includes a learning unit that generates a learning model through machine learning using environmental data indicating the cultivation conditions of the plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data regarding the harvested fruit as a target variable, an acquisition unit that acquires an estimation result by inverse estimation when desired fruit data is input into the generated learning model, and a presentation unit that presents the environmental data corresponding to the desired fruit data based on the estimation result.

[0007] The plant cultivation management method disclosed herein includes the steps of generating a learning model through machine learning using environmental data indicating the cultivation conditions of a plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data regarding the harvested fruit as a target variable, obtaining an estimation result by inverse estimation when desired fruit data is input into the generated learning model, and presenting the environmental data corresponding to the desired fruit data based on the estimation result.

[0008] The plant cultivation management program of the present disclosure causes a computer to perform the following processes: generating a learning model through machine learning using environmental data indicating the cultivation conditions of a plant when the plant is cultivated and the fruit is harvested as an explanatory variable, and fruit data regarding the harvested fruit as a target variable; obtaining an estimation result by inverse estimation when desired fruit data is input into the generated learning model; and presenting the environmental data corresponding to the desired fruit data based on the estimation result.

[0009] According to the present disclosure, appropriate cultivation management can be carried out when cultivating plants and harvesting fruits.

[0010] FIG. 1 is a diagram schematically illustrating an example of a plant cultivation facility according to this embodiment. FIG. 2 is a functional block diagram schematically illustrating an example of a plant cultivation management device. FIG. 3 is a conceptual schematic diagram of a learning model M. FIG. 4 is a flowchart illustrating an example of a plant cultivation management method. FIG. 5 is a diagram schematically illustrating the relationship between the cultivation period and yield of strawberries grown year-round. FIG. 6 is a diagram schematically illustrating an example of the relationship between the cultivation period and yield of strawberries grown year-round when flower thinning management is performed. FIG. 7 is a diagram schematically illustrating another example of the relationship between the cultivation period and yield of strawberries grown year-round when flower thinning management is performed. FIG. 8 is a diagram schematically illustrating an example of the relationship between the cultivation period and yield of strawberries grown year-round when temperature management is performed. FIG. 9 is a diagram schematically illustrating another example of the relationship between the cultivation period and yield of strawberries grown year-round when temperature management is performed. FIG. 10 is a diagram schematically illustrating an example of the relationship between the cultivation period and yield of strawberries grown year-round when photoperiod management is performed. Fig. 11 is a diagram schematically showing another example of the relationship between the cultivation period and yield of strawberries cultivated year-round when photoperiod is controlled. Fig. 12 is a diagram schematically showing an example of the relationship between the cultivation period and yield of strawberries cultivated year-round when red light irradiation time is controlled. Fig. 13 is a diagram schematically showing another example of the relationship between the cultivation period and yield of strawberries cultivated year-round when red light irradiation time is controlled.

[0011] Hereinafter, embodiments of a plant cultivation management device according to the present disclosure will be described with reference to the drawings. Note that the present invention is not limited to these embodiments. Furthermore, the components in the following embodiments include those that are easily replaceable by those skilled in the art, or those that are substantially identical.

[0012] FIG. 1 is a diagram schematically illustrating an example of a plant cultivation facility 1 according to this embodiment. The plant cultivation facility 1 is a facility for cultivating plants and harvesting fruit. In this embodiment, strawberries will be used as an example of fruit. Note that the plant cultivation facility 1 may also be used to cultivate other fruits, such as mandarin oranges. As shown in FIG. 1 , the plant cultivation facility 1 includes a greenhouse 101, electric lamps 102, red LED lamps 103, an environmental adjustment device 104, and a plant cultivation management device 100.

[0013] The greenhouse 101 has a greenhouse main body 101a and a soil portion 101b. The greenhouse main body 101a is made of a translucent material that allows sunlight to pass through, and a space for accommodating plants is formed inside. The soil portion 101b is laid below the greenhouse main body 101a. The soil portion 101b is made of soil or sand containing ingredients suitable for plant growth. The plants grow upward from this soil portion 101b, exposed to sunlight that has passed through the ceiling and walls of the greenhouse main body 101a.

[0014] An electric lamp 102 and a red LED lamp 103 are provided inside the greenhouse 101. The electric lamp 102 has an output of several watts / m 2 The red LED lamps 102 and 103 are lighting devices (normal lighting) that emit weak light of about 100 Hz, and specifically, incandescent bulbs or the like are preferably used. The red LED lamps 103 emit visible light (red light) with a wavelength of 600 to 700 nm. The red LED lamps 103 each emit light from above onto the plants in the greenhouse 101. The red LED lamps 103 output the irradiation time of red light to the plant cultivation management device 100. The electric lamps 102 and red LED lamps 103 both emit light from above onto the plants in the greenhouse 101.

[0015] The environmental adjustment device 104 adjusts the environment inside the greenhouse 101. The environmental adjustment device 104 includes, for example, a temperature adjustment mechanism that adjusts the temperature inside the greenhouse 101, a humidity adjustment mechanism that adjusts the humidity inside the greenhouse 101, a carbon dioxide concentration adjustment mechanism that adjusts the carbon dioxide concentration inside the greenhouse 101, and the like.

[0016] An information acquisition unit 105 that acquires information about the cultivation status is provided inside the greenhouse 101. The information acquisition unit 105 includes, for example, a temperature sensor 105a that detects the temperature inside the greenhouse 101, a photoperiod measurement sensor 105b that detects the duration of sunlight on the plants, a carbon dioxide concentration sensor 105c that detects the carbon dioxide concentration inside the greenhouse 101, a humidity sensor 105d that detects the humidity inside the greenhouse 101, and a flower size measurement system 105e that measures the size of the flowers on the plants. The temperature sensor 105a, the photoperiod measurement sensor 105b, the carbon dioxide concentration sensor 105c, and the humidity sensor 105d each output their detection results to the plant cultivation management device 100. The flower size measurement system 105e also captures images of the inside of the greenhouse 101 with, for example, a camera, performs image processing on the captured images to detect flowers, calculates the size of the detected flowers, and outputs the calculation result to the plant cultivation management device 100.

[0017] 2 is a functional block diagram schematically illustrating an example of the plant cultivation management device 100. As shown in FIG. 2, the plant cultivation management device 100 is a computer, and includes a processing unit 10 and a storage unit 20.

[0018] The processing unit 10 performs various types of information processing and includes a processor such as a CPU (Central Processing Unit) and memories such as a ROM (Read Only Memory) and a RAM (Random Access Memory).

[0019] The processing unit 10 has a learning unit 11, an acquisition unit 12, a presentation unit 13, and an environment control unit 14. The learning unit 11 trains an untrained learning model to generate the above-mentioned learning model M that has undergone machine learning, and stores the generated learning model M in the storage unit 20. In this embodiment, the learning unit 11 generates the learning model M through machine learning using environmental data indicating the cultivation conditions of the plants when the plants are cultivated and the fruits are harvested as explanatory variables and fruit data related to the harvested fruits as objective variables.

[0020] The acquisition unit 12 acquires an estimation result by inverse estimation when desired fruit data is input to the generated learning model M. The learning model M will be described later.

[0021] The presentation unit 13 presents the environmental data corresponding to the desired fruit data based on the estimation results of the learning model M.

[0022] The environmental control unit 14 controls the operation of the environmental adjustment device 104 to control the environment inside the greenhouse 101 .

[0023] The storage unit 20 stores information such as various programs and data, etc. The storage unit 20 includes storage such as a hard disk drive (HDD) and a solid state drive (SSD).

[0024] The storage unit 20 stores, for example, various programs, data, etc. for performing processing in the processing unit 10. The storage unit 20 also stores the environmental data, fruit data, and learning model M described above.

[0025] The memory unit 20 stores a plant cultivation management program that causes a computer to perform the following processes: obtaining an estimation result by inverse estimation when desired fruit data is input into a learning model M that has been machine-learned using environmental data indicating the cultivation conditions of the plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data regarding the harvested fruit as a target variable; and presenting environmental data corresponding to the desired fruit data based on the estimation result.

[0026] In the plant cultivation management device 100, the processor in the processing unit 10 reads out various programs and loads them in memory, thereby executing information processing corresponding to the functions of each of the above units. Examples of the various programs include a program received by a communication unit (not shown), a program stored in the storage unit 20, and a program recorded on an external recording medium. The plant cultivation management device 100 functions as an information processing device (computer) that executes various information processes. Note that the various programs may be executed by an information processing device other than the plant cultivation management device 100, or the plant cultivation management device 100 and the other information processing device may cooperate to execute the various programs.

[0027] FIG. 3 is a conceptual schematic diagram of a learning model M. The learning model M refers to a learning model in AI (Artificial Intelligence). Specifically, the learning model M is a learning model trained by deep learning, and is composed of variables and a model (neural network configuration information) that defines a neural network that constitutes a classifier trained by deep learning. The learning model M can compare input data with trained data to determine the degree of similarity, for example, the percentage of similarity. Deep learning is one type of machine learning technique, and in a narrow sense, is composed of, for example, a neural network with four or more layers. In the example of this embodiment, the learning model M is a CNN (Conventional Neural Network) model. As shown in FIG. 3, the learning model M includes an intermediate layer MN including multiple convolutional layers and multiple pooling layers, and a final layer FN including a fully connected layer. When input data IN is input, the learning model M performs calculations in the intermediate layer MN and the final layer FN and outputs output data ON. However, the learning model M is not limited to a CNN model and may be a learning model of any type.

[0028] The learning model M in this embodiment is a learning model in which environmental data indicating the cultivation conditions of plants when they are cultivated and the fruits are harvested is used as an explanatory variable, and fruit data regarding the harvested fruits is used as a target variable to machine-learn the correspondence.

[0029] The environmental data includes various data such as inflorescence data indicating the number of inflorescences of the plant, temperature data indicating the temperature around the plant, day length data indicating the hours of sunlight per unit period for the plant, red light data indicating the hours of red light irradiation per unit period for the plant, carbon dioxide concentration data indicating the carbon dioxide concentration around the plant, humidity data indicating the humidity around the plant, fertilizer data indicating the amount of fertilizer applied, etc. The environmental data may also include other conditions, such as nutrients in the soil portion 101b.

[0030] The learning unit 11 generates environmental data based on the irradiation time of the red LED lamp 103, the detection results of the temperature sensor 105a, the day length measurement sensor 105b, the carbon dioxide concentration sensor 105c and the humidity sensor 105d, the calculation results of the flower size measurement system 105e, the amount of fertilizer applied to the plants, etc.

[0031] The fruit data includes yield data indicating the fruit yield and quality data indicating the quality of the fruit. The yield data includes, for example, data specifying the fruit harvest amount in weight. The yield data includes data indicating the harvest amount for a predetermined period (e.g., one year) and data indicating the harvest amount for each unit period (e.g., one month) within the predetermined period. The quality data includes, for example, data indicating the grade of each fruit (appearance such as shape and color, degree of pest infestation, ripeness, etc.) and data indicating the size (weight, etc.). The quality data can be associated with the harvest time of the fruit.

[0032] The learning unit 11 generates fruit data (yield data and quality data) based on data input from outside regarding fruit yield and quality.

[0033] The learning unit 11 can generate the learning model M by setting weighting coefficients and bias values ​​in the learning model using the generated environmental data and fruit data as training data. In this case, the learning model M can be said to be a learning model in which the weighting coefficients and bias values ​​have been learned by the learning unit 11.

[0034] The generated learning model M is a learning model for estimating (forward estimation) fruit data based on environmental data. Specifically, when environmental data is input as input data IN, the learning model M outputs fruit data as output data ON.

[0035] The learning model M according to this embodiment can estimate (inversely estimate) environmental data from fruit data based on the correspondence between the environmental data and fruit data. Specifically, when fruit data is input as input data IN, the learning model M can output environmental data as output data ON.

[0036] Next, a plant cultivation management method using the plant cultivation management device 100 configured as described above will be described. Fig. 4 is a flowchart showing an example of the plant cultivation management method. As shown in Fig. 4, the plant cultivation management method includes a learning model generation step S10, an estimation result acquisition step S20, a presentation step S30, and a management step S40.

[0037] In the learning model generation step S10, the learning unit 11 generates a learning model M by machine learning the correspondence relationship using environmental data indicating the cultivation conditions of the plants when they are cultivated and the fruits are harvested as explanatory variables and fruit data regarding the harvested fruits as objective variables.

[0038] In this plant cultivation method, plants are cultivated while adjusting the environment inside the greenhouse 101 using the environmental adjustment device 104. The plant growth period is divided into days, and the following operations can be performed for each day. Specifically, before the photoperiod (i.e., the time period during which sunlight is available), for example, the electric lamp 102 provides illumination for approximately two hours (referred to as the illumination period). During this illumination period, only the electric lamp 102 is turned on, and the red LED lamp 103 is turned off. After the photoperiod has elapsed, the electric lamp 102 is turned on again, and after a specified time (e.g., approximately two hours) has elapsed, the electric lamp 102 is turned off, and only the red LED lamp 103 is turned on. After the red LED lamp 103 has been turned on for approximately one to two hours, the red LED lamp 103 is turned off. The red LED lamp 103 outputs the irradiation time of red light to the plant cultivation management device 100. By repeating the above steps daily during the growth period, plants can be cultivated.

[0039] During the plant cultivation period, the information acquisition unit 105 detects the cultivation status using the temperature sensor 105a, the day length measurement sensor 105b, the carbon dioxide concentration sensor 105c, and the humidity sensor 105d. The temperature sensor 105a, the day length measurement sensor 105b, the carbon dioxide concentration sensor 105c, and the humidity sensor 105d output the detection results to the plant cultivation management device 100. Furthermore, when flowers bloom on the plant, the flower size measurement system 105e calculates the size of each flower and outputs the calculation results to the plant cultivation management device 100. Furthermore, the operator inputs the amount of fertilizer applied to the plant to the plant cultivation management device 100.

[0040] In the plant cultivation management device 100, the learning unit 11 generates environmental data based on the irradiation time of the red LED lamp 103, the detection results of the temperature sensor 105a, the day length measurement sensor 105b, the carbon dioxide concentration sensor 105c and the humidity sensor 105d, the calculation results of the flower size measurement system 105e, the amount of fertilizer applied to the plants, etc.

[0041] When a plant bears fruit and the fruit is harvested, an operator associates information on the fruit yield and quality with the harvest date and inputs the information into the plant cultivation management device 100. In the plant cultivation management device 100, the learning unit 11 generates fruit data (yield data and quality data) on the basis of the input data on the fruit yield and quality.

[0042] The learning unit 11 generates a learning model M by machine learning a correspondence relationship between the generated environmental data and fruit data, using the environmental data as an explanatory variable and the fruit data as a target variable. The learning unit 11 stores the generated learning model M in the storage unit 20.

[0043] Next, in an estimation result acquisition step S20, the acquisition unit 12 acquires an estimation result by inverse estimation when desired fruit data is input to the generated learning model M. The desired fruit data may be, for example, yield data indicating the yield of fruit. In this case, data in which a target yield per unit period (e.g., one month) is set so that the yield for each unit period is leveled over a predetermined period (e.g., one year) may be used.

[0044] In addition, data can be used in which the target yield per unit period is set so that the yield in one unit period (e.g., December) over a specified period (e.g., one year) is significantly higher than the yield in other unit periods (January to November).

[0045] In the presentation step S30, the presentation unit 13 presents environmental data corresponding to the desired fruit data based on the estimation result in the learning model M.

[0046] In the management step S40, the plant cultivation management device 100 or the operator manages the cultivation status of the plant based on the presented environmental data. Figure 5 is a diagram schematically illustrating the relationship between the cultivation period and yield of year-round strawberries. As shown in Figure 5, when producing year-round strawberries, if appropriate cultivation management is not performed in terms of flower thinning, temperature, sunshine hours, etc., yield variations will occur, for example, depending on the production month. Appropriate cultivation management is necessary. An example of cultivation status management will be described below. In Figure 5, the actual yield value Qr is shown by a dashed line, and the interpolated value Q obtained by machine learning is shown by a solid line. In the following, the interpolated value Q will be described as the yield value.

[0047] <Example 1: Flower Thinning Management> Figure 6 is a diagram schematically illustrating an example of the relationship between the cultivation period and yield of strawberries grown year-round when flower thinning management is performed. The presentation unit 13 can present inflorescence data as environmental data. The number of inflorescences corresponds to the strawberry fruit yield. In other words, the greater the number of inflorescences, the greater the yield. Conversely, the fewer the number of inflorescences, the lower the yield. As shown in Figure 6, based on the presented inflorescence data, the worker performs tasks such as cluster thinning, flower thinning, and fruit thinning between one week and three weeks before the peak yield date (timings T1, T2, T3, and T4 in Figure 6). This task can suppress the peak yield. The proportion of inflorescences and fruit remaining after cluster thinning, flower thinning, and fruit thinning can be, for example, 75% or more of the total.

[0048] The effect of suppressing the peak yield increases in the order of cluster thinning, flower thinning, and fruit thinning. It is preferable to carry out the work early relative to the peak yield (for example, three weeks before the peak). If it is done too early relative to the peak yield (for example, four weeks before the peak), the clusters will be too small to carry out the work.

[0049] It is also known that in strawberry cultivation, the larger the flower size, the larger the fruit size. Therefore, when thinning flowers, by thinning small flowers based on the calculation results of the flower size measurement system 105e, more large flowers can be left. In this case, more large fruits will grow. Since large fruits generally have a higher unit price per weight, performing this operation will increase the farmer's income per unit area.

[0050] Figure 7 is a diagram showing another example of the relationship between the cultivation period and yield of strawberries grown year-round when thinning flowers is performed. As shown in Figure 7, among multiple yield peaks, cluster, flower, and fruit thinning are not performed for a specific peak, while cluster, flower, and fruit thinning are performed for other peaks between one week and three weeks before the peak date (timings T5, T6, and T7 in Figure 6), significantly reducing the proportion of flower clusters and fruits left after cluster, flower, and fruit thinning (for example, approximately 10% of the total). This allows the yield of a specific peak when cluster, flower, and fruit thinning are not performed to be significantly increased. By using this, the yield of one unit period within a specified period can be significantly increased compared to the yield of other unit periods.

[0051] <Example 2: Temperature Management> Figure 8 is a diagram schematically illustrating an example of the relationship between the cultivation period and yield of strawberries grown year-round when temperature is managed. The presentation unit 13 can present temperature data as environmental data. When plants are exposed to high temperatures, for example, during the daytime, the development of inflorescences is suppressed. Utilizing this phenomenon, as shown in Figure 8, the environmental control unit 14 repeatedly switches the daytime temperature between low temperature mode L (e.g., 25°C) and high temperature mode H (e.g., 28°C) every two weeks, for example. Setting the temperature to the low temperature mode promotes the development of inflorescences. Furthermore, setting the temperature to the high temperature mode can suppress the development of inflorescences, thereby suppressing the peak yield.

[0052] FIG. 9 is a diagram schematically illustrating another example of the relationship between the cultivation period and yield of year-round strawberries when temperature is controlled. As shown in FIG. 9, the environmental control unit 14 sets the daytime temperature to high temperature mode H for the period from six weeks to ten weeks before a predetermined peak among multiple yield peaks, thereby suppressing the development of inflorescences. Thereafter, the environmental control unit 14 sets the daytime temperature to low temperature mode L to promote the development of inflorescences. After the inflorescences appear (buds emerge), the environmental control unit 14 sets the temperature to high temperature mode H within two weeks to again suppress the development of inflorescences. As a result, the yield at the predetermined peak can be significantly increased. By using this, the yield of one unit period within a predetermined period can be significantly increased relative to the yield of other unit periods.

[0053] It is also possible to perform the tasks of cluster thinning, flower thinning, and fruit thinning at the same time. When performing cluster thinning, flower thinning, and fruit thinning, by thinning small flowers based on the calculation results of the flower size measurement system 105e, it is possible to leave more large flowers. In this case, more large fruits will grow.

[0054] <Example 3: Photoperiod> FIG. 10 is a diagram schematically illustrating an example of the relationship between the cultivation period and yield of strawberries grown year-round when photoperiod management is performed. The presentation unit 13 can present photoperiod data as environmental data. Photoperiod can be adjusted by turning on the electric lamp 102. Photoperiod may induce flower bud formation in plants. In the case of long-day plants, the longer the day, the more likely they are to flower. Taking advantage of this phenomenon, the environmental control unit 14 repeatedly switches the photoperiod between short-day mode S (e.g., 14 hours) and long-day mode L (e.g., 16 hours) every two weeks, as shown in FIG. 10 . Setting the photoperiod to the short-day mode can suppress inflorescence development and peak yield compared to when the photoperiod is set to the long-day mode. Note that the photoperiod in the long-day mode may be longer than 16 hours, for example, 18 hours or 24 hours.

[0055] FIG. 11 is a diagram schematically illustrating another example of the relationship between the cultivation period and yield of strawberries grown year-round when photoperiod management is performed. As shown in FIG. 11 , the environmental control unit 14 sets the photoperiod to short-period mode for the period from six weeks to ten weeks before a predetermined peak among multiple yield peaks, thereby suppressing the development of inflorescences. Thereafter, the environmental control unit 14 sets the photoperiod to long-period mode to promote the development of inflorescences. After the inflorescences appear (buds emerge), the environmental control unit 14 sets the photoperiod to short-period mode within two weeks to again suppress the development of inflorescences. As a result, the yield at the predetermined peak can be significantly increased. By using this, the yield of one unit period within a predetermined period can be significantly increased relative to the yield of other unit periods.

[0056] It is also possible to perform the tasks of cluster thinning, flower thinning, and fruit thinning at the same time. When performing cluster thinning, flower thinning, and fruit thinning, by thinning small flowers based on the calculation results of the flower size measurement system 105e, it is possible to leave more large flowers. In this case, more large fruits will grow.

[0057] Example 4: Red Light Irradiation Time FIG. 12 is a diagram illustrating an example of the relationship between the cultivation period and yield of strawberries grown year-round when red light irradiation time is managed. The presentation unit 13 can present red light data as environmental data. Red light from the red LED lamp 103 has the effect of weakening the effect of sunlight. The longer the red light irradiation time, the weaker the effect of sunlight. Taking advantage of this phenomenon, the environmental control unit 14 repeatedly switches the red light between a non-irradiation mode (OFF: no red light irradiation) and an irradiation mode (ON: red light irradiation for 1 to 2 hours) every two weeks, for example, after the end of the photoperiod, as shown in FIG. 12 . By using red light in the irradiation mode, inflorescence development can be suppressed compared to the non-irradiation mode, thereby suppressing the peak yield.

[0058] FIG. 13 is a diagram schematically illustrating another example of the relationship between the cultivation period and yield of strawberries grown year-round when the red light irradiation time is managed. As shown in FIG. 13 , the environmental control unit 14 switches the red light irradiation mode to the red light irradiation mode for the period from 6 weeks to 10 weeks before a predetermined peak among multiple yield peaks, thereby suppressing the development of inflorescences. Thereafter, the environmental control unit 14 switches the red light irradiation mode to the red light non-irradiation mode to promote the development of inflorescences. After the inflorescences appear (buds emerge), the environmental control unit 14 switches the red light irradiation mode to the red light irradiation mode within two weeks to again suppress the development of inflorescences. As a result, the yield at the predetermined peak can be significantly increased. By using this, the yield of one unit period within a predetermined period can be significantly increased compared to the yield of other unit periods.

[0059] It is also possible to perform the tasks of cluster thinning, flower thinning, and fruit thinning at the same time. When performing cluster thinning, flower thinning, and fruit thinning, by thinning small flowers based on the calculation results of the flower size measurement system 105e, it is possible to leave more large flowers. In this case, more large fruits will grow.

[0060] <Example 5: Combination of environmental data> The presentation unit 13 can extract at least one data from inflorescence data, temperature data, day length data, and red light data based on predetermined extraction conditions, and present the extracted data as environmental data.

[0061] The predetermined extraction conditions include, for example, effectively adjusting the yield (optimization mode), reducing costs (cost reduction mode), reducing the number of workers required (personnel cost reduction mode), etc. In this case, data corresponding to the extraction conditions, such as data associating the cultivation period with the incurred costs, data associating the cultivation period with the number of workers, etc., are input to the plant cultivation management device 100 as environmental data.

[0062] The predetermined extraction conditions may be configured to be preset in the presentation unit 13, for example, or may be input together with the fruit data input to the learning model M.

[0063] As described above, according to the first aspect of the present disclosure, there is provided a plant cultivation management device 100 including a learning unit 11 that generates a learning model M through machine learning using environmental data indicating the cultivation conditions of a plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data related to the harvested fruit as a target variable, an acquisition unit 12 that acquires an estimation result by inverse estimation when desired fruit data is input to the generated learning model M, and a presentation unit 13 that presents environmental data corresponding to the desired fruit data based on the estimation result.

[0064] This configuration makes it possible to obtain appropriate environmental data for achieving the desired fruit data, thereby enabling appropriate cultivation management based on the obtained environmental data when cultivating plants and harvesting the fruits. Furthermore, the plant cultivation management device 100 controls the cultivation conditions so as to approach the desired fruit data while operating and managing the plant cultivation facility 1. Because plant growth is strongly influenced by region and location, constructing a learning model M at different cultivation locations makes it possible to propose more reliable environmental data.

[0065] According to a second aspect of the present disclosure, in the plant cultivation management device according to the first aspect, the fruit data includes at least one of yield data indicating a yield of the fruit and quality data indicating a quality of the fruit.

[0066] This configuration makes it possible to obtain appropriate environmental data for achieving at least one of the desired yield and quality.

[0067] According to a third aspect of the present disclosure, in the plant cultivation management device according to the second aspect, the environmental data includes inflorescence data indicating the number of inflorescences of the plant, and the presentation unit 13 presents the inflorescence data as the environmental data.

[0068] According to this configuration, it is possible to obtain appropriate inflorescence data to achieve the desired content of the fruit data.

[0069] According to a fourth aspect of the present disclosure, in the plant cultivation management device of the second or third aspect, the environmental data includes temperature data indicating the temperature around the plant, and the presentation unit 13 presents the temperature data as the environmental data.

[0070] According to this configuration, it is possible to obtain appropriate temperature data to achieve the desired fruit data content.

[0071] According to a fifth aspect of the present disclosure, in a plant cultivation management device relating to any of the second to fourth aspects, the environmental data includes day length data indicating the hours of sunlight per unit period for the plant, and the presentation unit 13 presents the day length data as the environmental data.

[0072] According to this configuration, appropriate day length data for achieving the desired fruit data content can be obtained.

[0073] According to a sixth aspect of the present disclosure, in the plant cultivation management device relating to any of the second to fifth aspects, the environmental data includes red light data indicating the irradiation time of red light to the plant per unit period, and the presentation unit 13 presents the red light data as the environmental data.

[0074] This configuration makes it possible to obtain appropriate red light data to achieve the desired fruit data content.

[0075] According to a seventh aspect of the present disclosure, in a plant cultivation management device relating to any of the second to sixth aspects, the fruit data includes yield data, and the yield data includes data set so that the yield for each unit period within a predetermined period is leveled out, or data set so that the yield for one unit period within the predetermined period is significantly higher than the yield for other unit periods.

[0076] According to this configuration, by properly managing the cultivation conditions, it is possible to standardize the yield for each unit period, or to make the yield for one unit period significantly higher than the yield for other unit periods.

[0077] According to an eighth aspect of the present disclosure, in the plant cultivation management device relating to any of the second to seventh aspects, the environmental data includes inflorescence data indicating the number of inflorescences on the plant, temperature data indicating the temperature around the plant, day length data indicating the number of hours of sunlight per day on the plant, and red light data indicating the number of hours of red light irradiation per day on the plant, and the presentation unit 13 extracts at least one data item from the inflorescence data, temperature data, day length data, and red light data based on predetermined extraction conditions, and presents the extracted data as environmental data.

[0078] According to this configuration, environmental data can be obtained according to predetermined extraction conditions, so that the cultivation conditions can be flexibly managed according to, for example, costs, the number of workers available, etc.

[0079] According to a ninth aspect of the present disclosure, the plant cultivation management device according to the first aspect further includes an environment control unit 14 that controls the environment around the plant based on the environmental data presented by the presentation unit 13.

[0080] According to this configuration, the environment around the plants can be automatically controlled based on the presented environmental data, thereby reducing the number of workers required.

[0081] According to a tenth aspect of the present disclosure, there is provided a plant cultivation management method including a step of obtaining an estimation result by inverse estimation when desired fruit data is input into a learning model M that has been machine-trained using environmental data indicating the cultivation conditions of a plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data regarding the harvested fruit as a target variable, and a step of presenting environmental data corresponding to the desired fruit data based on the estimation result.

[0082] With this configuration, appropriate environmental data can be obtained to achieve the desired fruit data content, so that when cultivating plants and harvesting fruit, appropriate cultivation management can be carried out based on the obtained environmental data.

[0083] According to an eleventh aspect of the present disclosure, there is provided a plant cultivation management program that causes a computer to perform the following processes: obtaining an estimation result by inverse estimation when desired fruit data is input into a learning model M that has been machine-learned using environmental data indicating the cultivation conditions of a plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data regarding the harvested fruit as a target variable; and presenting environmental data corresponding to the desired fruit data based on the estimation result.

[0084] With this configuration, appropriate environmental data can be obtained to achieve the desired fruit data content, so that when cultivating plants and harvesting fruit, appropriate cultivation management can be carried out based on the obtained environmental data.

[0085] The technical scope of the present invention is not limited to the above-described embodiments, and appropriate modifications can be made without departing from the spirit of the present invention. For example, environmental data may include harvest history data indicating the date and time of harvest when harvesting is performed by hand. In addition to the existence of periodic events in the growth of the plant itself, manual harvesting by workers may cause fluctuations in yield. Therefore, by including harvest history data as environmental data and training the learning model M, more accurate learning is possible.

[0086] Furthermore, in the above embodiment, the fruit data is mainly explained using yield data as an example, but the present invention is not limited to this configuration, and similar explanations are possible using other data such as quality data.

[0087] REFERENCE SIGNS LIST 1 Plant cultivation facility 10 Processing unit 11 Learning unit 12 Acquisition unit 13 Presentation unit 14 Environmental control unit 20 Memory unit 100 Plant cultivation management device 101 Greenhouse 101a Greenhouse main body 101b Soil unit 102 Electric lamp 103 Red LED lamp 104 Environmental adjustment device 105 Information acquisition unit 105a Temperature sensor 105b Day length measurement sensor 105c Carbon dioxide concentration sensor 105d Humidity sensor 105e Flower size measurement system M Learning model FN Final layer IN Input data MN Intermediate layer ON Output data

Claims

1. A plant cultivation management device comprising: a learning unit that generates a learning model through machine learning using environmental data indicating the cultivation conditions of a plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data regarding the harvested fruit as a target variable; an acquisition unit that acquires an estimation result by inverse estimation when desired fruit data is input to the generated learning model; and a presentation unit that presents the environmental data corresponding to the desired fruit data based on the estimation result.

2. The plant cultivation management device according to claim 1, wherein the fruit data includes at least one of yield data indicating a yield of the fruit and quality data indicating a quality of the fruit.

3. The plant cultivation management device according to claim 2, wherein the environmental data includes inflorescence data indicating a number of inflorescences of the plant, and the presentation unit presents the inflorescence data as the environmental data.

4. The plant cultivation management device according to claim 2, wherein the environmental data includes temperature data indicating a temperature around the plant, and the presentation unit presents the temperature data as the environmental data.

5. The plant cultivation management device according to claim 2, wherein the environmental data includes day length data indicating the hours of sunlight per unit period for the plant, and the presentation unit presents the day length data as the environmental data.

6. The plant cultivation management device according to claim 2, wherein the environmental data includes red light data indicating a time period during which the plant is irradiated with red light per unit period, and the presentation unit presents the red light data as the environmental data.

7. The plant cultivation management device according to claim 2, wherein the fruit data includes the yield data, and the yield data includes data set so that the yield for each unit period within a specified period is leveled out, or data set so that the yield for one of the unit periods within the specified period is exceptionally high compared to the yields for other unit periods.

8. The plant cultivation management device according to claim 2, wherein the environmental data includes inflorescence data indicating the number of inflorescences of the plant, temperature data indicating the temperature around the plant, day length data indicating the number of hours of sunlight per day on the plant, and red light data indicating the number of hours of red light irradiation per day on the plant, and the presentation unit extracts at least one data item from the inflorescence data, the temperature data, the day length data, and the red light data based on predetermined extraction conditions, and presents the extracted data item as the environmental data.

9. The plant cultivation management device according to claim 1, further comprising an environmental control unit that controls an environment around the plant based on the environmental data presented by the presentation unit.

10. A plant cultivation management method comprising: a step of obtaining an estimation result by inverse estimation when desired fruit data is input into a learning model that has been machine-trained using environmental data indicating the cultivation conditions of the plant when the plant is cultivated and the fruit is harvested as an explanatory variable and fruit data regarding the harvested fruit as a target variable; and a step of presenting the environmental data corresponding to the desired fruit data based on the estimation result.

11. A plant cultivation management program that causes a computer to execute the following processes: a process of obtaining an estimation result by inverse estimation when desired fruit data is input into a machine-learned learning model in which environmental data indicating the cultivation conditions of the plant when the plant is cultivated and the fruit is harvested is used as an explanatory variable, and fruit data regarding the harvested fruit is used as a target variable; and a process of presenting the environmental data corresponding to the desired fruit data based on the estimation result.

Citation Information

Patent Citations

  • Culture support device, culture support method, computer program and recording medium

    JP2019030253A

  • Crop production assisting program, crop production assisting method, and crop production assisting device

    JP2021056572A

  • Crop rating prediction method and crop rating prediction program

    JP2023003705A

  • Yield improvement system, cultivation information processing system, cultivation system, yield improvement method, and yield improvement program

    JP2023094836A