Harvest assistance system, processing device, and computer program

Abstract

A harvest assistance system 1 according to the present disclosure comprises: a first measurement device 2 that measures the quality of fruit cultivated in a field F; and a management server 8. The management server 8 comprises a processing unit 32 that executes: a process for acquiring a plurality of types of quality values on the basis of outputs from the first measurement device 2; a process for acquiring a plurality of pieces of time-series data in which each among the plurality of types of quality values is accumulated in a time series; and a process for determining, on the basis of the plurality of pieces of time-series data, a predicted period that is predicted to be an appropriate time for harvesting the fruit.

Description

Harvest support system, processing device, and computer program

[0001] This application claims priority to Japanese Patent Application No. 2023-215577, filed December 21, 2023, and incorporates by reference all of the contents of that application.

[0002] Patent Document 1 discloses a measuring device that irradiates a measurement object such as a fruit with light and performs spectroscopic analysis of the transmitted light to obtain quality values ​​such as sugar content and acidity of the measurement object.

[0003] Japanese Patent Application Laid-Open No. 2020-101409

[0004] The present disclosure provides a harvest support system that includes a measuring device for measuring the quality of fruit grown in a field, and a processing device. The processing device includes a processing unit that executes a first process for acquiring multiple quality values ​​based on the output of the measuring device, a second process for acquiring multiple time-series data in which the multiple quality values ​​are accumulated in chronological order, and a third process for calculating a predicted optimum harvest period for the fruit based on the multiple time-series data.

[0005] FIG. 1 is a diagram showing an example of the overall configuration of a harvest support system according to an embodiment. FIG. 2 is a block diagram showing essential parts of an agricultural machine. FIG. 3 is a block diagram showing an example of a first measurement device. FIG. 4A is a block diagram showing an example of a second measurement device. FIG. 4B is a cross-sectional view of a second measurement head. FIG. 5 is a block diagram showing an example of the configuration of a management server. FIG. 6 is a flowchart showing an example of fruit quality measurement work in a farm field. FIG. 7 is a plan view of a farm field. FIG. 8 is a flowchart showing an example of quality value acquisition processing. FIG. 9 is a diagram showing an example of a plurality of first calibration curve data. FIG. 10 is a diagram showing an example of a plurality of second calibration curve data. FIG. 11 is a diagram showing an example of scaling data. FIG. 12 is a diagram showing a mode in which a processing unit calculates a measured quality value of a fruit with ID = 1. FIG. 13 is a flowchart showing an example of a prediction processing. FIG. 14 is a diagram showing an example of a quality value database. FIG. 15 is a graph showing an example of a predicted quality value. FIG. 16 is a diagram showing a first example of a sunny / rainy forecast and a predicted precipitation amount included in the environmental prediction information. FIG. 17 is a diagram showing a second example of a sunny / rainy forecast and a predicted precipitation amount included in the environmental prediction information. Fig. 18 is a diagram showing a third example of a sunshine / rain forecast and a forecast precipitation amount included in the environmental forecast information Fig. 19 is a diagram showing an example of a disease risk included in the environmental forecast information.

[0006] [Problem to be Solved by the Present Disclosure] The above-mentioned measuring device can grasp the current quality of fruit, but cannot grasp the future quality state. Therefore, while it is possible to determine whether the current state of fruit is suitable for harvesting, it is not possible to predict the optimum harvest time, such as how long it will take for the fruit to become suitable for harvesting, if the current state of fruit has not yet reached a state suitable for harvesting.

[0007] [Effects of the Present Disclosure] According to the present disclosure, it is possible to predict the optimum harvest time for fruit.

[0008] First, the contents of the embodiment will be listed and explained.

[0009] (1) A harvesting support system according to the present disclosure includes a measuring device that measures the quality of fruit grown in a field, and a processing device. The processing device includes a processing unit that executes a first process of acquiring multiple quality values ​​based on the output of the measuring device, a second process of acquiring multiple time-series data in which the multiple quality values ​​are accumulated in chronological order, and a third process of calculating a predicted period for the optimum harvest time of the fruit based on the multiple time-series data.

[0010] According to the above configuration, multiple types of quality values ​​can be predicted using multiple time-series data, and a predicted period for the optimum harvest time can be calculated based on the multiple predicted quality values. As a result, the predicted period can be output as a prediction result for the optimum harvest time of fruit.

[0011] (2) In the harvest support system of (1), the third process preferably includes: a process of calculating a predicted quality value for each of the plurality of quality values ​​based on the plurality of time-series data; a process of acquiring environmental prediction information indicating an environmental prediction for the field; and a process of calculating the prediction period based on the environmental prediction information and the plurality of predicted quality values. In this case, the prediction period can be calculated by taking into account the environmental prediction information for the field in addition to the predicted quality value of the fruit.

[0012] (3) In the harvest support system of (2), if the environmental prediction information includes a predicted value indicating a predicted environmental state, the process of calculating the predicted period preferably includes a process of calculating a first period during which the plurality of predicted quality values ​​satisfy a predetermined condition, and a process of calculating a second period during which the environmental state is determined to be suitable for harvesting based on a comparison result between the predicted value and a predetermined threshold, and a process of calculating the predicted period based on the first period and the second period. In this case, the predicted period can be appropriately calculated based on the relationship between an appropriate period for the quality value and an appropriate period for the environment.

[0013] (4) In the harvest support system of (3), the process of calculating the predicted period based on the first period and the second period preferably includes a process of determining an overlapping period between the first period and the second period as the predicted period. In this case, a period appropriate for the environment and quality values ​​can be determined as the predicted period.

[0014] (5) Immediately after a period that is not the second period, the field condition may be poor due to rainfall, etc. In this regard, in the harvest support system described above in (4), the process of calculating the prediction period based on the first period and the second period may further include a process of providing an interval of at least one day between the period that is not the second period and the prediction period if a period that is not the second period exists immediately before the overlapping period. This makes it possible to calculate the prediction period while avoiding times when the field condition is poor.

[0015] (6) In the harvest support system of any one of (2) to (5) above, the environmental prediction information may include at least one of meteorological information and disease risk information related to diseases of the fruit.

[0016] (7) In the harvest support system of (4), the process of determining the prediction period based on the first period and the second period may include a process of, when there is no overlap between the first period and the second period, determining the prediction period to be a date in the second period that is closest to the first period. In this case, the prediction period can be set to a timing in the second period when a plurality of predicted quality values ​​are closest to a predetermined condition.

[0017] (8) In the harvest support system of (4), the process of determining the prediction period may further include a process of outputting the plurality of predicted quality values ​​to an external device if the first period does not exist. In this case, the plurality of predicted quality values ​​can be provided to an operator as information necessary for determining the prediction period.

[0018] (9) In addition, in any one of the harvesting support systems (1) to (8) above, the measuring device may include a first measuring unit that measures the quality of the fruit and a second measuring unit that measures the quality of the fruit with a configuration different from that of the first measuring unit, and the first process may include a process of determining the multiple types of quality values ​​when the quality of the fruit is measured by the first measuring unit based on scaling data that indicates a correlation between the output of the first measuring unit and multiple reference quality values ​​based on the output of the second measuring unit. In this case, if the measurement accuracy of the second measuring unit is higher than that of the first measuring unit, the output of the first measuring unit can be corrected to a quality value equivalent to that of the second measuring unit, which has higher measurement accuracy than the first measuring unit, thereby easily improving the quality measurement accuracy.

[0019] (10) In the harvesting support system of (1) or (9), the fruit preferably includes grapes. A plurality of quality values ​​are taken into consideration when determining the optimum harvest time for grapes. Therefore, this embodiment can be suitably used for determining the optimum harvest time for grapes.

[0020] (11) In addition, in the harvest support system of (1) or (10) above, it is preferable that the multiple types of quality values ​​include at least two of sugar content, acidity, pH, and polyphenol content.

[0021] (12) From another perspective, the present disclosure provides a processing device for supporting the harvesting of fruit grown in a farm field, the processing device including a processing unit for acquiring multiple types of quality values ​​based on the output of a measurement device that measures the quality of fruit grown in the farm field, acquiring multiple time-series data in which the multiple types of quality values ​​are accumulated in a chronological order, and calculating a predicted period for the optimum harvest time of the fruit based on the multiple time-series data.

[0022] (13) From another perspective, the present disclosure provides a computer program for causing a computer to execute a process for supporting the harvesting of fruit grown in a field, the computer program causing the computer to execute the steps of acquiring multiple types of quality values ​​based on the output of a measurement device that measures the quality of the fruit, acquiring multiple pieces of time-series data in which the multiple types of quality values ​​are accumulated in chronological order, and calculating a predicted period for which the fruit is optimal for harvesting based on the multiple pieces of time-series data.

[0023] [Details of the embodiment] Preferred embodiments will now be described with reference to the drawings. Note that at least some of the embodiments described below may be combined in any desired manner.

[0024] [Overall Configuration of the Harvest Support System] FIG. 1 is a diagram showing an example of the overall configuration of a harvest support system according to an embodiment. In FIG. 1, the harvest support system 1 has a function of measuring the quality value of fruit grown in a field F. Furthermore, the harvest support system 1 has a function of calculating a predicted period for the optimum harvest time of the fruit. In this embodiment, the field F is, for example, a vineyard for growing grapes, which are used to make wine. Therefore, multiple trees T are grown in the field F. The multiple trees T are fruit trees, i.e., grapevines. The multiple trees T are grown in multiple rows. The harvest support system 1 has a function of measuring the quality value of grapes bearing fruit on the trees T. More specifically, the harvest support system 1 determines the quality value of the fruit through analysis using near-infrared light. The quality value of the fruit is a value based on various components that represent the quality of the fruit. The quality values ​​include sugar content, acidity, pH, polyphenols, etc.

[0025] The harvest support system 1 includes a measuring device 5, an agricultural machine 6, a management server 8, a management terminal 10, and a weather information server 7. The measuring device 5 includes a first measuring device 2 (first measurement unit) and a second measuring device 4 (second measurement unit). The first measuring device 2 and the second measuring device 4 are mounted on the agricultural machine 6. The first measuring device 2 and the second measuring device 4 are devices for measuring fruit quality. The first measuring device 2 and the second measuring device 4 have the function of acquiring spectral information in the near-infrared region contained in reflected light and transmitted light from the fruit. The spectral information in the near-infrared region includes information indicating the quality value of the fruit. Therefore, the first measuring device 2 and the second measuring device 4 output the spectral information in the near-infrared region as a quality measurement result. The first measuring device 2 and the second measuring device 4 will be described in detail later. The harvest support system 1 further includes a manipulator 12. The manipulator 12 has the function of moving a second measuring head 26 (described later) of the second measuring device 4.

[0026] The agricultural machine 6, the management server 8, and the management terminal 10 are connected to each other so that they can communicate with each other via a public network NW such as the Internet. The agricultural machine 6 has a communication function using, for example, a mobile communication system. The agricultural machine 6 is connected to the public network NW via a wireless base station BS of the mobile communication system.

[0027] The management server 8 has a function of performing processing to determine the quality value of fruit based on the output of the first measuring device 2 and the output of the second measuring device 4. The management terminal 10 is a terminal operated by an operator 14 of the harvest support system 1. The management terminal 10 has a function of accepting operations on the harvest support system 1 by the operator 14 and a function of outputting quality values ​​and the like determined by the management server 8 to the operator 14. The weather information server 7 collects and stores weather information issued by the Japan Meteorological Agency and the like. The weather information includes weather forecast information. The weather forecast information includes forecasts of sunshine and rain, predicted precipitation, predicted temperature, and predicted humidity for a period of approximately one week to ten days. The weather information server 7 provides the management server 8 and the management terminal 10 with the necessary information in response to requests from the management server 8 and the management terminal 10.

[0028] The agricultural machine 6 is, for example, a tractor. The agricultural machine 6 is capable of traveling within a field F. The agricultural machine 6 travels between rows of multiple trees T within the field F and can approach all of the trees T within the field F. The agricultural machine 6 can travel within the field F by being manually driven by an operator, or can travel within the field F by being automatically driven based on control commands from the management server 8 or control commands from the management terminal 10 based on input from an operator 14.

[0029] 2 is a block diagram showing the on-board network of the agricultural machine 6. The agricultural machine 6 has an on-board network 6a that complies with a communication standard such as CAN (Controller Area Network). The on-board network 6a includes a control device 18, an input / output device 19, and a communication device 20.

[0030] The communication device 20 functions as a mobile terminal in a mobile communication system. Therefore, the communication device 20 communicates wirelessly with a radio base station BS. The control device 18 is an ECU (Electronic Control Unit) that controls the driving system and work system of the agricultural machine 6. The control device 18 controls each part of the agricultural machine 6 based on the driving operation of the operator. The control device 18 is connected to the public network NW via the communication device 20. Therefore, the control device 18 is connected to the management server 8 and the management terminal 10 so that they can communicate with each other. The control device 18 exchanges necessary information with the management server 8 and the management terminal 10 via the public network NW. If the agricultural machine 6 is capable of autonomous driving, the control device 18 controls cameras and sensors that grasp the surroundings of the agricultural machine 6, as well as the drive system and steering system of the agricultural machine 6, based on control commands and the like from the management server 8 and the management terminal 10, and performs processing to execute autonomous driving. The input / output device 19 has a function to accept operations from the operator and a function to output information and the like to the operator.

[0031] In addition, the first measuring device 2, the second measuring device 4, and the manipulator 12 are connected to the in-vehicle network 6a. The first measuring device 2, the second measuring device 4, and the manipulator 12 are connected to the public network NW by a communication device 20. Therefore, the first measuring device 2, the second measuring device 4, and the manipulator 12 are connected to the management server 8 and the management terminal 10 so that they can communicate with each other. This allows the first measuring device 2, the second measuring device 4, and the manipulator 12 to provide necessary information to the management server 8 and the management terminal 10. In addition, the first measuring device 2, the second measuring device 4, and the manipulator 12 can be controlled by the management server 8 and the management terminal 10.

[0032] [Regarding the First Measuring Device 2 and the Second Measuring Device 4] Figure 3 is a block diagram showing an example of the first measuring device 2. As described above, the first measuring device 2 is a device that measures the quality of fruit and outputs spectral information in the near-infrared region as a quality measurement result. The first measuring device 2 receives light reflected from the fruit and outputs spectral information obtained by dispersing the received reflected light. Note that spectral information refers to information indicating the relationship between wavelength and light intensity in the near-infrared region.

[0033] As shown in FIG. 3 , the first measurement device 2 includes a first measurement head 22 and a first control unit 24. The first measurement head 22 includes, for example, a hyperspectral camera. A hyperspectral camera is a spectroscopic camera with multiple detection wavelength bands capable of detecting light intensity. The first measurement head 22 (hyperspectral camera) has multiple detection wavelength bands in the near-infrared region. The first measurement head 22 captures an image of a predetermined imaging region to acquire a two-dimensional image. That is, the first measurement head 22 receives reflected light when sunlight is irradiated onto an object present in the imaging region to acquire a two-dimensional image. The pixels constituting this two-dimensional image have luminance information indicating the light intensity of multiple detection wavelength bands. Therefore, the image acquired by the first measurement head 22 has spectral information in the near-infrared region for each pixel. Hereinafter, the image acquired by the first measurement head 22 will also be referred to as a spectral image. In this way, the first measurement head 22 can acquire spectral information in the near-infrared region for each imaging region.

[0034] The first measuring head 22 is used to capture an image of a grape berry K, which is the object to be measured. The first measuring head 22 is a camera capable of capturing two-dimensional images. Therefore, when capturing an image of the fruit K using the first measuring head 22, a predetermined first distance D1 is provided between the first measuring head 22 and the grape berry K. In this embodiment, the grape berry K refers to a bunch containing multiple fruit kernels k1. When the first distance D1 is provided, the imaging area A of the first measuring head 22 is large enough to capture an image of multiple fruits K scattered among the branches and leaves of a single tree T. Therefore, the spectral image captured by the first measuring head 22 includes multiple fruits K. The first measuring head 22 outputs a spectral image in which multiple fruits K are captured. In other words, the first measuring head 22 receives reflected light from the multiple fruits K and outputs a spectral image.

[0035] The first measuring head 22 provides the captured spectral image to the first control unit 24. The first control unit 24 is connected to the communication device 20 of the agricultural machine 6. The first control unit 24 has a function of controlling the first measuring head 22 based on control commands and the like provided from the management server 8 or the management terminal 10. The first control unit 24 also has an input unit (not shown) that accepts operational inputs and has a function of controlling the first measuring head 22 based on operational inputs from the operator. The first control unit 24 also transmits the spectral image provided from the first measuring head 22 to the management server 8. In other words, the spectral image acquired by the first measuring head 22 is provided to the management server 8 as the output (measurement result) of the first measuring device 2.

[0036] 4A is a block diagram showing an example of the second measurement device 4. As described above, the second measurement device 4 is a device for measuring the quality of the fruit K and outputs spectral information in the near-infrared region as a quality measurement result. The second measurement device 4 irradiates the fruit K with light, receives the transmitted light that has passed through the fruit K, and outputs spectral information obtained by separating the transmitted light.

[0037] 4A , the second measuring device 4 includes a second measuring head 26, a light source 27, a spectroscope 28, and a second control unit 30. The second measuring head 26 has a light-projecting unit 26a and a light-receiving unit 26b. The light-projecting unit 26a and the light source 27 are connected by an optical fiber. The light source 27 is, for example, a halogen lamp. Light from the light source 27 is guided to the light-projecting unit 26a of the second measuring head 26 through the optical fiber. The light-projecting unit 26a irradiates the light from the light source 27 toward the fruit K, which is the object to be measured.

[0038] The light receiving unit 26b receives transmitted light that has passed through the fruit K. The transmitted light received by the light receiving unit 26b is generated when light from the light projecting unit 26a is irradiated onto the fruit K. The light receiving unit 26b and the spectroscope 28 are connected by an optical fiber. The transmitted light received by the light receiving unit 26b is guided to the spectroscope 28 through the optical fiber.

[0039] 4B is a cross-sectional view of the second measuring head 26. As shown in FIG. 4B, the second measuring head 26 has a head body 26c. The head body 26c is an annular member formed of resin or the like. The head body 26c is fixed to the tip of the manipulator 12. Therefore, the second measuring head 26 can be moved by the manipulator 12.

[0040] The light-projecting unit 26a is provided on the annular tip surface 26c1 of the head main body 26c. The light-projecting unit 26a is annular. When light from the light source 27 is applied to the light-projecting unit 26a, it is scattered internally and then uniformly emitted from the light-projecting unit 26a. The light from the light-projecting unit 26a is irradiated onto the fruit kernel k1 of the fruit K, producing transmitted light that passes through the fruit kernel k1 and reflected light that is reflected on the surface of the fruit kernel k1. The light-receiving unit 26b is provided in the hole 26c2 of the head main body 26c. The light-receiving unit 26b receives transmitted light that has passed through the hole 26c2. Note that the light-receiving unit 26b mainly receives transmitted light, but also receives reflected light.

[0041] When measuring fruit K using the second measuring device 4, it is necessary to irradiate fruit grain k1 with light using the light projecting unit 26a, and therefore it is necessary to bring the second measuring head 26 close to the fruit grain k1 of the fruit K. When measuring fruit K using the second measuring device 4, the second measuring head 26 is placed at a measurement position as shown in Fig. 4B. The measurement position shown in Fig. 4B refers to a position where there is a second distance D2 between the second measuring head 26 and the fruit grain k1. The second distance D2 is shorter than the first distance D1 (see Fig. 3).

[0042] The second measuring head 26 is placed at the measurement position by the manipulator 12. In this way, the manipulator 12 functions as a movement mechanism that moves the second measuring head 26 to the measurement position.

[0043] When the second measurement head 26 is positioned close to the fruit k1 at the measurement position, light from the light-emitting unit 26a is irradiated onto the fruit k1. The light irradiated onto the fruit k1 passes through the fruit k1 as shown by the arrows in Figure 4B, producing transmitted light. The transmitted light passes through the hole 26c2 and is received by the light-receiving unit 26b.

[0044] 4A, the spectroscope 28 has a function of separating transmitted light and a light-receiving function of converting the separated spectrum into a signal. When the spectroscope 28 receives the transmitted light received by the light-receiving unit 26b, the spectroscope 28 outputs spectral information obtained by separating the transmitted light. The spectral information output by the spectroscope 28 is provided to the second control unit 30.

[0045] The second control unit 30 is connected to the communication device 20 of the agricultural machine 6. The second control unit 30 has a function of controlling the light source 27 and the spectroscope 28 based on control commands and the like provided from the management server 8 or the management terminal 10. The second control unit 30 also has an input unit (not shown) that accepts operational input, and also has a function of controlling the light source 27 and the spectroscope 28 based on operational input from the operator.

[0046] Furthermore, the second control unit 30 transmits the spectral information provided by the spectrometer 28 to the management server 8. That is, the spectral information acquired by the spectrometer 28 is provided to the management server 8 as the output (measurement result) of the second measuring device 4.

[0047] As described above, the configuration of the first measuring device 2 and the configuration of the second measuring device 4 differ in terms of the presence or absence of a spectroscopic camera and a light source. Furthermore, the first measuring device 2 measures the quality of multiple fruits K collectively, while the second measuring device 4 measures the quality of a single fruit k1. Here, the measurement accuracy of the quality value indicated by the spectral information output by the second measuring device 4 is higher than the measurement accuracy of the quality value indicated by the spectral image (spectral information) output by the first measuring device 2. Measurement by the second measuring device 4 involves bringing the second measuring head 26 close to the fruit k1 and acquiring spectral information from the transmitted light when irradiating the fruit k1 with light, making it less susceptible to the influence of the surrounding environment. On the other hand, measurement by the first measuring device 2 involves acquiring spectral information from the reflected light from the fruit K, making it more susceptible to the influence of the surrounding environment. As a result, the measurement accuracy of the second measuring device 4 is higher than that of the first measuring device 2.

[0048] In this embodiment, the first measuring device 2 and the second measuring device 4 are mounted on the agricultural machine 6, but this is not limiting. The first measuring device 2 and the second measuring device 4 may be carried by the worker or may be mounted on a cart or the like. Alternatively, only the first measuring device 2 may be mounted on the agricultural machine 6, and the second measuring device 4 may be placed in a predetermined position. In this case, observation by the first measuring device 2 is performed while patrolling, and observation by the second measuring device 4 is performed at a fixed point.

[0049] [Regarding the Management Server 8] Fig. 5 is a block diagram showing an example configuration of the management server 8. As shown in Fig. 5, the management server 8 (processing device) is a type of information processing device having a processing unit 32, a storage unit 34, and a communication unit 36. The communication unit 36 ​​is a communication interface capable of communicating with external devices via a public network NW. The processing unit 32 is, for example, one of various processors suitable for computer control, such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), or an FPGA (Field Programmable Gate Array).

[0050] The storage unit 34 is, for example, a flash memory, a hard disk, a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The storage unit 34 stores computer programs and necessary information to be executed by the processing unit 32. The processing unit 32 executes computer programs stored in a computer-readable, non-transitory recording medium such as the storage unit 34 to realize various processing functions of the processing unit 32.

[0051] The memory unit 34 also stores a plurality of first calibration curve data C1, a plurality of second calibration curve data C2, a plurality of scaling data S, and a quality value database 40. The plurality of first calibration curve data C1 is data for calculating a quality value (first quality value) based on the output of the first measurement device 2. The plurality of second calibration curve data C2 is data for calculating a quality value (second quality value) based on the output of the second measurement device 4. The plurality of scaling data S is data indicating the correlation between the first quality value and the second quality value (reference quality value). The plurality of scaling data S is generated in a quality value acquisition process 32a executed by the processing unit 32. The quality value database 40 is a database for registering and storing the measured quality values ​​obtained by the quality value acquisition process 32a executed by the processing unit 32.

[0052] The processing unit 32 has the function of executing a quality value acquisition process 32a and a prediction process 32b. The quality value acquisition process 32a (first process) is a process for determining a measured quality value of the target fruit KT based on the scaling data S and the first quality value of the target fruit KT. The target fruit KT is a fruit among the fruits K for which a measured quality value is to be determined. The measured quality value is a quality value determined by the quality value acquisition process 32a, which is a quality value obtained by correcting the first quality value of the target fruit KT determined based on the output of the first measuring device 2, and which has higher accuracy than the first quality value. The processing unit 32 determines multiple types of measured quality values ​​based on the output of the measuring device 5 (the first measuring device 2 and the second measuring device 4). The multiple types of quality values ​​include sugar content, acidity, pH, polyphenol content, etc.

[0053] The prediction process 32b includes a process (second process) for acquiring a plurality of time-series data in which the plurality of measured quality values ​​obtained by the quality value acquisition process 32a are accumulated in time series, and a process (third process) for calculating a predicted period for which the optimum harvest time for the fruit is predicted based on the plurality of time-series data. These processes will be described in detail later.

[0054] [Quality Measurement Work of Fruit K in Farm Field F] Fig. 6 is a flowchart showing an example of quality measurement work of fruit K in farm field F. As shown in Fig. 6, in the quality measurement work of fruit K, first, quality measurement is performed using the first measuring device 2 and the second measuring device 4 (step S1 in Fig. 6).

[0055] FIG. 7 is a plan view of a field F. In FIG. 7, two mutually orthogonal directions are defined as the X direction and the Y direction. As shown in FIG. 7, one of the X directions is defined as the X1 direction, and the opposite direction of the X1 direction is defined as the X2 direction. One of the Y directions is defined as the Y1 direction, and the opposite direction of the Y1 direction is defined as the Y2 direction. Quality measurement by the measuring devices 2 and 4 is performed by having the agricultural machine 6 travel within the field F. In the field F shown in FIG. 7, for example, multiple trees T (15 trees in the illustrated example) are cultivated. Five trees T are arranged in the Y direction and three trees T are arranged in the X direction. If the Y direction is defined as the row direction, the trees T are arranged in three rows. Each row consists of five trees T. Each of the 15 trees T cultivated in the field F is assigned an ID as identification information. Integers from 1 to 15 are assigned to the 15 trees T as IDs. The quality measurement work of the fruits K in the field F is performed on all 15 trees T in the field F. Of the 15 trees T, the hatched trees T (trees T with IDs = 2, 4, 8, 12, and 14) are trees T for which the quality measurement of the fruits K is performed by both the first measuring device 2 and the second measuring device 4. The other trees T are trees T for which the quality measurement of the fruits K is performed only by the first measuring device 2. In other words, the first measuring device 2 performs quality measurement of the fruits K on all trees T, and the second measuring device 4 performs quality measurement of the fruits K on some of the trees T in the field F.

[0056] The quality measurement by the first measuring device 2 and the second measuring device 4 in step S1 may be performed by operation of the worker driving the agricultural machine 6, or may be performed by remote control based on a command from the operator 14. When the quality measurement is performed by the worker, the agricultural machine 6 travels by manual operation by the worker, and the first measuring device 2 and the second measuring device 4 perform quality measurement by manual operation by the worker. When the quality measurement is performed by remote control based on a command from the operator 14, the agricultural machine 6 travels by automatic driving, and the first measuring device 2 and the second measuring device 4 perform quality measurement by control based on a command from the operator 14. In this embodiment, a case where the worker driving the agricultural machine 6 performs quality measurement will be described.

[0057] A worker rides the agricultural machine 6 and travels along the side of a row of multiple trees T in the field F along the dashed line in FIG. 7 . Thus, the agricultural machine 6 approaches each tree T in order, starting with the tree T with ID=1. When the worker reaches the side of each tree T, he stops the agricultural machine 6 and performs quality measurement using the first measuring device 2 and the second measuring device 4. For example, as shown in FIG. 7 , when the worker stops the agricultural machine 6 next to the tree T with ID=1 (on the X1 direction), the worker operates the first measuring device 2 to capture an image of multiple fruits K on the tree T with ID=1, including the multiple fruits K in the imaging area. As a result, the first measuring device 2 outputs a spectral image of the multiple fruits K on the tree T with ID=1 and provides it to the management server 8. At this time, the first measuring device 2 adds the ID of the tree T to the output spectral image and transmits it. The first measuring device 2 adds the ID of the tree T to the spectral image in response to the operator's operation input. When the output (spectral image) of the first measuring device 2 is provided to the management server 8, the quality measurement of the fruit K on the tree T with ID=1 is completed.

[0058] After completing the quality measurement of the fruit K on the tree T with ID=1, the worker moves the agricultural machine 6 in the Y2 direction and stops the agricultural machine 6 to the side (X1 direction side) of the tree T with ID=2. Here, the tree T with ID=2 is the tree T for which quality measurement of the fruit K will be performed by both the first measuring device 2 and the second measuring device 4. Therefore, the worker operates the first measuring device 2 to capture an image of the multiple fruits K on the tree T with ID=2, including them in the imaging area. As a result, the first measuring device 2 outputs a spectral image of the multiple fruits K on the tree T with ID=2, and provides the spectral image with the ID of the tree T added to it to the management server 8.

[0059] The worker then operates the second measuring device 4 to perform quality measurement. Using the manipulator 12, the worker brings the second measuring head 26 of the second measuring device 4 close to the tree T. The worker brings the second measuring head 26 close to one of the multiple fruits K on the tree T that has been imaged by the first measuring device 2. The worker then operates the second measuring device 4 to cause the second measuring device 4 to receive light transmitted through the fruit K. The second measuring device 4 then outputs spectral information for the fruit K on the tree T with ID = 2 and provides it to the management server 8. In the following description, the fruit K measured by the second measuring device 4 is also referred to as the reference fruit KR. The second measuring device 4 also adds the ID of the tree T to the output spectral image and transmits it. The second measuring device 4 adds the ID of the tree T to the spectral information in response to a user operation. The second measuring device 4 may also add location information indicating the location of the reference fruit KR to the spectral information in response to a user operation. When the output of the first measuring device 2 and the output of the second measuring device 4 are given to the management server 8, the quality measurement of the tree T with ID=2 is completed.

[0060] As described above, the worker measures the quality of fruits K on multiple trees T in the order of the IDs. The outputs of the first measuring device 2 and the second measuring device 4 (spectral images and spectral information) are stored in the storage unit 34 of the management server 8. Therefore, the storage unit 34 stores spectral images assigned with IDs 1 to 15 and spectral information assigned with IDs 2, 4, 8, 12, and 14.

[0061] As shown in FIG. 6, after the quality measurements by the first measurement device 2 and the second measurement device 4 are completed, the management server 8 then performs a quality value acquisition process (step S2 in FIG. 6).

[0062] [Quality Value Acquisition Process] Fig. 8 is a flowchart showing an example of the quality value acquisition process. In the quality value acquisition process, the measured quality values ​​of the fruits K of the 15 trees T are calculated.

[0063] 8, first, the processing unit 32 of the management server 8 calculates a first quality value and a second quality value (step S21 in FIG. 8). The processing unit 32 calculates the first quality value based on the output (spectral image) of the first measurement device 2. The processing unit 32 also calculates the second quality value based on the output (spectral information) of the second measurement device 4.

[0064] The processing unit 32 determines the first quality value and the second quality value by a spectroscopic analysis technique using near-infrared light. Quantitative analysis by spectroscopic analysis using near-infrared light is an analytical method that utilizes the fact that the degree of light absorption in a wavelength band where a specific component significantly absorbs light varies depending on the content of the specific component. The processing unit 32 determines the sugar content, acidity, pH, and polyphenol content as the first quality value and the second quality value. The wavelength bands in the near-infrared region used when analyzing the sugar content, acidity, pH, and polyphenol content are set in advance.

[0065] First, the process of calculating the first quality value will be described. The processing unit 32 obtains spectral images of wavelength bands corresponding to each quality value from the spectral image output by the first measurement device 2. In this embodiment, the processing unit 32 obtains a spectral image of a wavelength band corresponding to sugar content, a spectral image of a wavelength band corresponding to acidity, a spectral image of a wavelength band corresponding to pH, and a spectral image of a wavelength band corresponding to polyphenols. Note that a spectral image is an image in which the brightness of pixels constituting the image represents the light intensity in a predetermined wavelength band.

[0066] The following describes the process performed by the processing unit 32 when determining the sugar content as the first quality value from the spectral image in the wavelength band corresponding to the sugar content. The processing unit 32 identifies image regions of multiple fruit K portions from the spectral image in the wavelength band corresponding to the sugar content. The brightness of pixels included in these multiple image regions indicates the light intensity in the wavelength band corresponding to the sugar content. In other words, the brightness of a pixel indicates the amount of light absorption.

[0067] The processing unit 32 calculates a representative value of the luminance obtained from one spectral image based on the luminance of pixels included in the image regions of the plurality of fruits K. For example, the processing unit 32 may use the average value of the luminance of pixels included in the image regions of the plurality of fruits K as the representative value. Furthermore, a threshold value can be set for the luminance in order to identify the image regions of the plurality of fruits K in the spectral image. In this case, the processing unit 32 can identify the region equal to or greater than the threshold as the image region of the fruit K.

[0068] Furthermore, when the reference fruit KR can be identified from the spectral images of the trees T with IDs 2, 4, 8, 12, and 14 using position information indicating the position of the reference fruit KR, the processing unit 32 may calculate a representative value of brightness using pixels included in an image area that includes only the portion of the reference fruit KR. In this case, the brightness of the reference fruit KR can be calculated.

[0069] Next, the processing unit 32 references the first calibration curve data C1 and calculates the sugar content based on the brightness (representative value). FIG. 9 is a diagram showing an example of multiple first calibration curve data C1. FIG. 9 schematically shows the first calibration curve data C1. As shown in FIG. 9, among the multiple first calibration curve data C11 to C14, the first calibration curve data C11 shows the correlation between the brightness of pixels in the spectral image in a wavelength band corresponding to the sugar content and the sugar content. Although omitted from FIG. 9, the first calibration curve data C12 shows the correlation between the brightness of pixels in the spectral image in a wavelength band corresponding to the acidity and the acidity. The first calibration curve data C13 shows the correlation between the brightness of pixels in the spectral image in a wavelength band corresponding to the pH and the pH. The first calibration curve data C14 shows the correlation between the brightness of pixels in the spectral image in a wavelength band corresponding to the polyphenols and the polyphenol content.

[0070] The multiple first calibration curve data C1 may be any data that can be used to calculate a first quality value from the brightness of the pixels of the spectroscopic image, and may be a table obtained by experiment, simulation, etc., or a linear or nonlinear mathematical formula.

[0071] When determining the sugar content as the first quality value, the processing unit 32 refers to the first sugar content calibration curve data C11 and determines the sugar content corresponding to the brightness determined above. This sugar content is the first quality value.

[0072] The processing unit 32 also determines the acidity, pH, and polyphenol content in addition to the sugar content using a method similar to that for sugar content. As described above, the processing unit 32 determines the first quality values ​​(sugar content, acidity, pH, and polyphenol content) for each of the trees T with IDs 1 to 15 based on the output of the first measurement device 2.

[0073] Next, a process for calculating the second quality value will be described. The processing unit 32 calculates the absorbance of a wavelength band corresponding to each quality value from the spectral information output by the second measurement device 4. The absorbance is a value indicating the degree of absorption of light in a predetermined wavelength band when the light from the light source 27 is used as a reference.

[0074] The following describes the process performed by the processing unit 32 when determining the sugar content as the first quality value. The processing unit 32 references the second calibration curve data C2 and determines the sugar content based on the absorbance (representative value).

[0075] FIG. 10 is a diagram showing an example of multiple second calibration curve data C2. FIG. 10 schematically shows the second calibration curve data C2. As shown in FIG. 10, among the multiple second calibration curve data C21 to C24, the second calibration curve data C21 shows the correlation between the sugar content and the absorbance in the wavelength band corresponding to the sugar content. Although omitted from FIG. 10, the second calibration curve data C22 shows the correlation between the acidity and the absorbance in the wavelength band corresponding to the acidity. The second calibration curve data C23 shows the correlation between the pH and the absorbance in the wavelength band corresponding to the pH. The second calibration curve data C24 shows the correlation between the polyphenol content and the absorbance in the wavelength band corresponding to the polyphenols.

[0076] The multiple second calibration curve data C2 may be any data that allows the second quality value to be calculated from the absorbance, and may be a table obtained by experiment, simulation, etc., or a linear or nonlinear mathematical formula.

[0077] When determining the sugar content as the second quality value, the processing unit 32 refers to the second sugar content calibration curve data C21 and determines the sugar content corresponding to the absorbance determined above. This sugar content is the second quality value.

[0078] The processing unit 32 also determines the acidity, pH, and polyphenol content in addition to the sugar content using a method similar to that for sugar content. As described above, the processing unit 32 determines the second quality values ​​(sugar content, acidity, pH, and polyphenol content) for each of the trees T with IDs 2, 4, 8, 12, and 14 based on the output of the second measurement device 4.

[0079] In this way, the processing unit 32 calculates the first quality value for each of the trees T with IDs 1 to 15 and the second quality value for each of the trees T with IDs 2, 4, 8, 12, and 14 (step S21 in FIG. 8). Next, the processing unit 32 generates scaling data S as shown in FIG. 8 (step S22 in FIG. 8).

[0080] The scaling data S is calculated based on the first quality value and the second quality value obtained when the reference fruit KR is measured by both the first measuring device 2 and the second measuring device 4. That is, the scaling data S is calculated based on the first quality value for each of the trees T with IDs = 2, 4, 8, 12, and 14 and the second quality value for each of the trees T with IDs = 2, 4, 8, 12, and 14. Hereinafter, the first quality values ​​for IDs = 2, 4, 8, 12, and 14 will also be referred to as reference first quality values, and the second quality values ​​for IDs = 2, 4, 8, 12, and 14 will also be referred to as reference second quality values.

[0081] The first quality value for each of the trees T with IDs = 2, 4, 8, 12, and 14 may be the average of the quality values ​​of multiple fruits K including the reference fruit KR. Also, the first quality value for each of the trees T with IDs = 2, 4, 8, 12, and 14 may include only the quality value of the reference fruit KR.

[0082] 11 is a diagram showing an example of scaling data S. As shown in FIG. 11, among the multiple scaling data S1 to S4, scaling data S1 shows the correlation between the sugar content measured by the first measurement device 2 and the sugar content measured by the second measurement device 4. Although omitted from FIG. 11, scaling data S2 shows the correlation between the acidity measured by the first measurement device 2 and the acidity measured by the second measurement device 4. Scaling data S3 shows the correlation between the pH measured by the first measurement device 2 and the pH measured by the second measurement device 4. Scaling data S4 shows the correlation between the polyphenol content measured by the first measurement device 2 and the polyphenol content measured by the second measurement device 4.

[0083] The scaling data S1 includes a line L showing the correlation between the sugar content measured by the first measuring device 2 and the sugar content measured by the second measuring device 4. The horizontal axis of the scaling data S1 is the sugar content measured by the first measuring device 2 (first quality value), and the vertical axis is the sugar content measured by the second measuring device 4 (second measured value). Line L is a straight line passing through points VU and VL. The five points in FIG. 11 are points plotted for a reference fruit KR having a reference first quality value and a reference second quality value. Of the five points, point VU is the point on the reference fruit KR having the maximum sugar content measured by the multiple first measuring devices 2 (multiple reference first quality values) and the maximum sugar content measured by the multiple second measuring devices 4 (multiple reference second quality values). Point VL is a point on the reference fruit KR that has the smallest sugar content (multiple reference first quality values) measured by multiple first measuring devices 2 and the smallest sugar content (multiple reference second quality values) measured by multiple second measuring devices 4.

[0084] In this manner, in this embodiment, the scaling data S1 is calculated based on a plurality of reference first quality values ​​and a plurality of reference second quality values ​​obtained when a plurality of reference fruits KR are measured by both the first measuring device 2 and the second measuring device 4. In this case, the first quality values ​​and the second quality values ​​are associated with each other by the plurality of reference fruits KR. This allows for further improvement in the accuracy of the scaling data S.

[0085] Furthermore, the scaling data S1 is calculated based on the maximum and minimum values ​​of the multiple reference first quality values ​​(outputs of the first measurement device 2) and the maximum and minimum values ​​of the multiple reference second quality values. In this case, the reference first quality value and the reference second quality value are associated with each other based on at least the respective maximum and minimum values. Therefore, the scaling data S1 can be easily calculated without reducing accuracy.

[0086] The other scaling data S2 to S4 are calculated in the same manner as the scaling data S1. The scaling data S may be any data indicating the correlation between the first quality value (output of the first measurement device 2) and the second quality value, and may be a table or a mathematical formula.

[0087] In this way, the processing unit 32 generates the scaling data S (step S22 in FIG. 8). Next, the processing unit 32 calculates a measurement quality value as shown in FIG. 8 (step S23 in FIG. 8).

[0088] The processing unit 32 selects a target fruit KT from the fruits K on the trees T with IDs 1 to 15, and determines a measured quality value of the target fruit KT based on the scaling data S and the first quality value of the target fruit KT. The processing unit 32 selects the target fruit KT in order from the fruits K on the trees T with IDs 1 to 15, and repeatedly determines the measured quality values. In this way, the processing unit 32 determines the measured quality value of each of the fruits K on the trees T with IDs 1 to 15. The processing unit 32 references the scaling data S and determines a second quality value corresponding to the first quality value of the target fruit KT. The processing unit 32 determines the second quality value corresponding to this first quality value as the measured quality value of the target fruit KT.

[0089] For example, a case will be described in which the sugar content is calculated as a measured quality value of fruit K on tree T with ID = 1. FIG. 12 is a diagram showing the manner in which the processing unit 32 calculates the measured quality value of fruit K on tree T with ID = 1. The processing unit 32 selects fruit K on tree T with ID = 1 as target fruit KT. Next, the processing unit 32 references the scaling data S1 and calculates the sugar content (vertical axis in FIG. 12) measured by the second measuring device 4 on diagram L that corresponds to the sugar content (horizontal axis in FIG. 12), which is the first quality value of fruit K (target fruit KT) with ID = 1. The processing unit 32 sets the sugar content measured by the second measuring device 4 that corresponds to the first quality value as the measured quality value. In this way, the vertical axis of the scaling data S1 indicates the sugar content measured by the second measuring device 4 as well as the measured quality value calculated from the first quality value.

[0090] The processing unit 32 also determines the acidity, pH, and polyphenol content of the measured quality values ​​other than sugar content using a method similar to that for sugar content. The processing unit 32 repeats the same process for the fruits K on the trees T with IDs 1 to 15, and similarly determines the measured quality values ​​for each of the fruits K on the trees T with IDs 1 to 15. Note that for the fruits K on the trees T with IDs 2, 4, 8, 12, and 14, including the reference fruit KR, the second quality values ​​measured by the second measuring device 4 may be used as the measured quality values ​​as they are.

[0091] According to this quality value acquisition process, a second quality value corresponding to the first quality value (output of the first measuring device) of the target fruit KT can be calculated based on the scaling data S, and the calculated second quality value can be used as the measured quality value of the target fruit KT. As a result, the first quality value of the target fruit KT can be corrected to a value equivalent to the quality value of the second measuring device 4, which has higher measurement accuracy than the first measuring device 2, and the quality measurement accuracy can be easily improved.

[0092] Furthermore, because the first measuring head 22 includes a hyperspectral camera (spectroscopic camera), the first measuring device 2 can measure a wider range at once and is more convenient than the second measuring device 4, which is equipped with a light source 27 and a spectroscope 28 that disperses transmitted light. Therefore, by using the first measuring device 2, convenience can be improved while improving the accuracy of quality measurement.

[0093] In this way, the processing unit 32 calculates the measurement quality value (step S23 in FIG. 8). Next, the processing unit 32 proceeds to step S3 in FIG. 6 and executes the prediction process (step S3 in FIG. 6).

[0094] [Regarding the Prediction Process] FIG. 13 is a flowchart illustrating an example of the prediction process. In the prediction process, the processing unit 32 of the management server 8 calculates a predicted period for which the optimum fruit harvesting period is predicted, as described above. In this embodiment, the prediction period is calculated using a one-day period as the minimum unit. As shown in FIG. 13 , the processing unit 32 first registers multiple types of measured quality values ​​(sugar content, acidity, pH, and polyphenol content) acquired by the quality value acquisition process 32a in the quality value database 40 (step S31 in FIG. 13 ). FIG. 14 is a diagram illustrating an example of the quality value database 40. The quality value database 40 includes a sugar content database 40a, an acidity database 40b, a pH database 40c, and a polyphenol database 40d. In FIG. 14 , the sugar content database 40a stores the ID of the tree T and the sugar content in association with each other. Furthermore, the sugar content is registered in chronological order for each measurement date. That is, the sugar content database 40a stores time-series data in which the sugar content of each tree T is accumulated in time series.

[0095] The acidity database 40b, pH database 40c, and polyphenol database 40d are similar to the sugar content database 40a. Thus, the acidity database 40b, pH database 40c, and polyphenol database 40d store time-series data (multiple time-series data) of the acidity, pH, and polyphenol content of each tree T.

[0096] The processing unit 32 registers the four measured quality values, namely, sugar content, acidity, pH, and polyphenol content, in the quality value database 40. That is, the processing unit 32 acquires four time-series data in which the four measured quality values ​​are accumulated in time series (second processing).

[0097] Next, the processing unit 32 calculates a prediction period based on multiple pieces of time-series data (third process). More specifically, the processing unit 32 calculates a predicted quality value as shown in FIG. 13 (step S32 in FIG. 13). FIG. 15 is a graph showing an example of a predicted quality value. FIG. 15 shows time-series data and predicted sugar content for sugar content, as well as time-series data and predicted acidity for acidity. In FIG. 15, the horizontal axis represents the month and date, and the vertical axis represents sugar content and acidity. Also, in FIG. 15, circles represent the measured sugar content (measurement quality value), and squares represent the measured acidity (measurement quality value). Note that the following explanation will be given of a prediction process using two measurement quality values ​​(two time-series data) for sugar content and acidity.

[0098] In this embodiment, the processing unit 32 calculates the average value of the measured quality values ​​for the entire field F on each measurement date, and uses the average value as the quality value. For example, the sugar content value on date d0 is the average sugar content value of the trees T with IDs 1 to 15 measured on date d0. Similarly, the acidity value on date d0 is the average acidity value of the trees T with IDs 1 to 15 measured on date d0.

[0099] The processing unit 32 calculates a prediction curve L10 based on the sugar content values ​​(time-series data of sugar content). For example, the processing unit 32 calculates an approximate curve L11 based on time-series data of sugar content from before date d0, and calculates a sugar content prediction curve L10 by extending the approximate curve L11 in the direction toward the future from date d0. This prediction curve L10 shows the predicted sugar content values ​​in the future after date d0. In other words, the prediction curve L10 is a predicted sugar content (predicted quality value).

[0100] The processing unit 32 also calculates a predicted curve L20 based on the quality value of the acidity (time-series data of acidity). For example, the processing unit 32 calculates an approximate curve L21 based on time-series data of sugar content from before date d0, and then calculates a predicted curve L20 of acidity by extending the approximate curve L21 in the future direction from date d0. This predicted curve L20 indicates a predicted value of acidity in the future after date d0. In other words, the predicted curve L20 is a predicted acidity (predicted quality value). Note that a known method, such as regression analysis using the least squares method, can be used to calculate the approximate curve.

[0101] As described above, the processing unit 32 obtains the predicted curve L10 as the predicted sugar content and the predicted curve L20 as the predicted acidity.

[0102] Next, the processing unit 32 acquires environmental prediction information as shown in Fig. 13 (step S33 in Fig. 13). The processing unit 32 accesses the weather information server 7 and acquires the environmental prediction information. In this embodiment, the environmental prediction information includes weather information for a period of up to nine days from the present. In this embodiment, the weather information includes a forecast of sunny or rainy weather and a forecast precipitation amount for a period of up to nine days from the present. The forecast precipitation amount is a predicted value that indicates the predicted environmental state.

[0103] When the environment prediction information is acquired, the processing unit 32 proceeds to step S34, and obtains the first period and the second period (step S34 in FIG. 13).

[0104] The first period is a period during which the predicted sugar content and predicted acidity (multiple types of predicted quality values) satisfy predetermined conditions. The predetermined conditions are quality conditions that the fruit must meet at the time of harvest. The predetermined conditions include a condition for the predicted sugar content and a condition for the predicted acidity. The condition for the predicted sugar content is that the sugar content must be equal to or greater than a first threshold value Th1 (see FIG. 15). Furthermore, the condition for the predicted acidity is that the sugar content must be equal to or greater than a second threshold value Th2 (see FIG. 15). The first threshold value Th1 and the second threshold value Th2 are set to values ​​required for grapes that are used as raw materials for wine.

[0105] In the first period, both the predicted sugar content and the predicted acidity satisfy the predetermined conditions. Therefore, in Figure 15, the first period is the period in which the sugar content prediction curve L10 is equal to or greater than the first threshold value Th1 and the acidity prediction curve L20 is equal to or greater than the second threshold value Th2 overlap.

[0106] As shown by the approximate curve L11 and the predicted curve L10 in FIG. 15 , the sugar content tends to gradually increase over time. Therefore, the period in which the sugar content is equal to or greater than the first threshold value Th1 on the predicted curve L10 is the period after point P1 on the predicted curve L10. Point P1 is the point at which the predicted curve L10 reaches the first threshold value Th1. Also, as shown by the approximate curve L21 and the predicted curve L20 in FIG. 15 , the acidity tends to gradually decrease over time. Therefore, the period in which the acidity is equal to or greater than the second threshold value Th2 on the predicted curve L20 is the period before point P2 on the predicted curve L20. Point P2 is the point at which the predicted curve L20 reaches the second threshold value Th2. Thus, the sugar content and the acidity have different increasing and decreasing tendencies over time. Therefore, the first period in which both the sugar content and the acidity satisfy the specified conditions is the period from date d1 to date d2. Date d1 corresponds to point P1. Furthermore, the date d2 is the date corresponding to the point P2.

[0107] In other words, in this embodiment, the first period is the period from date d1 to date d2. In this way, the processing unit 32 calculates the first period based on the sugar content prediction curve L10 and the acidity prediction curve L20.

[0108] If the first period does not exist, the processing unit 32 outputs the prediction curves L10 and L20 to the operator 14 via the management terminal 10. If the first period does not exist, the prediction period cannot be determined. However, by outputting the prediction curves L10 and L20, it is possible to provide the operator 14 with information necessary for the operator 14 to determine the prediction period.

[0109] The second period is a period determined to be suitable for harvesting based on the predicted precipitation. More specifically, the second period is a period determined based on the results of comparing the predicted precipitation with a predetermined threshold. In this embodiment, the predetermined threshold is set to a predicted precipitation of 20 mm per day. The processing unit 32 determines that the weather is suitable for harvesting on days when the predicted precipitation per day is 20 mm or less.

[0110] Fig. 16 is a diagram showing a first example of a sunny / rainy forecast and predicted precipitation included in the environmental forecast information. Fig. 16 shows the sunny / rainy forecast and predicted precipitation for the next nine days from September 21st to September 29th. The sunny / rainy forecast and predicted precipitation shown in Fig. 16 are as follows: September 21st to September 24th: sunny / rainy forecast is "sunny", predicted precipitation is 0 mm September 25th: sunny / rainy forecast is "cloudy with occasional sunny spells", predicted precipitation is 20 mm September 26th to September 29th: sunny / rainy forecast is "sunny", predicted precipitation is 0 mm

[0111] In this case, the predicted precipitation for the nine days from September 21 to September 29 is 20 mm or less. Therefore, the processing unit 32 determines the nine days from September 21 to September 29 as the second period. As described above, the processing unit 32 determines the first period and the second period (step S34 in FIG. 13). Next, the processing unit 32 determines the prediction period based on the first period and the second period (step S35 in FIG. 13).

[0112] In principle, the processing unit 32 sets the overlapping period, in which the first period and the second period overlap, as the prediction period. In FIG. 16 , for example, the first period is assumed to be from September 26th to September 29th. In this case, the entire first period overlaps with the second period. Therefore, the entire first period is an overlapping period. Therefore, the processing unit 32 sets September 26th to September 29th as the prediction period.

[0113] Fig. 17 is a diagram showing a second example of fair / rain forecasts and predicted precipitation amounts included in the environmental forecast information. The fair / rain forecasts and predicted precipitation amounts shown in Fig. 17 are as follows: September 21st to September 24th: fair / rain forecast is "fine", predicted precipitation amount is 0 mm September 25th: fair / rain forecast is "cloudy with occasional fine weather", predicted precipitation amount is 20 mm September 26th: fair / rain forecast is "cloudy with occasional rain", predicted precipitation amount is 30 mm September 27th to September 28th: ​​fair / rain forecast is "rain", predicted precipitation amount is 80 mm September 29th: fair / rain forecast is "cloudy with occasional rain", predicted precipitation amount is 40 mm

[0114] In this case, the processing unit 32 determines the five days from September 21st to September 25th as the second period. Also, in FIG. 17, the first period is assumed to be from September 26th to September 29th. In this case, there is no overlapping period between the first period and the second period. If there is no overlapping period between the first period and the second period and the second period exists before the first period, the processing unit 32 sets the day of the second period that is closest to the first period as the prediction period. Therefore, in the case of FIG. 17, the processing unit 32 sets September 25th as the prediction period.

[0115] In this embodiment, if there is no overlapping period and the second period exists before the first period, the day in the second period that is closest to the first period is set as the predicted period, so that the predicted period can be set to a time in harvest-friendly weather that is as close as possible to the required quality conditions.

[0116] Fig. 18 is a diagram showing a third example of fair weather forecasts and predicted precipitation amounts included in the environmental forecast information. The fair weather forecasts and predicted precipitation amounts shown in Fig. 18 are as follows: September 21st to September 23rd: fair weather forecast is "fine", predicted precipitation amount is 0 mm September 24th: fair weather forecast is "cloudy with occasional rain", predicted precipitation amount is 40 mm September 25th: fair weather forecast is "rain", predicted precipitation amount is 80 mm September 26th to September 28th: ​​fair weather forecast is "fine", predicted precipitation amount is 0 mm September 29th: fair weather forecast is "cloudy with occasional rain", predicted precipitation amount is 40 mm

[0117] In this case, the processing unit 32 determines that the three days from September 21 to September 23 and the three days from September 26 to September 28 are the second period. Also, in Fig. 17, the first period is assumed to be from September 26 to September 29. In this case, the three days from September 26 to September 28 are the overlapping period.

[0118] Here, if there is a period that is not the second period immediately before the overlapping period, the processing unit 32 performs processing to provide an interval of at least one day between the non-second period and the prediction period. In Figure 18, September 24th and September 25th, which are immediately before the overlapping period, are periods that are not the second period. Therefore, the processing unit 32 sets September 27th and September 28th, of the three days from September 24th to September 28th, which are the overlapping period, to the prediction period. As a result, September 26th becomes the interval period.

[0119] Immediately after a period that is not the second period, the condition of the field F may be poor due to rainfall. In this regard, in the present embodiment, as described above, an interval period of at least one day is provided between a period that is not the second period and the prediction period, so that the prediction period can be set to avoid times when the condition of the field F is poor.

[0120] The processing unit 32 that has calculated the prediction period provides information indicating the prediction period to the management terminal 10 as necessary, and outputs the prediction period as the optimum fruit harvest time to the operator 14 via the management terminal 10 (step S36 in Fig. 13). After outputting the prediction period, the processing unit 32 returns to the flowchart shown in Fig. 6 and repeats the same processing again. The processing unit 32 performs quality value acquisition processing and prediction processing every time quality measurement is performed in the field F.

[0121] As described above, the process (third process) of calculating a predicted period based on two time series data (time series data of sugar content and time series data of acidity) includes a process of calculating predicted sugar content and predicted acidity (multiple types of predicted quality values), a process of acquiring environmental prediction information, and a process of calculating a predicted period based on the environmental prediction information and multiple types of predicted quality values.

[0122] According to the above configuration, future quality values ​​can be predicted using the time-series data of sugar content and time-series data of acidity (multiple time-series data), and a predicted period for the optimum harvest time can be calculated based on the predicted quality values. As a result, the predicted period can be output as a prediction result of the optimum harvest time of fruit.

[0123] The prediction process of this embodiment also includes a process of calculating a sugar content prediction curve L10 and an acidity prediction curve L20 (predicted sugar content and predicted acidity) based on the time-series data of sugar content and the time-series data of acidity (step S32 in FIG. 13), a process of acquiring environmental prediction information indicating an environmental prediction for the field F (step S33 in FIG. 13), and a process of calculating a prediction period based on the environmental prediction information, predicted sugar content, and predicted acidity. In this case, the prediction period can be calculated by taking into account the environmental prediction information for the field F in addition to the predicted quality value of the fruit.

[0124] The process of determining the predicted period based on the environmental prediction information, predicted sugar content, and predicted acidity includes a process of determining a first period in which the predicted sugar content and predicted acidity satisfy predetermined conditions, and a process of determining a second period in which the weather is judged to be suitable for harvesting based on the results of comparing the predicted precipitation with a predetermined threshold (step S34 in FIG. 13), and a process of determining the predicted period based on the first period and the second period (step S35 in FIG. 13). This makes it possible to appropriately determine the predicted period based on the relationship between an appropriate period for the quality value and an appropriate period for the weather.

[0125] [Other] The embodiments disclosed herein should be considered illustrative in all respects and not restrictive. For example, in the above embodiment, the environmental prediction information includes a sunshine / rain forecast and a predicted precipitation amount. However, disease risk information related to fruit diseases can also be used as the environmental prediction information. When disease risk information is used as the environmental prediction information, a disease prediction model that outputs disease risk information is stored in the memory unit 34 of the management server 8. For example, the Goidanich model for downy mildew is used as this disease prediction model. The Goidanich model is a model for predicting the growth of downy mildew fungi and is sometimes used for disease prediction. The disease prediction model is configured to output the risk of downy mildew fungus disease (disease risk information) as a percentage when temperature, humidity, and precipitation are given. This risk value expressed as a percentage is used as a prediction value indicating the predicted environmental state.

[0126] The processing unit 32 may determine the prediction period using disease risk information instead of the sunshine / rain forecast and predicted precipitation. In this case, when acquiring environmental forecast information (step S33 in FIG. 13 ), the processing unit 32 of the management server 8 accesses the weather information server 7 to acquire weather forecast information. As described above, the weather forecast information includes a sunshine / rain forecast, predicted precipitation, predicted temperature, and predicted humidity for a period of approximately one week to ten days. The processing unit 32 provides the predicted precipitation, predicted temperature, and predicted humidity included in the weather forecast information to the disease prediction model, and causes the disease prediction model to output disease risk information.

[0127] 19 is a diagram showing an example of disease risk included in environmental prediction information. Fig. 19 shows disease risk information for the nine-day period from September 21st to September 29th.

[0128] Here, the processing unit 32 determines that days when the disease risk information is less than 20% are periods in which there are no problems with disease (third period), and determines that days when the disease risk information is 20% or more are periods in which there are problems with disease.

[0129] In the case of Fig. 19, the disease risk information for the five days from September 21 to September 25 is 10% or less. Therefore, the processing unit 32 determines that the five days from September 21 to September 25 are the third period. The processing unit 32 also determines that the four days from September 26 to September 29 are a period in which there are problems with disease.

[0130] The processing unit 32 performs the same process as the determination using the first period and the second period in the above embodiment. That is, the processing unit 32, in principle, sets the overlapping period in which the first period and the third period overlap as the prediction period. Furthermore, if there is no overlapping period between the first period and the third period and the third period exists before the first period, the processing unit 32 sets the day in the third period that is closest to the first period as the prediction period.

[0131] 19 , if the first period is from September 26 to September 29, the processing unit 32 sets the prediction period to September 25. In this way, prediction information can also be obtained using disease risk information related to fruit diseases as environmental prediction information.

[0132] In the above embodiment, the prediction process is performed based on two types of measured quality values ​​(sugar content and acidity) obtained by the quality value acquisition process, but the prediction process can be performed using at least two measured quality values, and in addition to sugar content and acidity, pH and polyphenol content can also be used in the prediction process. As shown in this embodiment, it is particularly preferable to use sugar content and acidity as the measured quality values ​​used in the prediction process.

[0133] Furthermore, in the quality value acquisition process in the above embodiment, the case where the measured quality value is calculated based on the first quality value (output of the first measurement device) of the first measurement device 2 and the second quality value based on the output of the second measurement device 4 has been exemplified, but the second quality value from the second measurement device 4 may also be used as the measured quality value. In other words, the prediction process may be performed using the second quality value from the second measurement device 4. In this case, it is not necessary to perform quality measurement work using the first measurement device 2, and the second measurement device 4 may be used to measure the quality of fruits K for some or all of the multiple trees T, and the prediction process may be performed using the obtained second quality value.

[0134] Furthermore, when performing prediction processing using the second quality value obtained by the second measurement device 4, the second quality value may be obtained after adjusting the second calibration curve data C2 depending on the time and place. In this case, the accuracy of the quality measurement can be further improved.

[0135] In addition, in this embodiment, spectroscopic analysis using near-infrared light has been used as an example of a method for obtaining fruit quality values, but fruit quality values ​​may also be obtained by other methods as long as similar quality values ​​can be obtained.

[0136] The scope of the present invention is defined by the claims, not by the meaning described above, and is intended to include meanings equivalent to the claims and all modifications within the scope thereof.

[0137] DESCRIPTION OF SYMBOLS 1 Harvest support system 2 First measuring device (first measuring unit) 4 Second measuring device (second measuring unit) 5 Measuring device 6 Agricultural machine 6a In-vehicle network 7 Weather information server 8 Management server (processing device) 10 Management terminal 12 Manipulator 14 Operator 18 Control device 19 Input / output device 20 Communication device 22 First measuring head 24 First control unit 26 Second measuring head 26a Light projecting unit 26b Light receiving unit 26c Head body 26c1 Annular tip surface 26c2 Hole 27 Light source 28 Spectrometer 30 Second control unit 32 Processing unit 32a Quality value acquisition process 32b Prediction process 34 Memory unit 36 ​​Communication device 40 Quality value database 40a Sugar content database 40b Acidity database 40c pH database 40d Polyphenol database A Imaging area BS Wireless base station C1, C11, C12, C13, C14 First calibration curve data C2, C21, C22, C23, C24 Second calibration curve data D1 First interval D2 Second interval F Field K Fruit KT Target fruit KR Reference fruit L Diagram L10 Prediction curve L11 Approximation curve L20 Prediction curve L21 Approximation curve NW Public network S, S1, S2, S3, S4 Scaling data T Tree d0, d1, d2 Date k1 Fruit grain

Claims

1. A harvesting support system comprising: a measuring device for measuring the quality of fruit grown in a field; and a processing device, wherein the processing device has a processing unit which executes a first process of acquiring multiple types of quality values ​​based on the output of the measuring device, a second process of acquiring multiple time series data in which each of the multiple types of quality values ​​is accumulated in a chronological order, and a third process of calculating a predicted period for which the fruit is optimally harvested based on the multiple time series data.

2. The harvest support system of claim 1, wherein the third process includes: a process of calculating a predicted quality value for each of the multiple types of quality values ​​based on the multiple time series data; a process of acquiring environmental prediction information indicating an environmental prediction in the field; and a process of calculating the prediction period based on the environmental prediction information and the multiple predicted quality values.

3. The harvesting support system of claim 2, wherein the environmental prediction information includes a predicted value indicating a predicted environmental state, and the process of determining the predicted period includes a process of determining a first period in which the plurality of predicted quality values ​​satisfy a predetermined condition, and a process of determining a second period in which the environmental state is determined to be suitable for harvesting based on a comparison result between the predicted value and a predetermined threshold value, and a process of determining the predicted period based on the first period and the second period.

4. The harvest support system according to claim 3, wherein the process of determining the predicted period based on the first period and the second period includes a process of determining an overlapping period in which the first period and the second period overlap as the predicted period.

5. The harvesting support system of claim 4, wherein the process of determining the predicted period based on the first period and the second period further includes a process of, when a period that is not the second period exists immediately before the overlapping period, providing an interval of at least one day between the period that is not the second period and the predicted period.

6. A harvest support system as described in any one of claims 2 to 5, wherein the environmental prediction information includes at least one of meteorological information and disease risk information relating to diseases of the fruit.

7. A harvesting support system as claimed in any one of claims 3 to 6, wherein the process of determining the predicted period based on the first period and the second period includes a process of, when there is no overlapping period between the first period and the second period, determining the day in the second period that is closest to the first period as the predicted period.

8. A harvesting support system as claimed in any one of claims 3 to 7, wherein the process of determining the predicted period further includes a process of outputting the plurality of predicted quality values ​​to the outside when the first period does not exist.

9. A harvesting support system as described in any one of claims 1 to 8, wherein the measuring device comprises: a first measuring unit that measures the quality of the fruit; and a second measuring unit that measures the quality of the fruit with a configuration different from that of the first measuring unit, and the first process includes a process of determining the multiple types of quality values ​​when the quality of the fruit is measured by the first measuring unit based on scaling data indicating a correlation between the output of the first measuring unit and multiple reference quality values ​​based on the output of the second measuring unit.

10. The harvesting support system according to any one of claims 1 to 9, wherein the fruit includes grapes.

11. A harvesting support system as described in any one of claims 1 to 10, wherein the multiple types of quality values ​​include at least two of sugar content, acidity, pH, and polyphenol content.

12. A processing device comprising a processing unit that executes the following processes: a process of acquiring multiple types of quality values ​​based on the output of a measuring device that measures the quality of fruit grown in a field; a process of acquiring multiple time series data in which each of the multiple types of quality values ​​is accumulated in a chronological order; and a process of calculating a predicted period during which the fruit is optimally harvested based on the multiple time series data.

13. A computer program for causing a computer to execute a process for assisting in the harvesting of fruit grown in a field, the computer program causing the computer to execute the steps of: acquiring multiple types of quality values ​​based on the output of a measuring device for measuring the quality of the fruit; acquiring multiple time series data in which each of the multiple types of quality values ​​is accumulated in a chronological order; and calculating a predicted period for which the fruit is optimally harvested based on the multiple time series data.

Images