Farmland management device, farmland management method, and program

Abstract

SOLUTION: A farm field management device comprises: a.n acquisition unit that acquires a first measurement value of a first point measured by a first sensor installed at the first point in a farm field, and a second measurement value of a second point measured by a second sensor installed at the second point in the farm field; a determination unit that determines, based on the first measurement value and the second measurement value, whether to move a mobile robot equipped with a third sensor to a third point between the first point and the second point and cause the mobile robot to measure a third measurement value of the third point with the third sensor; and an instruction unit that instructs the mobile robot to measure the third point when the determination unit determines that the third measurement value of the third point is to be measured.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present invention relates to a farmland management device, a farmland management method, and a program. [Background technology]

[0002] Patent document 1 discloses a leaf surface environment sensor that detects the illuminance (or sunlight intensity) on the surface of a leaf, or the temperature and humidity (either temperature or humidity, or both) on the backside, leaf color, and the concentration of carbon dioxide emitted from the leaf. [Prior art document] [Patent documents] [Patent Document 1] JP 2022-100732 A Summary of the Invention

[0003] A farm field management device according to one aspect of the present invention may include an acquisition unit that acquires a first measurement value of a first point within a farm field measured by a first sensor installed at the first point and a second measurement value of a second point within the farm field measured by a second sensor installed at the second point. The farm field management device may include a determination unit that determines, based on the first measurement value and the second measurement value, whether to move a mobile robot equipped with a third sensor to a third point between the first point and the second point and cause the mobile robot to measure a third measurement value of the third point with the third sensor. The farm field management device may include an instruction unit that instructs the mobile robot to measure the third point when the determination unit determines that the third measurement value of the third point should be measured.

[0004] In the farm field management device, the judgment unit may determine to cause the mobile robot to measure the third measurement value at the third location using the third sensor if the difference between the first measurement value and the second measurement value is greater than or equal to a threshold value.

[0005] Any of the farm field management devices may further include an estimation unit that estimates the third measurement value of the third location by interpolation between the first measurement value and the second measurement value when the difference between the first measurement value and the second measurement value is smaller than a threshold value.

[0006] In any of the farm field management devices, the judgment unit may determine to have the mobile robot measure the third measurement value at the third location using the third sensor if the difference between the first measurement value or the second measurement value measured in the same time period is greater than or equal to a threshold value.

[0007] Any of the farm field management devices may further include an estimation unit that estimates the third measurement value of the third location by interpolation between the first measurement value and the second measurement value when the difference between the first measurement value or the second measurement value measured in the same time period is smaller than a threshold value.

[0008] In any of the farm field management devices, the estimation unit may perform the interpolation using a learned prediction model in which the temperature, humidity, and light intensity at the first location and the second location measured by the first sensor and the second sensor, respectively, are used as explanatory variables, and the temperature, humidity, and light intensity at the third location measured by the third sensor are used as objective variables.

[0009] In any of the farm field management devices, the instruction unit may instruct the mobile robot to measure the third location so that the greater the difference between the first measurement value and the second measurement value, the more frequently the third measurement value at the third location is measured by the third sensor.

[0010] In any of the farm field management devices, the judgment unit may determine that, if the difference between the first measurement value and the second measurement value is greater than or equal to a threshold value, to cause the mobile robot to measure the third measurement value at each of the plurality of third locations using the third sensor.

[0011] The instruction unit may instruct the mobile robot to measure the third locations such that the greater the difference between the first measurement value and the second measurement value, the greater the number of third locations the mobile robot is instructed to measure.

[0012] In any of the farm land management devices, the instruction unit may determine the position of the third point based on the first measurement value and the second measurement value.

[0013] Any of the farm land management devices may further include a generation unit that generates a measurement value distribution of the farm field based on the first measurement value, the second measurement value, and the third measurement value.

[0014] In any of the farm field management devices, the measurement value distribution may include a distribution of at least one of the temperature, humidity, electrical conductivity, and pH of the root portion of the plant, the surface temperature of the stem, leaf, and fruit portions of the plant, the ambient temperature, humidity, light intensity, and carbon dioxide concentration.

[0015] In any of the farm field management devices, the instruction unit may instruct an environmental control equipment in the farm field to adjust at least one of the water content, electrical conductivity, hydrogen ion exponent of the soil in the farm field, the temperature, humidity, amount of light, and carbon dioxide concentration in the farm field, so that the plants are in a predetermined growth state, based on a predictive model that indicates the relationship between the measurement value distribution and the occurrence of damage to the plants or the harvest results of the plants.

[0016] In any of the farm land management devices, the instruction unit may identify the growth state of the plant based on three-dimensional position information of the plant obtained from a detection result by an optical sensor present in the farm field.

[0017] In any of the farm land management devices, the optical sensor may be mounted on the mobile robot.

[0018] A farm field management method according to one aspect of the present invention may include acquiring a first measurement value of a first point within a farm field measured by a first sensor installed at the first point, and a second measurement value of a second point within the farm field measured by a second sensor installed at the second point. The farm field management method may include determining, based on the first measurement value and the second measurement value, whether to move a mobile robot equipped with a third sensor to a third point between the first point and the second point, and have the mobile robot measure a third measurement value of the third point with the third sensor. If it is determined in the determining step that the third measurement value of the third point should be measured, the farm field management method may include instructing the mobile robot to measure the third point.

[0019] A program according to one aspect of the present invention may cause a computer to function as an acquisition unit that acquires a first measurement value of a first point in a field measured by a first sensor installed at the first point and a second measurement value of a second point in the field measured by a second sensor installed at the second point. The program may cause the computer to function as a determination unit that determines, based on the first measurement value and the second measurement value, whether to move a mobile robot equipped with a third sensor to a third point between the first point and the second point and cause the mobile robot to measure a third measurement value of the third point with the third sensor. The program may cause the computer to function as an instruction unit that instructs the mobile robot to measure the third point when the determination unit determines that the third measurement value of the third point should be measured.

[0020] The above summary of the invention does not list all of the features of the present invention, and subcombinations of these features may also be inventions. [Brief explanation of the drawings]

[0021] [Figure 1] FIG. 1 is a diagram showing a mobile robot moving through a field where plants are cultivated. [Figure 2] 1 is a diagram illustrating an example of the overall configuration of a farm land management system according to an embodiment of the present invention. [Figure 3] FIG. 1 is a diagram showing an example of a cultivation shelf. [Figure 4] FIG. 2 is a diagram illustrating an example of functional blocks of the farm land management device. [Figure 5] 10 is a flowchart illustrating an example of a procedure for determining whether to obtain measurement values ​​between points by estimation or by measurement using a mobile robot. [Figure 6] FIG. 2 illustrates an example of a hardware configuration. DETAILED DESCRIPTION OF THE INVENTION

[0022] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the scope of the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0023] FIG. 1 shows a mobile robot 10 moving through a field where plants 50, such as vegetables or fruits, are cultivated. The mobile robot 10 is a vehicle that moves on the ground. The mobile robot 10 may also be an air vehicle, such as an unmanned aerial vehicle, that moves through the air, or a ship that moves on water. The mobile robot 10 may move between cultivation shelves 60 where the plants 50 are cultivated.

[0024] In this embodiment, the field is an artificial light type plant factory that uses artificial light such as LEDs or incandescent lamps as a light source to cultivate the plants 50. However, the field may also be a sunlight type plant factory that uses sunlight as a light source to cultivate the plants 50.

[0025] The mobile robot 10 includes an arm 12 and a sensor unit 20 attached to the tip of the arm 12. The arm 12 may be a multi-joint arm unit rotatably attached to the main body of the mobile robot 10. The sensor unit 20 includes various sensors that measure the environmental conditions around the plant 50 and the growth condition of the plant 50. The sensor unit 20 includes various sensors that measure the temperature, humidity, light intensity, and carbon dioxide concentration of the stem, leaf, and fruit of the plant 50. The sensor unit 20 may include an optical sensor that measures the growth condition of the plant 50. The optical sensor may be a camera or a laser scanner. The camera may be a twin-lens 3D camera or a ToF (Time of Flight) camera. Three-dimensional position information of the plant 50 may be generated using the detection results from the optical sensor.

[0026] The cultivation shelf 60 is also equipped with various sensors that measure the temperature, humidity (water content), electrical conductivity (EC), and hydrogen ion exponent (pH) of the root part (culture medium) of the plant 50, as well as various sensors that measure the surface temperature, ambient temperature, humidity, light intensity, and carbon dioxide concentration of the stem, leaf, and fruit parts of the plant 50.

[0027] In order to accurately grasp the environmental conditions around each plant 50 and the growth state of the plant 50, it is preferable to provide various sensors to each plant 50. However, providing various sensors to each plant 50 increases costs. Furthermore, when data is transmitted from a large number of sensors, communication may become congested, and the burden on the device that processes the data may increase.

[0028] Therefore, the farm land management system according to this embodiment reduces the number of sensors installed in the farm land while suppressing a decrease in accuracy in understanding the environmental conditions around each plant 50 and the growth conditions of the plants 50.

[0029] 2 shows an example of the overall configuration of a farmland management system according to this embodiment. The farmland management system includes a farmland management device 100, a mobile robot 10, multiple sensor units 30, a sensor management device 300, a nutrient solution supply system 400, a light source system 410, an air conditioning system 420, a ventilation system 430, and a carbon dioxide supply system 440. The nutrient solution supply system 400, the light source system 410, the air conditioning system 420, the ventilation system 430, and the carbon dioxide supply system 440 are examples of environmental control systems. Each of the multiple mobile robots 10 may move in different or the same areas of the farmland.

[0030] 3 shows an example of the installation location of the sensor unit 30. The sensor unit 30 is provided around some of the planting pots 62 of all the plants 50 on the cultivation shelf 60. The sensor unit 30 may be provided at equal intervals for each of the planting pots 62. In other words, the sensor unit 30 is not provided for each of all the planting pots 62 on the cultivation shelf 60. Part of the sensor unit 30 may be provided on the stems or leaves of the plants 50. Part of the sensor unit 30 may be provided within the culture medium.

[0031] The sensor unit 30 includes various sensors that measure the temperature, humidity (water content), electrical conductivity, and hydrogen ion exponent of the root portion (culture medium portion) of the plant 50, and various sensors that measure the temperature, humidity, light intensity, and carbon dioxide concentration of the stem, leaf, and fruit portions of the plant 50. The sensor unit 30 may perform measurements periodically at predetermined time intervals.

[0032] The farm land management device 100 manages various devices in a plant factory. The farm land management device 100 is connected to and communicates with a mobile robot 10, a sensor management device 300, a nutrient solution supplying device 400, a light source device 410, an air conditioning device 420, a ventilation device 430, and a carbon dioxide supplying device 440 via a network 80. The farm land management device 100 may also be connected to and communicate with a plurality of sensor units 30 via the network 80.

[0033] The farm land management device 100 and the sensor management device 300 may be computers having a central processing unit (CPU) and memory. The nutrient solution supply equipment 400, the light source equipment 410, the air conditioning equipment 420, the ventilation equipment 430, and the carbon dioxide supply equipment 440 may each be equipped with a computer having a central processing unit (CPU) and memory.

[0034] The computer may be a personal computer, tablet computer, smartphone, workstation, server computer, general-purpose computer, or a computer system in which multiple computers are connected. Such a computer system is also a computer in a broad sense. The computer may be a dedicated computer designed for environmental control in a plant factory, or may be dedicated hardware realized by dedicated circuits. The computer may be implemented in a virtual computer environment. When a computer is used, the farm field management device 100, the sensor management device 300, the nutrient solution supply equipment 400, the light source equipment 410, the air conditioning equipment 420, the ventilation equipment 430, and the carbon dioxide supply equipment 440 are realized by the computer executing a program.

[0035] The field management device 100 controls the mobile robot 10, the sensor management device 300, the nutrient solution supply equipment 400, the light source equipment 410, the air conditioning equipment 420, the ventilation equipment 430, and the carbon dioxide supply equipment 440, thereby controlling the environment within the plant factory according to the growth state of the plants 50.

[0036] The sensor management device 300 collects various measurement values ​​from each of the multiple sensor units 30 and provides the collected values ​​to the farm land management device 100. The farm land management device 100 may be equipped with the sensor management device 300.

[0037] The nutrient solution supplying equipment 400 supplies a nutrient solution containing fertilizer components such as potassium or calcium via a pump to the cultivation shelves 60. The nutrient solution supplying equipment 400 may adjust the fertilizer concentration and amount of the nutrient solution in response to instructions from the farm field management device 100.

[0038] The light source equipment 410 includes a light source that emits artificial light, such as an LED or an incandescent lamp, that is installed on the cultivation shelf 60, and irradiates the plants 50 with the artificial light. The light source equipment 410 may control the light intensity and irradiation period of the light source in response to instructions from the farm field management device 100. If the farm field is a sunlight-based plant factory, the farm field management system may include, instead of the light source equipment 410, a solar radiation control device that controls the opening and closing of curtains installed on windows to adjust the amount of sunlight irradiating the plants 50.

[0039] The air conditioning equipment 420 adjusts the temperature and humidity of the air in the indoor space of the plant factory and circulates the temperature- and humidity-adjusted air within the indoor space. The air conditioning equipment 420 may control the temperature and humidity of the indoor space in response to instructions from the farm land management device 100.

[0040] The ventilation equipment 430 includes a circulator or an electric fan that supplies air to the indoor space of the plant factory. The ventilation equipment 430 may control the amount and direction of air supplied to the indoor space in response to an instruction from the farm land management device 100.

[0041] The carbon dioxide supplying equipment 440 supplies carbon dioxide from a carbon dioxide tank into the room of the plant factory. The carbon dioxide supplying equipment 440 may control the amount of carbon dioxide supplied to the room space in response to an instruction from the farm land management apparatus 100.

[0042] 4 shows an example of functional blocks of the farm land management device 100. The farm land management device 100 includes an acquisition unit 102, a determination unit 104, an instruction unit 106, an estimation unit 110, a generation unit 120, and a storage unit 130. A CPU included in the farm land management device 100 may function as the acquisition unit 102, the determination unit 104, the instruction unit 106, the estimation unit 110, and the generation unit 120.

[0043] The acquisition unit 102 acquires each measurement value measured by each sensor unit 30 installed in the field (plant factory). The acquisition unit 102 may periodically acquire each measurement value measured by each sensor unit 30 installed in the field at predetermined intervals. The acquisition unit 102 accumulates each measurement value in the storage unit 130. For example, the acquisition unit 102 acquires a measurement value M1 at point P1 measured by a sensor unit 30 (P1) installed at point P1 in the field, and a measurement value M2 at point P2 measured by a sensor unit 30 (P2) installed at point P2 in the field.

[0044] 3, points P1 and P2 are located at positions spaced apart with a plurality of planting pots 62 in between. That is, between the sensor units 30, there is at least one plant 50 that is not a measurement target of the sensor unit 30.

[0045] Based on the measurement values ​​M1 and M2, the judgment unit 104 judges whether to move the mobile robot 10 equipped with the sensor unit 20 to a point P3 between points P1 and P2 and have the mobile robot 10 measure the measurement value M3 at point P3 using the sensor unit 20.

[0046] If the difference between the measurement values ​​M1 and M2 is smaller than the threshold, the estimation unit 110 estimates the measurement value M3 of the point P3 by interpolation between the measurement values ​​M1 and M2. If the difference between the measurement values ​​M1 and M2 is smaller than the threshold, the estimation unit 110 may estimate the measurement values ​​of multiple points including the point P3 by interpolation between the measurement values ​​M1 and M2. On the other hand, if the difference between the measurement values ​​M1 and M2 is equal to or greater than the threshold, the estimation unit 110 may not be able to accurately estimate the measurement value M3 of the point P3 between the measurement values ​​M1 and M2. Therefore, the determination unit 104 may determine to have the mobile robot 10 measure the measurement value M3 of the point P3 with the sensor unit 20.

[0047] If the difference between measurement values ​​M1 and M2 is equal to or greater than a threshold, the instructing unit 106 may instruct the mobile robot 10 to measure multiple points between points P1 and P2. The instructing unit 106 may instruct the mobile robot 10 to measure point P3 so that the greater the difference between measurement values ​​M1 and M2, the more frequently the sensor unit 20 measures measurement value M3 at point P3. If the difference between measurement values ​​M1 and M2 is between a first threshold and a second threshold, the instructing unit 106 may instruct the mobile robot 10 to measure point P3 at a first frequency per unit period (e.g., one day, one hour, etc.), and if the difference between measurement values ​​M1 and M2 is equal to or greater than a second threshold, the instructing unit 106 may instruct the mobile robot 10 to measure point P3 at a second frequency per unit period (e.g., one day, one hour, etc.) that is greater than the first frequency. The instruction unit 06 may instruct the mobile robot 10 to measure multiple points between points P1 and P2 so that the greater the difference between the measurement values ​​M1 and M2, the greater the number of points measured by the sensor unit 20 between points P1 and P2.

[0048] The instruction unit 106 instructs the mobile robot 10 to measure point P3 via the network 80. The instruction unit 106 may instruct the mobile robot 10 to measure point P3, which is the midpoint between points P1 and P2.

[0049] The instruction unit 106 may determine the position of point M3 based on the measurement values ​​M1 and M2. If either measurement value M1 or measurement value M2 is outside a predetermined measurement range, the instruction unit 106 may determine that either measurement value M1 or measurement value M2 is an abnormal value, and may determine as point P3 a point closer to the point determined to be an abnormal value than the midpoint between points P1 and P2. If either measurement value M1 or measurement value M2 differs from the average value of previous measurement values ​​in the same time period by more than a predetermined percentage, the instruction unit 106 may determine that either measurement value M1 or measurement value M2 is an abnormal value, and may determine as point P3 a point closer to the point determined to be an abnormal value than the midpoint between points P1 and P2. The instruction unit 106 may compare measurement values ​​M1 and M2 with the average values ​​of measurement values ​​measured in the same time period by sensor units 30 at points other than points P1 and P2, and determine as point P3 a point closer to point P1 or point P2, whichever has a larger difference from the average value, than the midpoint between points P1 and P2. The instruction unit 106 may determine multiple points P3 so that more measurement points are closer to the point P1 or point P2 that is determined to be an abnormal value than to a midpoint between the points P1 and P2, and may instruct the mobile robot 10 to measure each of the points P3.

[0050] If the difference between the measurement values ​​M1 or M2 measured in the same time period is equal to or greater than a threshold, the determination unit 104 may determine to cause the mobile robot 10 to measure the measurement value M3 of point P3 with the sensor unit 20. If the difference between the measurement values ​​M1 or M2 measured in the same time period is smaller than the threshold, the estimation unit 110 may estimate the measurement value M3 of point P3 by interpolating between the measurement values ​​M1 and M2.

[0051] The estimation unit 110 may perform interpolation using a learned prediction model in which the temperature, humidity, and light intensity of points P1 and P2 measured by sensor unit 30 (P1) and sensor unit 30 (P2), respectively, are used as explanatory variables, and the temperature, humidity, and light intensity of point P3 measured by sensor unit 20 are used as objective variables.

[0052] The generation unit 120 may perform machine learning according to a supervised learning algorithm using the temperatures, humidity, and light intensity of points P1 and P2 measured by the sensor unit 30 (P1) and the sensor unit 30 (P2), respectively, as explanatory variables, and the temperature, humidity, and light intensity of point P3 measured by the sensor unit 20 as objective variables, to generate a trained prediction model that predicts the temperature, humidity, and light intensity of point P3 from the temperatures, humidity, and light intensity of points P1 and P2, and store the trained model in the storage unit 130. The algorithm may be any type of algorithm, such as a neural network, a support vector machine, multiple regression analysis, or a decision tree.

[0053] The generation unit 120 generates a measurement value distribution of the indoor space of the plant factory based on the measurement values ​​M1, M2, and M3. The generation unit 120 generates the measurement value distribution of the indoor space of the plant factory based on the measurement values ​​at each point measured by the sensor unit 20 and the sensor unit 30 and the measurement values ​​at each point estimated by the estimation unit 110. The measurement value distribution may include at least one distribution of the temperature, humidity (water content), electrical conductivity, and pH of the root portion of the plant 50, the surface temperature of the stem, leaf, and fruit portion of the plant 50, the ambient temperature, humidity, light intensity, and carbon dioxide concentration. The measurement value distribution may include at least one distribution of the temperature, humidity (water content), electrical conductivity, and pH of the root portion of the plant 50, and at least one distribution of the surface temperature, ambient temperature, humidity, light intensity, and carbon dioxide concentration of the stem, leaf, and fruit portion of the plant 50. The measurement value distribution may include the respective distributions of temperature, humidity, electrical conductivity, and pH of the root portion of the plant 50, as well as the respective distributions of surface temperature, ambient temperature, humidity, light intensity, and carbon dioxide concentration of the stem, leaf, and fruit portions of the plant 50.

[0054] Based on a prediction model showing the relationship between the distribution of measurement values ​​in the indoor space of the plant factory and the occurrence of damage to the plants 50 or the harvest results of the plants 50, the instruction unit 106 instructs at least one of the nutrient solution supply equipment 400, light source equipment 410, air conditioning equipment 420, ventilation equipment 430, and carbon dioxide supply equipment 440 in the field to adjust at least one of the water content, electrical conductivity, hydrogen ion exponent of the soil in the field, the temperature, humidity, light intensity, and carbon dioxide concentration in the field so that the plants 50 are in a predetermined growth state. The instruction unit 106 may instruct at least one of the nutrient solution supply equipment 400, light source equipment 410, air conditioning equipment 420, ventilation equipment 430, and carbon dioxide supply equipment 440 in the field to adjust at least one of the water content, electrical conductivity, hydrogen ion exponent, temperature, humidity, light intensity, and carbon dioxide concentration of the soil in the field so that the plant 50 reaches a predetermined growth state, based on a prediction model showing the relationship between the measurement value distribution of the root portion of the plant 50 and the measurement value distribution of the stem, leaf, and fruit portion of the plant 50 and the occurrence of damage to the plant 50 or the harvest result of the plant 50. The instruction unit 106 may identify the growth state of the plant 50 based on three-dimensional position information of the plant 50 obtained by detection results from an optical sensor such as a camera or laser scanner present in the field. The sensor unit 20 provided in the mobile robot 10 may have an optical sensor such as a camera or laser scanner for generating three-dimensional position information.

[0055] The generation unit 120 may perform machine learning according to a supervised learning algorithm, using the cultivation condition data indicating each measurement value distribution as an explanatory variable and the cultivation result data indicating the occurrence of damage to the plant 50 or the harvest result of the plant 50 as a target variable, to generate a trained prediction model that predicts the occurrence of damage to the plant 50 or the harvest result of the plant 50 from each measurement value distribution, and store the trained model in the storage unit 130. The algorithm may be any type of algorithm, such as a neural network, a support vector machine, multiple regression analysis, or a decision tree.

[0056] The cultivation result data may include at least one of damage during cultivation of the plant 50 and the harvest result of the plant 50. The damage during cultivation of the plant 50 may include at least one of physiological damage such as fruit end rot and fruit cracking, damage caused by disease, and damage caused by pests. The data indicating damage to the plant 50 may include at least one of the presence or absence of damage, the type of damage, the frequency of occurrence (e.g., the percentage of damaged plants in a cultivation unit), and the range. The harvest result of the plant 50 may include at least one of the weight, number, and quality (sugar content, moisture content, etc.) of the harvested plants.

[0057] The generation unit 120 may include a preprocessing unit 122 , a class estimation unit 124 , a model generation unit 126 , and a model update unit 128 .

[0058] The preprocessing unit 122 performs preprocessing on at least one of the plurality of cultivation condition data and the plurality of cultivation result data stored in the storage unit 130, and supplies the preprocessed data to the class estimation unit 124. The preprocessing unit 122 may perform preprocessing for learning.

[0059] The class estimation unit 124 classifies the plurality of cultivation condition data into a plurality of classes. The class estimation unit 124 may supply the plurality of classified cultivation condition data to the model generation unit 126, the estimation unit 110, and the model update unit 128.

[0060] The model generation unit 126 generates a trained model that predicts the cultivation result data from the cultivation condition data, using the cultivation condition data and the cultivation result data, and stores the trained model in the storage unit 130. The model update unit 128 updates the trained model, using the cultivation condition data and the cultivation result data.

[0061] The pre-processing unit 122 may associate the cultivation condition data with the cultivation result data according to the acquisition time of the measurement values ​​from the sensor unit 20 and the sensor unit 30. For example, the pre-processing unit 122 may associate the cultivation condition data with the cultivation result data for each acquisition date, or for different acquisition dates, or for each same time interval on the same acquisition date, according to the acquisition time of the measurement values.

[0062] The preprocessing unit 122 may perform interpolation for missing data such as time-series measurement values. For periods where no data exists, the preprocessing unit 122 may use an interpolation algorithm such as linear interpolation or spline interpolation to interpolate data. If the difference between the data and the average value of multiple data exceeds a threshold predetermined by the user, the preprocessing unit 122 may delete the data as an outlier or change the value to the same as the data immediately preceding or following it on the time axis. The preprocessing unit 122 may perform rounding processing such as truncating the data to a predetermined number of digits. The preprocessing unit 122 may discretize the cultivation condition data and cultivation result data and associate them with the cultivation area for each cultivation period (sowing, raising seedlings, planting, and greening).

[0063] The preprocessing unit 122 may perform preprocessing to extract feature quantities of the cultivation condition data and the cultivation result data. The preprocessing unit 122 may extract, as feature quantities, at least one of an integrated value, a differential value, an average value, a variance value, and data obtained by separating daytime and nighttime components from at least one of the cultivation condition data and the cultivation result data. The preprocessing unit 122 may calculate, as feature quantities, an integrated value of data for each predetermined period. For example, the preprocessing unit 122 may extract, as feature quantities, an integrated value of the number of occurrences of faults, temperature, humidity, or the like every three hours. By calculating the integrated value in this way, delayed effects can be identified. The preprocessing unit 122 may calculate differential values ​​of data at multiple data acquisition times and extract the differential values ​​as feature quantities. The preprocessing unit 122 may regard data that changes over time as a composite wave of daytime and nighttime components and separate one of the daytime and nighttime components by interpolation (e.g., spline interpolation) or a frequency decomposition method (e.g., Fourier transform). Then, the preprocessing unit 122 may separate the difference between one of the separated daytime and nighttime components and the sensor data as the other of the daytime and nighttime components.

[0064] The preprocessing unit 122 may divide the preprocessed data into intervals such as the cultivation period or time zone (day and night), and create a dataset using the preprocessed data for each interval (mean value, variance value, etc.) as explanatory variables. The preprocessing unit 122 supplies the preprocessed dataset to the class estimation unit 124.

[0065] The class estimation unit 124 classifies (clusters) the preprocessed plurality of cultivation condition data into a plurality of classes. The class estimation unit 124 may classify a plurality of identifiers of the plants 50 into a plurality of classes (groups) having similar sets of cultivation condition data. The class estimation unit 124 may classify the data based on at least one of the similarity of the data, such as time series and feature amounts. For example, the class estimation unit 124 may classify data obtained in the same time period on different acquisition dates into the same class. Furthermore, the class estimation unit 124 may classify data whose similarity between vectorized data is higher than a threshold (for example, the distance is smaller than a threshold) into the same class. For example, the class estimation unit 124 may classify corresponding data sets whose extracted feature amounts have a difference equal to or smaller than a threshold into the same class.

[0066] The class estimation unit 124 may perform classification using the k-means method, probabilistic latent semantic analysis (pLSA), or the like.

[0067] The class estimation unit 124 may create a class estimation model using the classification result as the objective variable and the preprocessed dataset as the explanatory variable. The class estimation model may be created using machine learning such as a Bayesian network. The class estimation unit 124 may estimate the class of the cultivation condition data using the class estimation model in a later operation of outputting the prediction result. Furthermore, the class estimation unit 124 may use class information used to generate the trained model in a later operation of outputting the prediction result.

[0068] After clustering, the class estimation unit 124 may supply a portion of the cultivation condition data and a portion of the corresponding cultivation result data to the model generation unit 126 as data for model generation. After clustering, the class estimation unit 124 may supply another portion of the cultivation condition data and another portion of the corresponding cultivation result data to the estimation unit 110 as data for reliability calculation. The class estimation unit 124 may separate the data for model generation and the data for reliability calculation in order to calculate the reliability using cross-validation or the like. The class estimation unit 124 may randomly separate the data for model generation and the data for reliability calculation. Furthermore, the class estimation unit 124 may separate the data for model generation and the data for reliability calculation by data acquisition period (e.g., day and night, day, month).

[0069] The model generation unit 126 generates a trained model by performing machine learning using the classification results (classified data). The model generation unit 126 may receive, from the class estimation unit 124, some of the cultivation condition data indicating various measurement values ​​acquired by the acquisition unit 102 and the plurality of cultivation result data identified from the various measurement values ​​as data for model generation, and generate a trained model using the data for model generation. The model generation unit 126 may generate a model with a Bayesian network model structure using the plurality of cultivation condition data and the plurality of cultivation result data classified into multiple classes. The model generation unit 126 may also generate other machine learning models, such as a neural network.

[0070] Furthermore, if a trained model is already stored in the storage unit 130, the model update unit 128 updates the trained model using the classification results (classified data). The model update unit 128 may perform machine learning in the same way as the model generation unit 126, and update the trained model having a Bayesian network model structure. The model update unit 128 may compare the cultivation results of the plant 50 under the executed cultivation conditions with the cultivation results predicted from the cultivation conditions by the trained model, and update the trained model.

[0071] The estimation unit 110 may use the trained model to predict at least one of a defect in the cultivation of the plant 50 and a cultivation result. The estimation unit 110 may receive data for reliability calculation including the classification result (classified data) from the class estimation unit 124, and use the trained model to predict at least one of a defect in the cultivation of the plant 50 and a cultivation result from the data. The estimation unit 110 may use the trained model to predict one defect or cultivation result that has the highest probability or whose probability exceeds a threshold, from cultivation conditions corresponding to one identifier.

[0072] The estimation unit 110 may calculate the reliability for each of multiple classes using the predicted failure or cultivation result and the reliability calculation data used for the prediction. For a class estimated to include the cultivation conditions used for the prediction, the estimation unit 110 may calculate the reliability (accuracy rate) by comparing the actual failure or cultivation result for the cultivation conditions with the predicted failure or cultivation result. The estimation unit 110 may calculate the reliability for each identifier and calculate one final reliability for each class from multiple reliabilities. For one class, the estimation unit 110 may calculate the average or total value of multiple reliabilities calculated for multiple identifiers as the final reliability for that class. The estimation unit 110 may calculate the reliability using at least one of the recall, precision, and F-measure.

[0073] FIG. 5 is a flowchart showing an example of a procedure for determining whether to obtain the measured values ​​between points by estimation or by measurement using the mobile robot 10.

[0074] The acquisition unit 102 acquires measurement values ​​detected by various sensors from the sensor units 30 installed at each point on the cultivation shelf 60 (S100). The determination unit 104 acquires the difference between the target measurement values ​​(S102). Information indicating the pair of measurement values ​​from which the determination unit 104 derives the difference, i.e., the pair of points, may be stored in advance in the storage unit 130. The pair of points may be a pair of adjacent sensor units 30. The pair of points may be a pair of sensor units 30 adjacent to each other along the longitudinal direction of the cultivation shelf 60. The pair of points may be the pair of points that are closest to each other among pairs of points between which there is at least one plant 50 that is not the target of measurement by the sensor unit 30.

[0075] If the difference between the pair of measurement values ​​of interest is equal to or greater than the threshold value ("Y" in S104), the instruction unit 106 identifies at least one additional measurement point based on the pair of measurement values ​​of interest (S106). The instruction unit 106 may identify a midpoint between the points of the pair of measurement values ​​of interest as the additional measurement point. The instruction unit 106 may identify an additional measurement point based on each value of the pair of measurement values ​​of interest. The instruction unit 106 may identify a point of the pair of measurement values ​​of interest that is closer to the point of the measurement value that has a larger difference from the average measurement value of the same point in the same time period in the past as the additional measurement point. The instruction unit 106 may identify multiple additional measurement points that are equally spaced between the points of the pair of measurement values ​​of interest.

[0076] The instruction unit 106 instructs the mobile robot 10 to measure the identified additional measurement point (S106). The memory unit 130 may store each planting pot 62 on the cultivation shelf 60 in association with a number for uniquely identifying each planting pot 62. The mobile robot 10 may retain map information in its memory indicating the position of the planting pot 62 corresponding to the number. The instruction unit 106 may instruct the mobile robot 10 to measure the additional measurement point by outputting a measurement instruction command indicating the number corresponding to the additional measurement point to the mobile robot 10.

[0077] Upon receiving the instruction, the mobile robot 10 may move to the additional measurement point and measure the temperature, humidity, electrical conductivity, and pH of the root portion of the target plant 50, as well as the temperature, humidity, light intensity, and carbon dioxide concentration of the stem, leaf, and fruit portions of the plant 50, using the sensor unit 20. The mobile robot 10 may identify the position of the plant 50 using a camera provided on the arm 12, and control the arm 12 to perform measurements using the various sensors of the sensor unit 20. The acquisition unit 102 acquires the measurement values ​​of the various sensors at the additional measurement point from the sensor unit 20 (S110).

[0078] If the difference between the pair of measurement values ​​of interest is less than the threshold value ("N" in S104), the estimation unit 110 estimates the measurement values ​​of additional measurement points between the measurement values ​​by interpolation based on the pair of measurement values ​​of interest (S112).

[0079] The judgment unit 104 judges whether additional measurement values ​​have been measured or estimated for all pairs of measurement values ​​measured in the same time period (S114), and if judgment has not been made for all pairs of measurement values, the judgment unit 104 repeats the processing from step S102 onwards.

[0080] As described above, according to the farm land management device 100 of this embodiment, when the difference between the measurement values ​​is large, the mobile robot 10 moves to another measurement point between the measurement points and actually measures that point using the sensor unit 20. On the other hand, when the difference between the measurement values ​​is small, the measurement value of the other measurement point between the measurement points is estimated by interpolation instead of measuring using the mobile robot 10. Therefore, it is possible to reduce the number of sensor units installed in the farm land while suppressing a decrease in the accuracy of understanding the environmental conditions around each plant 50 and the growth conditions of the plants 50.

[0081] 6 illustrates an example of a computer 1200 that may embody aspects of the present embodiment in whole or in part. A program installed on the computer 1200 may cause the computer 1200 to perform operations associated with an apparatus according to an embodiment of the present invention or to function as one or more “parts” of the apparatus. Alternatively, the program may cause the computer 1200 to execute the operations or one or more “parts.” The program may cause the computer 1200 to execute a process or steps of a process according to an embodiment of the present invention. Such a program may be executed by the CPU 1212 to cause the computer 1200 to perform specific operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.

[0082] The computer 1200 according to this embodiment includes a CPU 1212 and a RAM 1214, which are interconnected by a host controller 1210. The computer 1200 also includes a communication interface 1222 and an input / output unit, which are connected to the host controller 1210 via an input / output controller 1220. The computer 1200 also includes a ROM 1230. The CPU 1212 operates according to programs stored in the ROM 1230 and RAM 1214, thereby controlling each unit.

[0083] The communication interface 1222 communicates with other electronic devices via a network. A hard disk drive may store programs and data used by the CPU 1212 in the computer 1200. The ROM 1230 stores a boot program executed by the computer 1200 upon activation and / or programs dependent on the computer's hardware. The programs may be provided via a computer-readable storage medium such as a CD-ROM, a USB memory, or an IC card, or via a network. The programs may be installed in the RAM 1214 or the ROM 1230, which are also examples of computer-readable storage media, and executed by the CPU 1212. The information processing described in these programs is read by the computer 1200, resulting in cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be configured by implementing operations or processing of information in accordance with the use of the computer 1200.

[0084] For example, when communication is performed between computer 1200 and an external device, CPU 1212 may execute a communication program loaded in RAM 1214 and instruct communication interface 1222 to perform communication processing based on the processing described in the communication program. Under the control of CPU 1212, communication interface 1222 reads transmission data stored in a transmission buffer area provided in RAM 1214 or a storage medium such as a USB memory, and transmits the read transmission data to a network, or writes reception data received from the network to a reception buffer area or the like provided on the storage medium.

[0085] The CPU 1212 may also cause all or a necessary portion of a file or database stored in an external storage medium such as a USB memory to be read into the RAM 1214, and perform various types of processing on the data in the RAM 1214. The CPU 1212 may then write the processed data back to the external storage medium.

[0086] Various types of information, such as various types of programs, data, tables, and databases, may be stored in the storage medium and subjected to information processing. The CPU 1212 may perform various types of processing on data read from the RAM 1214, including various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, information search / replacement, etc., as described throughout this disclosure and specified by the instruction sequences of the programs, and write the results back to the RAM 1214. The CPU 1212 may also search for information in a file, database, etc. in the storage medium. For example, if multiple entries, each having an attribute value of a first attribute associated with an attribute value of a second attribute, are stored in the storage medium, the CPU 1212 may search for an entry that matches a condition specified by the attribute value of the first attribute from among the multiple entries, read the attribute value of the second attribute stored in the entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0087] The above-described programs or software modules may be stored in a computer-readable storage medium on or near the computer 1200. Also, a storage medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer-readable storage medium, thereby providing the programs to the computer 1200 via the network.

[0088] A computer-readable medium may include any tangible device capable of storing instructions that are executed by an appropriate device. As a result, the computer-readable medium with instructions stored thereon comprises an article of manufacture, including instructions that can be executed to create means for performing the operations specified in the flowchart or block diagram. Examples of computer-readable media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, etc. More specific examples of computer-readable media may include floppy disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray (RTM) disc, memory stick, integrated circuit card, etc.

[0089] The computer-readable instructions may include either source code or object code written in any combination of one or more programming languages. The source code or object code includes conventional procedural programming languages. The conventional procedural programming languages ​​may be assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, or object-oriented programming languages ​​such as Smalltalk®, JAVA®, C++, etc., and the “C” programming language or similar programming languages. The computer-readable instructions may be provided to a processor or programmable circuitry of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus locally or over a wide-area network (WAN) such as a local area network (LAN), the Internet, etc. The processor or programmable circuitry may execute the computer-readable instructions to create means for performing the operations specified in the flowcharts or block diagrams. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, etc.

[0090] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications and improvements can be made to the above embodiments. It is clear from the claims that such modifications and improvements can also be included within the technical scope of the present invention.

[0091] It should be noted that the execution order of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not specifically stated as "before," "prior to," etc., and that the processes can be performed in any order unless the output of a previous process is used in a later process. Even if the operational flow in the claims, specifications, and drawings is described using "first," "next," etc. for convenience, this does not mean that the processes must be performed in this order. [Explanation of symbols]

[0092] 10 Mobile Robot 12 Arm 20, 30 sensor unit 30 Sensor Unit 50 plants 60 cultivation rack 62 Planting pot 80 Network 100 Farm field management device 102 Acquisition Department 104 Judgment Department 106 Instruction section 110 Estimation part 120 Generation part 122 Pretreatment section 124 Class Estimation Unit 126 Model Generation Unit 128 Model Update Section 130 Storage section 300 Sensor management device 400 Nutrient solution supply equipment 410 Light source equipment 420 Air conditioning equipment 430 Ventilation equipment 440 Carbon dioxide supply equipment 1200 Computer 1210 host controller 1212 CPU 1214 RAM 1220 Input / Output Controller 1222 communication interface 1230 ROM

Claims

1. an acquisition unit that acquires a first measurement value at a first point in a farm field measured by a first sensor installed at the first point, and a second measurement value at a second point in the farm field measured by a second sensor installed at the second point; a determination unit that determines, based on the first measurement value and the second measurement value, whether to move a mobile robot equipped with a third sensor to a third point between the first point and the second point and cause the mobile robot to measure a third measurement value at the third point using the third sensor; an instruction unit that instructs the mobile robot to measure the third point when the determination unit determines that a third measurement value of the third point is to be measured; A farm field management device comprising:

2. 2. The farm management device according to claim 1, wherein the judgment unit determines to cause the mobile robot to measure the third measurement value at the third location using the third sensor when the difference between the first measurement value and the second measurement value is equal to or greater than a threshold value.

3. 2. The farmland management device according to claim 1, further comprising an estimation unit that, when a difference between the first measurement value and the second measurement value is smaller than a threshold, estimates the third measurement value of the third point by interpolation between the first measurement value and the second measurement value.

4. 2. The farm management device of claim 1, wherein the judgment unit determines to have the mobile robot measure the third measurement value at the third location using the third sensor when a difference between the first measurement value or the second measurement value measured in the same time period is greater than or equal to a threshold value.

5. 2. The farm management device according to claim 1, further comprising an estimation unit that, when a difference between the first measurement value or the second measurement value measured in the same time period is smaller than a threshold, estimates the third measurement value of the third location by interpolation between the first measurement value and the second measurement value.

6. 6. The farmland management device according to claim 3, wherein the estimation unit performs the interpolation using a trained prediction model in which the temperature, humidity, and light intensity at the first location and the second location measured by the first sensor and the second sensor, respectively, are used as explanatory variables, and the temperature, humidity, and light intensity at the third location measured by the third sensor are used as objective variables.

7. 3. The field management device of claim 2, wherein the instruction unit instructs the mobile robot to measure the third location so that the greater the difference between the first measurement value and the second measurement value, the more frequently the third measurement value at the third location is measured by the third sensor.

8. 2. The farm management device according to claim 1, wherein the judgment unit determines to cause the mobile robot to measure the third measurement value at each of the plurality of third locations using the third sensor when the difference between the first measurement value and the second measurement value is greater than or equal to a threshold value.

9. 9. The field management device of claim 8, wherein the instruction unit instructs the mobile robot to measure the third locations such that the greater the difference between the first measurement value and the second measurement value, the greater the number of third locations the mobile robot is instructed to measure.

10. The farm land management device according to claim 1 , wherein the instruction unit determines the position of the third point based on the first measurement value and the second measurement value.

11. The farmland management device according to claim 1 , further comprising a generation unit that generates a measurement value distribution of the farmland based on the first measurement value, the second measurement value, and the third measurement value.

12. 12. The farm field management device according to claim 11, wherein the measurement value distribution includes a distribution of at least one of the temperature, humidity, electrical conductivity, and pH of a root portion of a plant, the surface temperature of a stem, leaf, or fruit portion of the plant, the ambient temperature, humidity, light intensity, and carbon dioxide concentration.

13. 13. The farm field management device according to claim 12, wherein the instruction unit instructs an environmental control facility in the farm field to adjust at least one of the water content, electrical conductivity, and hydrogen ion exponent of soil in the farm field, the temperature, humidity, amount of light, and carbon dioxide concentration in the farm field, so that the plants are in a predetermined growth state, based on a prediction model that indicates a relationship between the measurement value distribution and the occurrence of damage to the plants or the harvest results of the plants.

14. The farm land management device according to claim 13 , wherein the instruction unit identifies the growth state of the plant based on three-dimensional position information of the plant obtained by a detection result from an optical sensor present in the farm land.

15. The farm land management device according to claim 14 , wherein the optical sensor is mounted on the mobile robot.

16. obtaining a first measurement value at a first location in the field measured by a first sensor installed at the first location and a second measurement value at a second location in the field measured by a second sensor installed at the second location; determining whether to move a mobile robot equipped with a third sensor to a third point between the first point and the second point and measure a third measurement value at the third point using the third sensor based on the first and second measurement values; instructing the mobile robot to measure the third point when it is determined that a third measurement value of the third point is to be measured in the determining step; A field management method comprising:

17. an acquisition unit that acquires a first measurement value at a first point in a farm field measured by a first sensor installed at the first point, and a second measurement value at a second point in the farm field measured by a second sensor installed at the second point; a determination unit that determines, based on the first measurement value and the second measurement value, whether to move a mobile robot equipped with a third sensor to a third point between the first point and the second point and cause the mobile robot to measure a third measurement value at the third point using the third sensor; an instruction unit that instructs the mobile robot to measure the third point when the determination unit determines that a third measurement value of the third point is to be measured; A program that allows a computer to function as a

Images