WO2014203664A1 - 収量予測システムおよび収量予測装置 - Google Patents
収量予測システムおよび収量予測装置 Download PDFInfo
- Publication number
- WO2014203664A1 WO2014203664A1 PCT/JP2014/063150 JP2014063150W WO2014203664A1 WO 2014203664 A1 WO2014203664 A1 WO 2014203664A1 JP 2014063150 W JP2014063150 W JP 2014063150W WO 2014203664 A1 WO2014203664 A1 WO 2014203664A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- field
- yield
- data
- unit
- yield prediction
- Prior art date
Links
- 238000005259 measurement Methods 0.000 claims abstract description 41
- 238000004364 calculation method Methods 0.000 claims description 32
- 238000010191 image analysis Methods 0.000 claims description 18
- 238000003860 storage Methods 0.000 claims description 15
- 238000004458 analytical method Methods 0.000 claims description 10
- 238000003306 harvesting Methods 0.000 abstract description 11
- 238000007689 inspection Methods 0.000 abstract 2
- 230000000875 corresponding effect Effects 0.000 description 66
- 238000000034 method Methods 0.000 description 51
- 238000012545 processing Methods 0.000 description 33
- 230000006870 function Effects 0.000 description 16
- 238000010586 diagram Methods 0.000 description 15
- 241000209094 Oryza Species 0.000 description 12
- 235000007164 Oryza sativa Nutrition 0.000 description 12
- 235000009566 rice Nutrition 0.000 description 12
- 238000009826 distribution Methods 0.000 description 10
- 230000005855 radiation Effects 0.000 description 9
- 238000000605 extraction Methods 0.000 description 8
- 238000001556 precipitation Methods 0.000 description 8
- 238000001228 spectrum Methods 0.000 description 7
- 239000003337 fertilizer Substances 0.000 description 6
- 239000000463 material Substances 0.000 description 5
- 230000003595 spectral effect Effects 0.000 description 5
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 4
- NBIIXXVUZAFLBC-UHFFFAOYSA-N Phosphoric acid Chemical compound OP(O)(O)=O NBIIXXVUZAFLBC-UHFFFAOYSA-N 0.000 description 4
- 239000000203 mixture Substances 0.000 description 4
- 238000007619 statistical method Methods 0.000 description 4
- 230000029777 axis specification Effects 0.000 description 3
- 230000002596 correlated effect Effects 0.000 description 3
- 238000009313 farming Methods 0.000 description 3
- 238000009472 formulation Methods 0.000 description 3
- ZLMJMSJWJFRBEC-UHFFFAOYSA-N Potassium Chemical compound [K] ZLMJMSJWJFRBEC-UHFFFAOYSA-N 0.000 description 2
- 229910000147 aluminium phosphate Inorganic materials 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 229910052757 nitrogen Inorganic materials 0.000 description 2
- 229910052700 potassium Inorganic materials 0.000 description 2
- 239000011591 potassium Substances 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 241000196324 Embryophyta Species 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000013479 data entry Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000002054 transplantation Methods 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01G—HORTICULTURE; CULTIVATION OF VEGETABLES, FLOWERS, RICE, FRUIT, VINES, HOPS OR SEAWEED; FORESTRY; WATERING
- A01G7/00—Botany in general
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Definitions
- the present invention relates to a system and apparatus for estimating the yield of agricultural products.
- Japanese public institutions and non-life insurance companies in other countries often adopt an agricultural disaster compensation system for major crops, and pay insurance money to farmers depending on the damage situation of the farmland.
- the investigator visually inspects the field that received the damage declaration from the farmer or the yield survey by actual measurement, and calculates the insurance money from the degree of damage.
- field surveys other than damaged fields to obtain the average yield of all fields, field surveys to estimate the yield of each field in advance, etc. are also conducted. This is a large and costly burden, and the aging of investigators is also a problem.
- the bias of evaluation results due to the intervention of human resources is also a major complaint of farmers.
- Patent Document 1 As a prior art for estimating the yield of crops using satellite images, a technique as shown in Patent Document 1 is known.
- the yield prediction formula is based on the correlation between the backscattering intensity in the sample field in the first half of the growing season obtained using the synthetic aperture radar image and the growth characteristics of paddy rice such as the number of stems obtained from the field survey.
- a rice yield prediction model generation method and a rice yield prediction method to be generated are disclosed.
- the yield estimation model differs depending on the presence or absence of damage, so there is a problem that accuracy is reduced when yield estimation is performed using the same model.
- a receiving unit that can receive an aerial image including a field, a storage unit that stores shape information of a plurality of fields including the field, and position information, respectively. Based on the received aerial image including the field, the shape information and position information of the field, an image analysis unit that calculates the image feature amount of the field, and the predicted yield of the field from the calculated image feature amount A yield estimation unit for calculating, based on a time-series pattern of weather data stored in advance, the yield estimation unit for each specific growth stage of the crop cultivated in the field A time series pattern analysis unit that outputs a first parameter group having a correlation with A predicted yield of the field is calculated using the image feature amount and the first parameter group.
- the present invention relates to a system for estimating crop yield and the like for each field.
- the field here refers to a farming area within a predetermined area, and includes spatially continuous farming areas that are assumed to have the same attribute information. Further, the present invention can be applied not only to yield but also to estimate quantitative parameters such as soil component values such as nitrogen, phosphoric acid and potassium, and plant height of agricultural crops.
- the parameter group that serves as a criterion for selecting the measured field to be investigated from the past image and field attribute information accumulated in the past and the aerial image of the field to be surveyed.
- the measured field is selected so that the parameter group has a variance as much as possible.
- the measured field candidates are selected as concentrated as possible. Also, by analyzing the time series pattern of weather data for each growth stage, a parameter group correlated with the growth situation is calculated, and the yield estimation using the image feature quantity, field attribute information and the parameter group as explanatory variables is performed. carry out.
- the present invention provides a group of parameters serving as a determination criterion for selecting an actually measured field to be investigated from an image of a survey area and field attribute information accumulated in the past and an aerial image obtained by photographing the field to be surveyed.
- the field candidates are selected so as to be concentrated as much as possible.
- the yield estimation model differs depending on the variety, planting date, damage, and weather data. Therefore, in the present invention, by analyzing the time series pattern of weather data for each growth stage, a parameter group correlated with the growth situation is calculated, and the image feature quantity, field attribute information, and the parameter group are used as explanatory variables. Yield estimation is performed.
- the investigator simply surveys the yield of the field designated by the system, and the remaining field is an aerial image of the survey area obtained by a sensor mounted on a satellite or aircraft. Therefore, it is possible to reduce the burden on the investigator and the cost for the survey and improve the accuracy of the yield estimation.
- FIG. 1 is a block diagram showing a basic configuration of a yield prediction system.
- the yield prediction system of this embodiment includes a yield prediction device 10.
- the yield prediction apparatus 10 uses the data acquired from the previous data 11 before the previous fiscal year, the image of the current year in which the field subject to yield estimation is taken from the image DB 12, and the field data subject to yield estimation from the field GIS 13. After receiving the field survey information selected from the field surveyor 15 and receiving the field survey results from the field surveyor 15, the current time series meteorological data was received from the weather DB 14. And a function of outputting the yield estimation result 17 later.
- the yield prediction apparatus 10 includes a parameter priority calculation unit 101, an image analysis unit 102, an actual field selection unit 103, field characteristic data 105, an actual field display unit 106, and actual measurement data.
- An input unit 107, measured field data 108, non-measured field data 109, a yield estimation unit 110, and a yield estimation result display unit 111 are provided.
- the parameter priority calculation unit 101 has a function of receiving data acquired from the previous data 11 before the previous fiscal year and transmitting the priority of parameters used for the determination material for selecting the measured field to the measured field selection unit 103. .
- the parameter priority calculation unit 101 includes a statistical analysis unit 1011, a priority calculation unit 1012, and statistical data 1013. Specific processing contents of each unit and data stored in the statistical data 1013 will be described later.
- the image analysis unit 102 receives the image of the current fiscal year in which the field that is the target of yield estimation was captured from the image DB 12, the field data that is the target of yield estimation from the field GIS 13, and was calculated from the image in the field characteristic data 105. Stores field spectral features.
- the image analysis unit 102 includes a field pixel extraction unit 1021 and a field feature calculation unit 1022. Specific processing contents of each unit will be described later.
- the measured field selection unit 103 receives the field spectrum feature quantity from the field feature data 105 and the parameter priority used for the determination material for selecting the measured field from the parameter priority calculation unit 101, and the measured field display unit 106 The information of the measured field selected from the fields existing in the survey area is transmitted.
- the measured field selection unit 103 includes a field selection axis determination unit 1031, a histogram creation unit 1032, an actual measurement candidate selection unit 1033, an actual cost calculation unit 1034, an actual field determination unit 1035, Is provided. Specific processing contents of each unit will be described later.
- the measured field display unit 106 displays the information on the measured field selected from the fields existing in the target area, received from the measured field selection unit 103, on the field investigator 15. A specific display screen example will be described later.
- the actual measurement data input unit 107 receives the actual measurement result from the field investigator 15 and the field spectrum feature quantity from the field characteristic data 105, and updates the data stored in the actual field data 108 and the non-measured field data 109.
- the yield estimation unit 110 includes the field data to be yield estimated from the field GIS 13, the time series pattern of the current year's weather data from the weather DB 14, and the field feature of the field from the measured field data 108 and the non-measured field data 109.
- the yield obtained by the actual measurement is received, the data stored in the non-actual field data 109 is updated, and the yield estimation end flag is transmitted to the yield estimation result display unit 111.
- the yield estimation unit 110 includes a growth stage estimation unit 1101, a time series pattern analysis unit 1102, a model formulation unit 1103, a yield estimation unit 1104, and a growth DB 1105. Specific processing contents of each unit will be described later.
- the yield estimation result display unit 111 After receiving the yield estimation end flag from the yield estimation unit 110, the yield estimation result display unit 111 receives the yield obtained by the actual measurement from the measured field data 108, and the yield obtained by the estimation from the non-measured field data 109. The yield estimation result 17 is output.
- the data stored in the past data 11, the field GIS 13, the weather DB 14, the field feature data 105, the measured field data 108, and the non-measured field data 109 will be described later.
- the yield prediction device 10 may include the past data 11, the image DB 12, the field GIS 13, and the weather DB 14.
- the image DB 12 stores image data obtained by a sensor mounted on a satellite or an aircraft that includes at least the shooting date and time and image position information in the header information.
- the shooting date / time represents when the image data was shot
- the image position information represents position information of the four corners of the image.
- the image position information may be, for example, latitude and longitude information used in the world geodetic system. Yield estimation is performed by mainly using aerial images included in the image DB 12.
- FIG. 2 is a sequence diagram showing an example of processing performed by the yield prediction apparatus 10. Specific processing will be described below.
- the field GIS 13 transmits the field ID and coordinate information of the field to be estimated for yield in the survey area to the image analysis unit 102 (S201).
- the image analysis unit 102 receives an image of the current year in which the field for yield estimation target is captured from the image DB 12 and calculates an image feature amount for each field for yield estimation target.
- the image feature quantity of the field for which the yield is to be estimated is stored in the field feature data 105 (S202).
- the field feature data 105 stores the image feature quantity of the yield estimation target field from the image analysis unit 102
- the field feature data 105 is transmitted to the measured field selection unit 103 (S204).
- the parameter priority calculation unit 101 receives data acquired from the previous data 11 before the previous fiscal year, and performs statistical analysis on the received data to determine the parameter priority used for the determination material for selecting the measured field. Calculate and transmit to the measured field selection unit 103 (S203). Note that the process of S201 ⁇ S202 ⁇ S204 and the process of S203 may be performed in parallel, or either process may be performed first.
- the measured field selection unit 103 receives information on the field to be estimated for yield in the survey area from the field GIS 13 (S205), receives the image feature amount from the field feature data 105, and receives the parameter priority from the parameter priority calculation unit 101.
- the measured field is selected, and information on the selected measured field is transmitted to the field investigator 15 through the measured field display unit 106 (S206).
- the field investigator 15 performs field measurement according to the received information regarding the field measured, and acquires the yield of each field (S207).
- the measured field data 108 and the non-measured field data 109 receive the field ID and the image feature amount from the field feature data 105 (S209).
- the yield of each field is received from the member 15 through the actual measurement data input unit 107 (S208).
- the measured field data 108 stores the pixel feature amount and yield corresponding to the field ID surveyed by the field investigator 15, and the non-measured field data 109 was not surveyed by the field surveyor 15.
- a field feature amount and yield corresponding to the field ID are stored. Since the yield stored in the non-actual field data 109 has no information at the stage of S208, NULL is stored. Thereafter, the yield estimation unit 110 receives the field ID, pixel feature amount, and yield from the measured field data 108 and the non-measured field data 109 (S210), and receives field data that is the target of yield estimation from the field GIS13. (S211), the time series pattern of the current year's weather data is received from the weather DB 14 (S212), and a yield estimation model is formulated. Thereafter, the yield of each field estimated in accordance with the established yield estimation model is stored in the measured field data 108 and the non-measured field data 109 (S213).
- the yield estimation unit 110 transmits a flag for which the yield estimation has been completed to the yield estimation result display unit 111 (S215), and the yield estimation result display unit 111 reads each field from the measured field data 108 and the non-measured field data 109. ID, yield, estimated yield, etc. are received (S214). Finally, the result is displayed by the yield estimation result display unit 111, and the process performed by the yield prediction device 10 is completed.
- FIG. 3 is a flowchart showing an example of processing performed by the parameter priority calculation unit 101.
- the parameter priority calculation unit 101 receives data acquired from the past data 11 before the previous fiscal year, and transmits the parameter priority used for the determination material for selecting the measured field to the measured field selection unit 103. Specific processing will be described below.
- S31 represents the start of processing of the parameter priority calculation unit 101.
- the data acquired from the previous data 11 and acquired before the previous fiscal year is analyzed by the statistical analysis unit 1011 and stored in the statistical data 1013.
- statistical values of parameters (1103 to 1111) of data acquired before the previous fiscal year are calculated for each year.
- the statistical value here means an average, variance, number of samples, histogram, maximum value, minimum value, and the like.
- the histogram here represents the frequency according to the class of the corresponding parameter.
- the method of determining the class is arbitrary, and for example, the area from the minimum value to the maximum value may be determined by dividing into ten. If the parameter is a discrete value such as damage or product type, only the average, variance, maximum value, and minimum value cannot be calculated, so NULL is stored.
- the priority calculation unit 1012 receives the annual statistical value of each parameter from the statistical data 1013, calculates the priority of each parameter, and selects a measured field from each parameter name and the calculated priority value pair. To the unit 103. Specifically, the annual distribution of each parameter is compared, and when there are many similar distributions, the priority is high, and when the annual distribution is significantly different, the priority is calculated to be low. For example, if the parameter is a continuous value, the total number of years I, the index of year i, average M i of the relevant parameters, a dispersion when combined all years V, when the number of samples and N i, priority The following (Equation 1) may be used.
- d (pi (x), pj (x)) is a scale for measuring the distance between the probability distributions pi (x), pj (x), and may be, for example, the amount of information of the Cullback librar.
- S34 represents the end of processing of the parameter priority calculation unit 101.
- FIG. 4 is a flowchart illustrating an example of processing performed by the image analysis unit 102.
- the image analysis unit 102 receives an aerial image of the current fiscal year in which the field that is the target of yield estimation was captured from the image DB 12, the field data that is the target of yield estimation from the field GIS 13, and is calculated from the image in the field characteristic data 105. Stores the feature quantity in the field. Specific processing will be described below.
- S41 represents the start of processing of the image analysis unit 102.
- S42 the pixel value of the aerial image included in the field is extracted from the aerial image received from the image DB 12 and the shape information of each field received from the field GIS13. A specific extraction method will be described later.
- S43 if the number of pixels extracted in the process of S42 is 1 or more, the process proceeds to S44, and if not, the process proceeds to S45.
- the field feature calculation unit 1022 calculates a feature amount from the set of extracted pixel values, the field ID 1051 corresponding to the field feature data 105, the pixel value in the field as the pixel value 1052, and the feature amount 1053.
- the feature quantity in the field calculated in is stored.
- the pixel value stored in the pixel value 1052 all pixel values in the field may be arranged, or an average pixel value in the field may be used.
- the feature amount 1053 is a parameter obtained by performing a predetermined calculation on the pixel value. For example, the feature amount is obtained by subtracting the red pixel value from the near infrared pixel value and the near infrared pixel value.
- the normalized vegetation index which is the ratio of the sum of the pixel value and the red pixel value, may be used as a feature quantity, or the set of pixel values included in the field is segmented and the average belonging to each group A parameter obtained by performing a predetermined calculation on the pixel value and a parameter obtained by dividing the number of pixels belonging to each group by the number of pixels in the field may be used as the feature amount.
- S45 since the number of extracted pixels is 0, the field ID 1051 corresponding to the field feature data 105 is stored, and NULL is stored in the pixel value 1052 and the feature value 1053.
- S46 end determination is performed. If S44 or S45 is performed for all fields included in the field GIS13, the process proceeds to S47. If there is a field where neither S44 nor S45 has been performed yet, the process proceeds to S42, and the next field is also in the field. Extract pixel values. S47 represents the end of the processing of the image analysis unit 102.
- FIG. 5 is a flowchart showing an example of processing performed by the measured field selection unit 103.
- the measured field selection unit 103 receives the field spectrum feature quantity from the field feature data 105 and the parameter priority used for the determination material for selecting the measured field from the parameter priority calculation unit 101, and the measured field display unit 106 The information of the measured field selected from the fields existing in the survey area is transmitted. Specific processing will be described below.
- S 51 represents the start of the process of the measured field selection unit 103.
- the storage data is received from the field feature data 105 and the field GIS 13, and the statistical value of this year for each parameter is calculated.
- the statistical value is the same as the data stored in the statistical data 1013.
- S53 a plurality of parameters that are regarded as important in actual measurement are selected. Even if a parameter whose priority received from the parameter priority calculation unit 101 is equal to or greater than a certain threshold and a parameter whose variance is equal to or greater than a certain threshold among the statistical values of each parameter calculated in S52 are selected. Alternatively, for each parameter, a value obtained by multiplying the priority and the variance value may be calculated, and a parameter that is equal to or greater than a certain threshold value may be selected. If the parameter for which the statistical value is calculated in S52 is not a continuous value but a discrete value, a statistical value such as entropy may be used instead of variance.
- the processes of S52 and S53 are performed by the field selection axis determination unit 1031.
- S54 a multivariate histogram with the plurality of parameters selected in S53 as axes is created. The process of S54 is performed by the histogram creation unit 1032.
- each axis of the multivariate histogram created in S53 is divided into a plurality of sections.
- the division method and the number of divisions are arbitrary.
- the area from the minimum value to the maximum value of each parameter may be divided into 10 equal parts, or the average or median value of each parameter is subtracted by a constant multiple of standard deviation.
- the area from the value to the value obtained by adding a constant multiple of the standard deviation may be divided into 10 equal parts.
- actual field candidates are selected from the multivariate histogram divided into a plurality of areas for each axis. Specific processing contents will be described later. The processes in S55 and S56 are performed by the actual measurement candidate selecting unit 1033.
- the actual measurement cost is calculated for all combinations of the actual field candidates selected in S56.
- the actual measurement cost can be determined according to the spatial distribution of the selected actual measurement field. Basically, the smaller the distance between the selected actual measurement fields, the smaller the actual measurement cost. When the distance is large, the actual measurement cost increases.
- a determinant of a covariance matrix calculated from a set of coordinate information of selected fields may be used as the actual measurement cost.
- the process of S57 is performed by the actual measurement cost calculation unit 1034.
- the measured field candidate that minimizes the calculated measured cost is selected as the final measured field, and is transmitted to the measured field display unit 106.
- the field list corresponding to the selected field ID among the data stored in the field GIS 13 is transmitted.
- the process of S58 is performed by the measured field determination unit 1035.
- S59 represents the end of the process of the measured field selection unit 103.
- FIG. 6 is a flowchart showing an example of processing performed by the yield estimation unit 110.
- Field data that is the target of yield estimation from the field GIS13, time series pattern of the current year's weather data from the weather DB14, spectral field features of the field from the measured field data 108 and non-measured field data 109, and the yield obtained by actual measurement.
- the data is received, the data stored in the non-measured field data 109 is updated, and the yield estimation end flag is transmitted to the yield estimation result display unit 111. Specific processing will be described below.
- S61 represents the start of processing of the yield estimation unit 110.
- time series weather data is received from the weather DB 14, and the stage of the growth stage is estimated for each field.
- the integrated temperature, integrated solar radiation, integrated precipitation, etc. are calculated for all the days from the date of rice planting to the date of image capture or harvest, and compared with the stored data in the growth DB 1105. . Thereby, it is estimated which growth stage corresponds to all days from the date of rice planting to the date of image capture or the date of harvest.
- the integrated temperature, the integrated solar radiation amount, the integrated precipitation amount and the like are explanatory variables, and the growth stage is an objective variable.
- a k-nearest neighbor method As an estimation method, a k-nearest neighbor method, a k-average method, a clustering method such as a mixed normal distribution, or the like may be applied.
- the meteorological data used may be meteorological data observed at a position closest to the corresponding field, or the meteorological data at the corresponding field position may be interpolated using meteorological data at a plurality of points.
- the process of S62 is performed by the growth stage estimation unit 1101.
- explanatory variables used for yield estimation are calculated from the time series weather data of the corresponding field.
- An explanatory variable is calculated by analyzing the time series pattern of each parameter of the meteorological data for each stage of the growth stage estimated in S61. Details will be described later. The process of S63 is performed by the time series pattern analysis unit 1102.
- a yield estimation model is formulated by using the spectral feature amount and yield of each field received from the measured field data 108 and explanatory variables based on the time series weather data calculated in S63.
- the explanatory variable in the yield estimation model is an explanatory variable based on the spectral feature quantity of each field received from the measured field data 108 and the time-series weather data calculated in S63, and the objective variable is the yield.
- the explanatory variable index is i
- the explanatory variable is Xi
- the objective variable is Y
- a linear model such as (Equation 3)
- a nonlinear model such as (Equation 4) or (Equation 5) is established.
- Each parameter may be obtained in a standing manner.
- a mixed model for each product type may be used, or a hierarchical Bayes model may be used.
- the process of S64 is performed by the model formulation unit 1103.
- S65 substitutes the explanatory variable based on the time-series meteorological data calculated in S63 with the spectral feature quantity of each field received from the measured field data 108 and the non-measured field data 109 into the yield estimation model formulated in S64. To estimate the yield of each field. The estimated yield for each field is stored in the estimated yield 1085 of the measured field data 108 and the non-measured field data 109. Thereafter, an end flag is transmitted to the yield estimation result display unit 111. The process of S65 is performed by the yield estimation unit 1104. S66 represents the end of the processing of the yield estimation unit 110.
- FIG. 7 shows an example of the data structure of the field GIS13.
- the field GIS 13 stores data 1301 to 1309 relating to the coordinates, shape, and attribute information of each field prepared in advance.
- the field ID 1301 is a label for identifying the field that is the target of yield estimation.
- the coordinate data 1302 is obtained by arranging the coordinates of each vertex when the shape of each field is a polygon.
- the coordinates stored in the coordinate data 1302 may be an arbitrary coordinate system. For example, the latitude and longitude of the world geodetic system may be stored.
- the fertilizer 1303 represents the amount of fertilizer applied to each field.
- the fertilizer 1303 may be, for example, the amount of nitrogen, phosphoric acid, or potassium applied before cultivation.
- the unit is arbitrary, and may be, for example, kilogram per 10 ares.
- the variety 1304 stores the variety name of the crop cultivated in each field. Variety 1304 may store a crop name.
- the damage 1305 stores the name of the damage that occurred in each field. When multiple damages occur in one field, the damage that has the greatest impact among the damages may be stored, or all the damages that have occurred may be stored.
- the standard yield 1306 stores a yield that can be averagely harvested in each field.
- the unit is arbitrary, and may be, for example, kilogram per 10 ares.
- the rice transplanting period 1307 the rice transplanting date of this year in each field is stored.
- the heading date 1308, the heading date of this year in each field is stored.
- the harvest date 1309 stores the harvest date of this year in each field.
- FIG. 8 shows an example of the data structure of the field feature data 105, the measured field data 108, and the non-measured field data 109.
- the field feature data 105 stores data 1051 to 1053 relating to pixel feature amounts and the like of each field.
- the measured field data 108 and the non-measured field data 109 store data 1081 to 1085 relating to the pixel feature amount, yield, and the like of each field.
- Field ID 1051 is a label for identifying the field that is the target of yield estimation, and matches field ID 1301 in field GIS13.
- the pixel value 1052 is a pixel value included in the corresponding field in the aerial image of the survey area. All the pixel values included in the field may be stored, or the average pixel value in the field may be stored.
- the feature amount 1053 is obtained by arranging parameters calculated by performing a predetermined calculation from the pixel value 1052. For example, a value obtained by subtracting a red pixel value from a near-infrared pixel value and a normalized vegetation index that is a ratio of a sum of a near-infrared pixel value and a red pixel value may be used as parameters. Segmentation is performed on the set of pixel values included in the parameter, and the parameters calculated by performing a predetermined operation on the average pixel value belonging to each group and the number of pixels belonging to each group are stored in the field. Parameters divided by the number of pixels may be arranged.
- Field ID 1081 is a label for identifying the field that is the target of yield estimation, and matches field ID 1301 in field GIS13.
- Pixel value 1082 which is a pixel value included in the corresponding field in the aerial image of the survey area, matches the pixel value 1052.
- the feature amount 1083 is an array of parameters calculated by performing a predetermined calculation from the pixel value 1052, and matches the feature amount 1053.
- Yield 1084 is the yield of the corresponding field.
- the unit is arbitrary, and may be, for example, kilogram per 10 ares. Since the field stored in the unmeasured field data 109 is a field that is not actually measured by the investigator, NULL is stored as the yield 1084 of the unmeasured field data 109.
- Estimated yield 1085 is the estimated yield of the corresponding field.
- the unit is arbitrary, and may be, for example, kilogram per 10 ares.
- the estimated yield 1085 stores the yield estimated by the yield estimating unit 110. Therefore, immediately after S209 is performed, NULL is stored, and the estimated yield after S213 is stored.
- FIG. 9 shows an example of the data structure of the weather DB 14.
- the weather DB 14 stores data 1400 to 1412 relating to time-series weather data observed in the survey area.
- the weather DB 14 stores data observed at a plurality of points.
- the observation ID 1400 is an ID for identifying an observation point.
- the observation place 1401 is coordinate information regarding the place of the corresponding observation point.
- the coordinates stored in the observation place 1401 may be an arbitrary coordinate system. For example, the latitude and longitude of the world geodetic system may be stored.
- the date 1402 represents the date when the weather data was observed.
- the average temperature 1403 represents the average temperature of the observation point within the corresponding date.
- the unit is arbitrary and may be, for example, ° C.
- the maximum temperature 1404 represents the maximum temperature of the observation point within the corresponding date.
- the unit is arbitrary and may be, for example, ° C.
- the minimum temperature 1405 represents the minimum temperature of the observation point within the corresponding date.
- the unit is arbitrary and may be, for example, ° C.
- the amount of solar radiation 1406 represents the average amount of solar radiation at the observation point on the corresponding date. As in the case of the temperature, the maximum solar radiation amount and the minimum solar radiation amount may be stored in the same manner.
- the unit is arbitrary, and may be, for example, W / m 2 .
- Humidity 1407 represents the average humidity at the observation point on the corresponding date. As in the case of the temperature, the maximum humidity and the minimum humidity may be stored in the same manner. The unit is arbitrary, and may be, for example,%.
- the atmospheric pressure 1408 represents the average atmospheric pressure at the observation point on the corresponding date. As in the case of the temperature, the maximum pressure and the minimum pressure may be stored in the same manner. The unit is arbitrary, and may be, for example,%.
- Precipitation 1409 represents the accumulated precipitation at the observation point on the corresponding date.
- the unit is arbitrary and may be, for example, mm.
- the sunshine time 1410 represents the sunshine time of the observation point on the corresponding date.
- the unit is arbitrary, and may be h, for example.
- the wind direction 1411 represents the wind direction of the observation point on the corresponding date.
- the wind speed 1412 represents the average wind speed at the observation point on the corresponding date.
- the instantaneous maximum wind speed and the like may be stored in the same manner.
- the unit is arbitrary, and may be, for example, m / s.
- the weather DB 14 may have a table for each observation point as in the above example, or the observation ID 1400 and the observation location 1402 may be added to the column to store all data in one table. .
- FIG. 10 shows an example of the data structure of past data 11, statistical data 1013, and growth DB 1105.
- the past data 11 stores data 1101 to 1112 such as attribute information and pixel feature amount for each field acquired before last year for each field.
- the field GIS 13 and the field feature data 105 coincide with each other, but the field GIS 13 and the field feature data 105 differ only in that they are acquired this year.
- the field ID 1101 is a label for identifying the field of the corresponding year that is the target of yield estimation, and matches the field ID 1301.
- the coordinate data 1102 is obtained by arranging the coordinates of each vertex when the shape of each field in the corresponding year is a polygon, and matches the coordinate data 1302.
- the fertilizer 1103 represents the amount of fertilizer applied to each field in the corresponding year and matches the fertilizer 1303.
- the variety 1104 stores the variety name of the crop cultivated in each field in the corresponding year, and matches the variety 1304.
- the damage 1105 stores the name of the damage that occurred in each field in the corresponding year and matches the damage 1305.
- the standard yield 1106 stores a yield that can be averagely harvested in each field in the corresponding year, and matches the standard yield 1306.
- the heading date 1108 stores the heading date for each field in the corresponding year, and matches the heading date 1308.
- the harvest date 1109 stores the harvest date in each field in the corresponding year, and matches the harvest date 1309.
- the pixel value 1110 is a pixel value included in the corresponding field in the aerial image in the survey area in the corresponding year, and matches the pixel value 1052.
- the feature value 1111 is a list of parameters calculated by performing a predetermined calculation from the pixel value 1110, and matches the feature value 1053.
- Year 1112 represents the year in which the corresponding data is acquired.
- the statistical data 1013 stores statistical data 10131 to 10138 for each year related to the past data 11.
- the year 10131 represents the acquisition year of the data for which the statistical value is calculated. This corresponds to the stored data 1112 of the past data 11.
- the parameter 10132 represents the type of parameter for calculating the statistical value. This corresponds to the column names of the stored data 1102 to 1111 of the past data 11.
- Average 10133 represents the average for the corresponding parameter for the corresponding year. Note that when the corresponding parameter corresponds to a discrete value such as a product type or damage, the average value cannot be calculated, and thus a NULL value is stored.
- Variance 10134 represents the variance for the corresponding parameter for the corresponding year. Similarly, when the corresponding parameter corresponds to a discrete value such as a product type or damage, since the variance cannot be calculated, a NULL value is stored. The number of samples 10135 represents the number of data related to the corresponding parameter in the corresponding year.
- the area definition 10136 represents a class definition area in the histogram regarding the corresponding parameter of the corresponding year. More specifically, when the corresponding parameter is a continuous value, the area definition 10136 corresponds to a combination of the maximum value and the minimum value of the definition area of each class by the number of divisions of the class. When the corresponding parameter is a discrete value, it corresponds to a value in which all the possible values of the corresponding parameter are arranged.
- the area A 10137 stores the number of data existing in the first area definition among the data regarding the corresponding parameter in the corresponding year.
- the area B10138 stores the number of data existing in the second area definition among the data regarding the corresponding parameter in the corresponding year.
- the growth DB 1105 stores, as a database, the relationship between each growth stage, accumulated temperature, accumulated solar radiation, accumulated precipitation, and variety.
- the growth stage 11051 represents each growth stage.
- the integrated temperature 11052 stores the integrated temperature at the corresponding growth stage.
- the unit is arbitrary and may be, for example, ° C.
- the accumulated solar radiation amount 11053 stores the cumulative solar radiation amount at the corresponding growth stage.
- the unit is arbitrary, and may be, for example, W / m 2 .
- the accumulated precipitation 11054 stores the accumulated precipitation at the corresponding growth stage.
- the unit is arbitrary and may be, for example, mm.
- the type 11055 represents the type information of the corresponding data. Information on the type of crop may be added here.
- FIG. 11 is an explanatory diagram of the process performed in S ⁇ b> 56 of the measured field selection unit 103. Specific processing contents will be described below.
- S56 the axis of the selected parameter is divided, and a predetermined number of data is extracted from each generated grid.
- FIG. 11 is an explanatory diagram when the selected parameters are the standard yield (axis 561) and the normalized vegetation index (axis 562).
- a scatter diagram 560 is created for the data obtained by combining the field feature data 105 and the field GIS 13 with respect to the axes 561 and 562. Then, the axes 561 and 562 are divided by a certain number (563).
- the division method and the number of division are arbitrary. For example, the area from the minimum value to the maximum value of the parameter may be divided into 10 equal parts, and from the value obtained by subtracting the standard deviation from the average, the median, and the standard deviation, to the value obtained by adding the constant from the average to the constant. The area may be divided into 10 equal parts. Since a plurality of grids 564 are generated by the division, a fixed number of data is extracted from each grid (565).
- the process proceeds to S57.
- the number of data to be extracted is 3, the total number of grids is 4, and the number of data existing in each grid is 10, 1, 2, 5, then 10! / (7! ⁇ 3!
- the second and third grids are one, and the fourth grid is 5! / (2! ⁇ 3!)
- the fourth grid is 5! / (2! ⁇ 3!)
- How to select these data is output as an actual field candidate. here,! Means factorial.
- FIG. 12 is an explanatory diagram of processing performed in the field pixel extraction unit 1021 (S42). Specific processing contents will be described below.
- FIG. 12 shows an example of extracting pixels included in the field 422 from the image 421.
- the pixel 423 is a set of pixels in the vicinity of the field 422.
- Position information of the four corners of the image 421 is upper left: (sx1, sy1), upper right: (sx2, sy1), lower left: (sx1, sy2), lower right: (sx2, sy2), and position information of each vertex of the field 422 Is (x1, y1), (x2, y2), (x3, y3), (x4, y4)...,
- the pixels included in the field are the output of the field pixel extraction unit 1021. Note that there are often pixels that straddle the field 422. In this case, it is added to the output only when the center of each pixel is included in the field.
- this extraction method may include pixels such as the road outside the field. Therefore, a buffer 424 having a fixed distance may be provided for the field 422, a new field 425 may be created, and the pixels in the field may be extracted.
- FIG. 13 is an explanatory diagram of processing performed in the time-series pattern analysis unit 1102 (S63). Specific processing contents will be described below.
- a graph 6301 is a plot of data with the date 6202 on the horizontal axis and the precipitation amount 6303 on the vertical axis. Since the growth stage is estimated for the date 6302, the time series data can be divided for each growth stage such as the rice planting period 6304.
- An image acquisition date 6305 is a shooting date of the aerial image acquired from the image DB 12, and is added to the header information of the data.
- a graph 6306 is a time series pattern obtained by removing high frequency components.
- the peak value 6307 of the time series pattern existing in each growth stage of the graph 6306, the width 6308 (for example, variance) of each peak, the peak position 6309, the integrated value 6310, etc. are output as explanatory variables used for yield estimation.
- the above parameters may be calculated separately before and after the image acquisition date 6305 and added as explanatory variables.
- a graph 6311 is a graph showing the time series change of the growth parameter for each field.
- the growth parameter referred to here is a parameter having a positive correlation with, for example, good growth, and for example, a normalized vegetation index calculated from an image may be used.
- a curve 6312 is a field growth curve of the field that has been damaged early, but the curve 6313 is a field growth curve of the field that has been steadily grown without being damaged.
- the curve 6313 is a field growth curve of the field that has been steadily grown without being damaged.
- parameters 6307 to 6310 calculated for each growth stage from the time series pattern are used as explanatory variables for yield estimation.
- yield prediction can be performed by simulation for various conditions such as varieties and cropping dates, and the results can be applied to farming support and the like.
- FIG. 14 is an example of an input screen of the actual measurement data input unit 107. Details of the input screen will be described below.
- a screen 10710 is a diagram displaying field polygons. When polygons are displayed, field polygons may be superimposed on an aerial image such as a satellite image. Field 10701 is the currently selected field, and field 10702 is the unselected field. Information on the field 10701 is displayed in a table 10703. Table 10703 displays the data stored in the field GIS 13 and the field measured yield 10704 investigated by the field investigator. Data can be input to the measured yield 10704 using a keyboard or the like.
- the range of the map displayed on the screen 10710 can be changed in the direction of the arrow.
- Each processing can be performed by selecting buttons 10710 to 10717.
- the button may be selected by simply touching the screen like a touch panel, or may be selected by a mouse cursor 10709 or keyboard input.
- the polygon display method displayed on the screen 10700 can be changed by selecting the button 10710.
- the polygon color and the thickness of the frame line can be changed according to the attribute values in the table 10703, or the polygon color and the frame line can be displayed in the field where the actual measurement yield 10704 has not been input. You may change the thickness etc.
- the field ID corresponding to the polygon is superimposed and displayed, but the type of attribute to be superimposed may be changed.
- the display scale of the screen 10700 can be enlarged.
- the display scale of the screen 10700 can be reduced.
- the field polygon can be searched for each attribute value.
- the screen 10700 displays the vicinity of the searched field.
- the button 10714 can newly add data to the field GIS 13, the measured field data 108, and the non-measured field data 109. Fields are added by adding to the table 10703 using a keyboard or the like.
- the button 10715 can update data to the field GIS 13, the measured field data 108, and the non-measured field data 109.
- the field is updated by editing the table 10703 with a keyboard or the like.
- the button 10716 can delete data in the field GIS 13, the measured field data 108, and the non-measured field data 109.
- the currently selected field data is deleted.
- the button 10717 can cancel data input. This completes the data entry by the field investigator.
- FIG. 15 is an example of an output screen of the yield estimation result display unit 111. Details of the output screen will be described below.
- Screen 11100, field 11101, 11102, table 11103, measured yield 11104, arrows 11105-11108, and buttons 11110-11113 are screen 10700, field 10701, 10702, table 10703, measured yield 10704, arrow 10705-10708, buttons 10710- It is the same as 10713.
- the estimated yield 11114 is data of the estimated yield 1085 stored in the measured field data 108 and the non-measured field data 109.
- the button 11115 performs accuracy evaluation of yield estimation.
- the field that has the measured yield 11104 and the estimated yield 11114 is only the field that is stored in the measured field data 108, and the accuracy is obtained by comparing the measured yield with the estimated yield for that field and calculating the statistical value.
- a correlation coefficient between the measured yield and the estimated yield may be calculated.
- the button 11116 generates a histogram regarding the measured yield 11104 and the estimated yield 11114.
- the histogram here is a graph representing the frequency according to the class of the corresponding parameter.
- the method of determining the class is arbitrary, and for example, the area from the minimum value to the maximum value may be determined by dividing into ten.
- the button 11117 outputs attribute names and attribute values included in the table 11103 as a CSV (Comma-Separated Values) file. At this time, the data to be output may be filtered by inputting the attribute value condition of the output data.
- CSV Common-Separated Values
- FIG. 19 is a block diagram illustrating a hardware configuration example of the yield prediction system.
- the yield prediction device 10 is a computer including an operation unit 1901, a display unit 1902, a processor 1903, a main memory 1904, and a storage device 1905.
- the processor 1903 executes a program stored in the main memory 1904.
- the main memory 1904 is a semiconductor memory, for example, and stores a program executed by the processor 1903 and data referred to by the processor 1903. Specifically, at least a part of the program and data stored in the storage device 1905 is copied to the main memory 1904 as necessary.
- the processor 1903 operates as a functional unit that realizes a predetermined function by operating according to the program of each functional unit.
- the processor 1903 functions as the yield estimation unit 110 by operating according to the yield prediction program.
- the yield prediction apparatus 10 is an apparatus including these functional units.
- the operation unit 1901 receives an input operation from the user.
- the operation unit may include, for example, a keyboard or a mouse.
- Display unit 1902 outputs information to the user.
- the display unit 1902 may be an image display device such as a liquid crystal display.
- the storage device 1905 is a non-volatile storage device such as a hard disk device (HDD) or a flash memory.
- the storage device 1905 of this embodiment stores at least the parameter priority calculation unit 101, the image analysis unit 102, the measured field selection unit 103, and the yield estimation unit 110.
- Each DB and storage unit may also be stored in the storage device 1905.
- Information such as a program that implements each function of the object recognition device 10 and a DB includes a storage device 1505, a nonvolatile semiconductor memory, a hard disk device, a storage device such as an SSD (Solid State Drive), an IC card, an SD card, It can be stored in a computer-readable non-transitory data storage medium such as a DVD.
- a storage device 1505 a nonvolatile semiconductor memory, a hard disk device, a storage device such as an SSD (Solid State Drive), an IC card, an SD card, It can be stored in a computer-readable non-transitory data storage medium such as a DVD.
- SSD Solid State Drive
- FIG. 20 is a sequence diagram showing an operation example of the present invention. A specific flow will be described below.
- the customer 304 places an order for yield estimation from the yield estimation contractor 303 (305).
- the field GIS 13 data held by the customer 304 is provided (306).
- the yield estimation company 303 makes an image photography request to the aerial photographer 302 (307).
- the shooting company 302 issues a shooting command to the satellite server 301 (308).
- image shooting 309 is started.
- the image is transmitted to the photographer 302 at the timing when the image is captured (310).
- an image is transmitted to the yield estimator 303 (311), and an image is also transmitted to the customer 304 (312).
- the customer 304 determines whether or not to estimate the yield with the captured image (313).
- the yield estimator 303 reorders the image capturing (315). Thereafter, the flow from 308 to 313 is repeated. If it is determined after 312 that the image is to be used (316), the yield estimation agent 303 performs the yield estimation using the yield estimation device 10 (317). Note that 306 may be performed after 316 and before 317. Thereafter, the result is transmitted to the customer 304 (318).
- the measured field is selected from the dispersion of data accumulated in the past, and the list of measured fields is transmitted to the field investigator.
- the measured field is selected using only the data of the year for which the yield estimation is performed.
- a time-series aerial image or altitude data is used to estimate the growth stage at the timing when the aerial image used for yield estimation is taken, and the growth stage is used as a parameter used when selecting a field.
- the second embodiment is the same as the first embodiment except for the parameter priority calculation unit 101 and the past data 11. Specifically, FIGS. 4 to 9 and 11 to 15 are also applied to the second embodiment. Hereinafter, the second embodiment will be described only with respect to differences from the first embodiment.
- FIG. 16 is a block diagram showing the basic configuration of the yield prediction system.
- the yield prediction system of this embodiment includes a yield prediction device 20.
- the yield prediction device 20 has captured the DEM (Digital Elevation Model) data of the target area from the elevation DB 21, the time series aerial image of the target area from the time series image DB 22, and the field that is the target of yield estimation from the image DB 12. After receiving the data of the field that is the target of yield estimation from the field GIS 13 for the image of the year, the function to send the field information selected as the actual measurement target to the field investigator 15 and the field survey result from the field investigator 15 A function of outputting the yield estimation result 17 after receiving the time series weather data of the current year from the weather DB 14 after receiving.
- DEM Digital Elevation Model
- the yield prediction apparatus 10 includes a growth stage classification unit 201, an image analysis unit 102, an actual field selection unit 103, field characteristic data 105, an actual field display unit 106, and actual data input.
- Image analysis unit 102 measured field selection unit 103, field feature data 105, measured field display unit 106, measured data input unit 107, measured field data 108, non-measured field data 109, and yield estimation unit 110
- the function of the yield estimation result display unit 111 is the same as that of the first embodiment.
- the growth stage classification unit 201 receives the DEM (Digital Elevation Model) data of the target area from the elevation DB 21 and the time series aerial image of the target area from the time series image DB 22, and selects the measured field to the measured field selection unit 103. It has a function to transmit the priority of parameters used at the time. In order to realize the above function, the growth stage classification unit 201 includes an air temperature fluctuation estimation unit 2011, an integrated temperature estimation unit 2012, a growth stage estimation unit 2013, and a field growth stage DB 2014. Specific processing contents will be described later.
- DEM Digital Elevation Model
- the altitude DB 21 stores DEM data of the target area
- the time-series image DB 22 stores time-series aerial images of the target area. Both data are data in the same format as the aerial image stored in the image DB 12. However, each pixel value of the DEM data represents an elevation value at the position coordinate.
- the unit is arbitrary, and may be m, for example.
- FIG. 17 is a flowchart showing an example of processing performed in the growth stage classification unit 201. Specific processing contents will be described below.
- S20101 represents the start of processing of the growth stage classification unit 201.
- time series weather data such as the temperature at each field location is converted to DEM data stored in the altitude DB 21, weather data stored in the weather DB 14, and time series aerial photography stored in the time series image DB 22. Estimated from the image.
- the observation location of the meteorological data stored in the meteorological DB 14 is not limited to the vicinity of the field, and even if the meteorological data corresponding to the nearest observation location of the corresponding field is used, it may differ from the actual meteorological data of the field. Many. Therefore, weather data from the rice planting date to the harvest date at the corresponding field position is estimated using DEM data and time-series aerial images.
- the time series meteorological data may be estimated by calculating ⁇ , ⁇ , and ⁇ by formulating the linear regression equation, and applying the above-described equation to all fields.
- a least square method or the like may be applied.
- the process of S20102 is performed by the temperature fluctuation estimation 2011.
- S20103 an integrated value of meteorological data from the date of planting and heading date of each field to the date of shooting the aerial image used for yield estimation is calculated.
- the process of S20103 is performed by the integrated temperature estimation unit 2012.
- the growth stage of each field is estimated using the weather data integrated value calculated in S20103.
- the estimation method may be the method described in S62.
- the field ID and the estimated growth stage are stored in the field growth stage DB 2014.
- a priority is calculated regarding the growth stage.
- the only parameter that exists is the growth stage, it is only necessary to determine whether or not to use it by looking at the values that the growth stage can take.
- the priority may be set to 0, and in other cases, the priority may be set to 1.
- the processing of S20104 and S20105 is performed by the growth stage estimation unit 2013.
- S20106 represents the end of the process of the growth stage classification unit 201.
- FIG. 18 shows an example of the data structure of the field growth stage data 2014.
- the field growth stage data 2014 stores the result of estimating the growth stage of each field when the aerial image acquired for yield estimation is taken.
- the field ID 20141 is a label for identifying the field that is the target of yield estimation, and matches the field ID 1301.
- the growth stage 20142 represents the growth stage of the corresponding field.
- the yield is estimated with high accuracy using aerial images and time-series weather data at a specific growth stage of a crop while suppressing the number of survey fields as much as possible. It becomes possible.
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Biodiversity & Conservation Biology (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Botany (AREA)
- Ecology (AREA)
- Game Theory and Decision Science (AREA)
- Forests & Forestry (AREA)
- Environmental Sciences (AREA)
- Agronomy & Crop Science (AREA)
- Animal Husbandry (AREA)
- Marine Sciences & Fisheries (AREA)
- Mining & Mineral Resources (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Image Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
さらに、全ほ場の平均的な収量を得るための被害ほ場以外における実測調査、各ほ場の基準となる単収をあらかじめ見積もるための実測調査なども行われ、調査員によるほ場実測、目視検査は多数回、広範囲におよび、コスト面で大きな負担となり、さらに調査員の高齢化も問題視されている。また人手を介在することによる評価結果の偏りも、農家の大きな不満となっている。
前記画像特徴量と前記第一パラメータ群を用いて前記ほ場の予測収量を算出する。
また収量に限らず、窒素、リン酸、カリウム等の土壌成分値や農作物の草丈等の定量的なパラメータを推定する際にも本発明を適用することができる。
図1は、収量予測システムの基本構成を表すブロック図である。
S41は、画像解析部102の処理開始を表す。
S42では、画像DB12から受信した空撮画像と、ほ場GIS13から受信した各ほ場の形状情報から、ほ場内に含まれる空撮画像の画素値を抽出する。具体的な抽出方法については、後述する。
S43では、S42の処理時に抽出された画素数が1以上であればS44へ進み、そうでなければS45へ進む。
S46では、終了判定を行う。ほ場GIS13に含まれる全てのほ場について、S44、もしくはS45を行った場合はS47に進み、まだS44、S45のいずれも行っていないほ場が存在する場合はS42に進み、次のほ場についてもほ場内画素値を抽出する。
S47は、画像解析部102の処理終了を表す。
S52では、ほ場特徴データ105とほ場GIS13から格納データを受信し、各パラメータに対する今年度の統計値を算出する。前記統計値は統計データ1013に格納されているデータと同様である。
S52、S53の処理は、ほ場選定軸決定部1031によって行われる。
S54では、S53にて選択された複数のパラメータを軸とする多変量ヒストグラムを作成する。
S54の処理は、ヒストグラム作成部1032によって行われる。
S56では、各軸について複数の区域に分割された多変量ヒストグラムから、実測ほ場候補の選定を行う。具体的な処理内容については、後述する。
S55、S56の処理は、実測候補選定部1033によって行われる。
S57の処理は、実測コスト算出部1034によって行われる。
S58の処理は、実測ほ場決定部1035によって行われる。
S59は、実測ほ場選択部103の処理終了を表す。
S62では、気象DB14から時系列気象データを受信し、各ほ場に関して生育ステージの時期を推定する。具体的には各ほ場に対し、田植日から画像撮影日もしくは収穫日までの全ての日について、積算気温、積算日射量、積算降水量等を算出し、生育DB1105の格納データとの比較を行う。これにより、田植日から画像撮影日もしくは収穫日までの全ての日が、どの生育ステージに相当するかを推定する。この際、積算気温、積算日射量、積算降水量等を説明変数、生育ステージが目的変数となる。推定方法としてk近傍法やk平均法、混合正規分布等のクラスタリング手法等を適用してもよい。なおこの際、用いる気象データは、対応するほ場から最も近い位置で観測される気象データでもよいし、複数地点の気象データを用いて対応するほ場の位置の気象データを補間してもよい。
S62の処理は、生育ステージ推定部1101によって行われる。
S63の処理は、時系列パターン解析部1102によって行われる。
S64の処理は、モデル策定部1103によって行われる。
S65の処理は、収量推計部1104によって行われる。
S66は、収量推計部110の処理終了を表す。
ほ場ID1301は、収量推計対象であるほ場を識別するためのラベルである。
座標データ1302は、各ほ場の形状を多角形とした時の、各頂点の座標を並べたものである。座標データ1302に格納される座標は任意の座標系でよく、例えば世界測地系の緯度、経度を格納してもよい。
品種1304は、各ほ場で栽培される作物の品種名が格納される。品種1304は、作物名も一緒に格納されてもよい。
被害1305は、各ほ場で起きた被害名が格納される。一つのほ場にて複数の被害が起きた際には、起きた被害の中で一番影響が大きいものを格納してもよいし、起きた被害全てを格納してもよい。
田植期1307は、各ほ場における今年度の田植日が格納される。
出穂揃日1308は、各ほ場における今年度の出穂揃日が格納される。
収穫日1309は、各ほ場における今年度の収穫日が格納される。
気象DB14は、調査地域にて観測された時系列気象データに関するデータ1400~1412を格納する。気象DB14は複数地点で観測されたデータを格納している。
観測ID1400は、観測地点を識別するためのIDである。
観測場所1401は、対応する観測地点の場所に関する座標情報である。観測場所1401に格納される座標は、任意の座標系でよく、例えば世界測地系の緯度、経度を格納してもよい。
平均気温1403は、対応する日付内での観測地点の平均気温を表す。単位は任意であり、例えば℃でもよい。
最高気温1404は、対応する日付内での観測地点の最高気温を表す。単位は任意であり、例えば℃でもよい。
最低気温1405は、対応する日付内での観測地点の最低気温を表す。単位は任意であり、例えば℃でもよい。
日射量1406は、対応する日付の観測地点の平均日射量を表す。なお気温の場合と同じく、最高日射量や最低日射量も同様に格納してもよい。単位は任意であり、例えばW/m2でもよい。
日照時間1410は、対応する日付の観測地点の日照時間を表す。単位は任意であり、例えばhでもよい。
風向1411は、対応する日付の観測地点の風向を表す。
風速1412は、対応する日付の観測地点の平均風速を表す。なお瞬間最大風速なども同様に格納してもよい。単位は任意であり、例えばm/sでもよい。
座標データ1102は、対応する年度の各ほ場の形状を多角形とした時の、各頂点の座標を並べたものであり、座標データ1302と一致する。
肥料1103は、対応する年度の各ほ場に対する施肥量を表し、肥料1303と一致する。
被害1105は、対応する年度に各ほ場で起きた被害名が格納され、被害1305と一致する。
基準単収1106は、対応する年度の各ほ場で平均的に収穫できると想定される収量が格納され、基準単収1306と一致する。
田植期1107は、対応する年度の各ほ場における田植日が格納され、田植期1307と一致する。
出穂揃日1108は、対応する年度の各ほ場における出穂揃日が格納され、出穂揃日1308と一致する。
収穫日1109は、対応する年度の各ほ場における収穫日が格納され、収穫日1309と一致する。
特徴量1111は、画素値1110から所定の演算をすることによって算出されたパラメータを並べたものであり、特徴量1053と一致する。
年1112は、対応するデータが取得される年度を表す。
統計データ1013は、過去データ11に関する各年の統計データ10131~10138を格納する。
年10131は、統計値を算出したデータの取得年を表す。過去データ11の格納データ1112に対応する。
平均10133は、対応する年の、対応パラメータに関する平均を表す。なお対応パラメータが品種や被害等の離散値に相当する場合は、平均を計算することができないために、NULL値を格納する。
サンプル数10135は、対応する年の、対応パラメータに関するデータ数を表す。
区域B10138は、対応する年の、対応パラメータに関するデータの中で、二番目の区域定義に存在するデータ数を格納する。同様に区域C、区域D・・・と、区域が階級数の分、存在しうる。例えば階級数が10だった場合、カラムとして区域Iまで存在する。これに対し、例えば対応パラメータが5種類しか存在しない離散値の場合、区域F~IにはNULL値を格納する。
生育ステージ11051は、各生育ステージを表す。
積算気温11052には、対応する生育ステージの時の積算気温が格納される。単位は任意であり、例えば℃でもよい。
積算日射量11053には、対応する生育ステージの時の積算日射量が格納される。単位は任意であり、例えばW/m2でもよい。
積算降水量11054には、対応する生育ステージの時の積算降水量が格納される。単位は任意であり、例えばmmでもよい。
品種11055は、該当データの品種情報を表す。なおここに作物の種類の情報を加えてもよい。
S56では選定されたパラメータの軸を分割し、生成された各グリッドから決められた数のデータを抽出する。図11は選定されたパラメータが基準単収(軸561)と、正規化植生指数(軸562)である場合の説明図である。
図12は画像421に対し、ほ場422内に含まれる画素を抽出する時の例である。画素423は、ほ場422近傍の画素の集合である。
画像421の四隅の位置情報を左上:(sx1、sy1)、右上:(sx2、sy1)、左下:(sx1、sy2)、右下:(sx2、sy2)とし、ほ場422の各頂点の位置情報を(x1、y1)、(x2、y2)、(x3、y3)、(x4、y4)・・・とすると、画像421の位置座標とほ場422の位置座標を比較することにより、画像内の各画素がほ場内に含まれるか否かを調べることができる。ほ場内に含まれる画素が、ほ場内画素抽出部1021の出力となる。なおほ場422をまたぐ画素が存在することが多々ある。この場合、各々の画素の中心がほ場内に含まれる場合にのみ、上記出力に加える。
グラフ6311はほ場ごとの生長パラメータの時系列変化をグラフに表したものである。ここでいう生長パラメータとは、例えば生育の良さと正の相関を持つパラメータで、例えば画像から算出された正規化植生指数等を用いてもよい。
ほ場10701は現在選択されているほ場であり、ほ場10702は選択されていないほ場である。ほ場10701の情報は表10703に表示される。表10703に表示されるのはほ場GIS13に格納されているデータと、現地調査員によって調査されたほ場の実測収量10704である。実測収量10704に、キーボード等を用いてデータ入力することが可能である。
以下、本発明の第2の実施形態について図面を参照して説明する。
S20103では、各ほ場の作付日や出穂日等から収量推計に用いる空撮画像の撮影日までの気象データ積算値を算出する。
S20103の処理は、積算気温推定部2012によって行われる。
S20104では、S20103で算出された気象データ積算値を用いて、各ほ場の生育ステージを推定する。推定方法は、S62に記述されている方法でもよい。その後、ほ場生育ステージDB2014にほ場IDと推定生育ステージを格納する。
S20105では、生育ステージに関して優先度を算出する。但し、存在するパラメータは生育ステージのみなので、生育ステージの取りうる値を見て、使うか否かだけを決定すればよい。例えば、ほ場生育ステージDB2014を見て、生育ステージが1種類しかない場合は、全く生育ステージにバラつきがないため、優先度を0とし、それ以外の場合は優先度1としてもよい。
S20104、S20105の処理は生育ステージ推定部2013によって行われる。
S20106は、生育ステージ分類部201の処理終了を表す。
ほ場生育段階データ2014は、収量推計用に取得された空撮画像が撮影された時に関して、各ほ場の生育ステージを推定した結果を格納する。
ほ場ID20141は、収量推計対象であるほ場を識別するためのラベルであり、ほ場ID1301と一致する。
生育ステージ20142は、対応するほ場の生育ステージを表す。
11 過去データ
12 画像DB
13 ほ場GIS
14 気象DB
15 現地調査員
17 収量推計結果
21 標高DB
22 時系列画像DB
101 パラメータ優先度算出部
102 画像解析部
103 実測ほ場選択部
105 ほ場特徴データ
106 実測ほ場表示部
107 実測データ入力部
108 実測ほ場データ
109 非実測ほ場データ
110 収量推計部
111 収量推計結果表示部
201 生育ステージ分類部
1011 統計解析部
1012 優先度算出部
1013 統計データ
1021 ほ場内画素抽出部
1022 ほ場内特徴算出部
1031 ほ場選定軸決定部
1032 ヒストグラム作成部
1033 実測候補選定部
1034 実測コスト算出部
1035 実測ほ場決定部
1101 生育ステージ推定部
1102 時系列パターン解析部
1103 モデル策定部
1104 収量推計部
1105 生育DB
1901 操作部
1902 表示部
1903 プロセッサ
1904 メインメモリ
1905 記憶装置
2011 気温変動推定部
2012 積算気温推定部
2013 生育ステージ推定部
2014 ほ場生育ステージDB
Claims (13)
- ほ場を含む空撮画像を受信できる受信部と、
当該ほ場を含む複数のほ場の形状情報と、位置情報とをそれぞれ記憶する記憶部と、
受信した前記ほ場を含む空撮画像と、当該ほ場の形状情報と位置情報とから、当該ほ場の画像特徴量を算出する画像解析部と、
算出された前記画像特徴量から前記ほ場の予測収量を算出する収量推計部とを有し、
前記収量推計部は、
予め記憶された気象データの時系列パターンを基に、前記ほ場で栽培される作物の特定の生育ステージごとに、前記作物の生育状況と相関を有する第一パラメータ群を出力する時系列パターン解析部を備え、
前記画像特徴量と前記第一パラメータ群を用いて前記ほ場の予測収量を算出することを特徴とする収量予測システム。 - 請求項1に記載の収量予測システムにおいて、
前記パラメータ群は、
高周波成分を除去した気象データの時系列パターンから抽出したピークの値、位置、数、前記時系列パターンの積算値のうち少なくとも一つを用いて算出されること、
を特徴とする収量予測システム。 - 請求項1に記載の収量予測システムにおいて、
前記記憶部に記憶された当年度以前のほ場ごとの属性情報から、実測ほ場を選定するためのパラメータの優先度を算出するパラメータ優先度算出部と、
前記優先度と、前記画像特徴量と、前記ほ場の位置情報から、実測ほ場を選定する実測ほ場選択部を有することを特徴とする収量予測システム。 - 請求項1に記載の収量予測システムにおいて、
前記空撮画像の取得時における前記作物の生育ステージを推定する生育ステージ分類部をさらに有し、
前記生育ステージの推定は、当年度の時系列空撮画像と標高データの少なくとも一つと、当年度の時系列気象データから算出され、
前記生育ステージと、前記画像特徴量と、前記ほ場の位置情報から、実測ほ場を選定する実測ほ場選択部を有することを特徴とする収量予測システム。 - 請求項3または4に記載の収量予測システムにおいて、
選定された前記実測ほ場の実測収量を入力できる実測データ入力部を有することを特徴とする収量予測システム。 - 請求項1に記載の収量予測システムにおいて、
前記時系列パターン解析部は、
前記気象データの時系列パターンを取得し、前記ほ場で栽培される作物の特定の生育ステージごとに、また前記ほ場の撮影された空撮画像の撮影日前後ごとに、前記作物の生育状況と相関を有する第二パラメータ群を出力することを特徴とする収量予測システム。 - 請求項3に記載の収量予測システムにおいて、
前記パラメータ優先度算出部は、
当年度以前の各年度におけるほ場の属性情報の統計情報を比較することにより、前記優先度を算出することを特徴とする収量予測システム。 - 請求項1に記載の収量予測システムにおいて、
前記収量推計部は、
当年度の時系列気象データから、前記ほ場の生育ステージが変化する日付を推定する生育ステージ推定部と、を備えることを特徴とする収量予測システム。 - 請求項1に記載の収量予測システムにおいて、
前記予測収量を表示する収量推計結果表示部をさらに有することを特徴とする収量予測システム。 - ほ場の形状情報と位置情報と、当該ほ場が撮影された空撮画像とから、前記ほ場の画像特徴量を出力する画像解析部と、
前記画像特徴量より、前記ほ場の予測収量を出力する収量推計部とを有し、
前記収量推計部は、
予め記憶された気象データの時系列パターンを基に、前記ほ場で栽培される作物の特定の生育ステージごとに、前記作物の生育状況との相関を有するパラメータ群を出力する時系列パターン解析部を備え、
前記画像特徴量と前記パラメータ群を用いて前記ほ場の予測収量を算出すること
を特徴とする収量予測装置。 - 請求項10に記載の収量予測装置において、
前記予測収量を表示する収量推計結果表示部をさらに有することを特徴とする収量予測装置。 - 請求項10に記載の収量予測装置において、
前記ほ場の形状情報と位置情報とを記憶する記憶部を備えることを特徴とする収量予測装置。 - 請求項10に記載の収量予測装置において、
前記空撮画像を受信する受信部を備えることを特徴とする収量予測装置。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
BR112015031241A BR112015031241B8 (pt) | 2013-06-18 | 2014-05-19 | sistema e dispositivo de previsão de colheitas |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2013-127094 | 2013-06-18 | ||
JP2013127094A JP6147579B2 (ja) | 2013-06-18 | 2013-06-18 | 収量予測システムおよび収量予測装置 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2014203664A1 true WO2014203664A1 (ja) | 2014-12-24 |
Family
ID=52104406
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2014/063150 WO2014203664A1 (ja) | 2013-06-18 | 2014-05-19 | 収量予測システムおよび収量予測装置 |
Country Status (3)
Country | Link |
---|---|
JP (1) | JP6147579B2 (ja) |
BR (1) | BR112015031241B8 (ja) |
WO (1) | WO2014203664A1 (ja) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104914828A (zh) * | 2015-04-22 | 2015-09-16 | 柳州易农科技有限公司 | 用于农业的智能系统 |
JP2017163956A (ja) * | 2016-03-18 | 2017-09-21 | 国立大学法人名古屋大学 | 予測方法、予測プログラム、及び、予測装置 |
JP2017169511A (ja) * | 2016-03-24 | 2017-09-28 | 株式会社日立ソリューションズ東日本 | 農作物の正常株率推定装置、農作物の収穫量予測装置および農作物の正常株率推定方法 |
US10028426B2 (en) | 2015-04-17 | 2018-07-24 | 360 Yield Center, Llc | Agronomic systems, methods and apparatuses |
WO2018151196A1 (ja) * | 2017-02-16 | 2018-08-23 | Necソリューションイノベータ株式会社 | 作物生育の推定装置、作物生育の推定方法、プログラム、および記録媒体 |
CN110443420A (zh) * | 2019-08-05 | 2019-11-12 | 山东农业大学 | 一种基于机器学习的作物产量预测方法 |
CN110892440A (zh) * | 2017-07-18 | 2020-03-17 | 株式会社植物生命系统 | 信息处理装置、信息处理方法和程序 |
WO2020160642A1 (en) * | 2019-02-05 | 2020-08-13 | Farmers Edge Inc. | Harvest confirmation system and method |
US10891482B2 (en) | 2018-07-10 | 2021-01-12 | Adroit Robotics | Systems, devices, and methods for in-field diagnosis of growth stage and crop yield estimation in a plant area |
CN112348229A (zh) * | 2020-10-12 | 2021-02-09 | 上海微亿智造科技有限公司 | 基于工业大数据的产品良品率预测方法和预测系统 |
JP7474031B2 (ja) | 2019-03-29 | 2024-04-24 | 株式会社トプコン | 圃場情報管理装置、圃場情報管理システム、圃場情報管理装置の制御方法及び圃場情報管理プログラム |
Families Citing this family (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6285875B2 (ja) * | 2015-01-06 | 2018-02-28 | 株式会社日立製作所 | 植物成長分析システム、及び植物成長分析方法 |
EP3861844B1 (en) | 2015-06-08 | 2023-08-30 | Climate LLC | Agricultural data analysis |
US10312986B2 (en) * | 2015-08-19 | 2019-06-04 | Intel Corporation | Flexible CSI RS configuration for FD-MIMO systems |
US11062223B2 (en) | 2015-12-02 | 2021-07-13 | The Climate Corporation | Forecasting field level crop yield during a growing season |
JP6679102B2 (ja) * | 2016-03-25 | 2020-04-15 | Necソリューションイノベータ株式会社 | 出荷数量予測装置、出荷数量予測方法、及びプログラム |
JP6685856B2 (ja) * | 2016-06-30 | 2020-04-22 | 株式会社日立製作所 | 農作業計画支援装置及び農作業計画支援方法 |
JP6895680B2 (ja) * | 2017-03-03 | 2021-06-30 | 国立研究開発法人農業・食品産業技術総合研究機構 | 作物の倒伏度評価装置及び作物の倒伏度評価方法 |
JP6333456B1 (ja) * | 2017-08-30 | 2018-05-30 | アビームコンサルティング株式会社 | 生産管理装置、生産管理方法及びプログラム |
JP7125704B2 (ja) * | 2018-05-01 | 2022-08-25 | 国立大学法人千葉大学 | 農業支援システム |
US20220189025A1 (en) * | 2019-03-08 | 2022-06-16 | Assest Corporation | Crop yield prediction program and cultivation environment assessment program |
JP6704148B1 (ja) * | 2019-03-08 | 2020-06-03 | Assest株式会社 | 農作物の収穫量予測プログラム及び農作物の品質予測プログラム |
JP7081536B2 (ja) * | 2019-03-12 | 2022-06-07 | コニカミノルタ株式会社 | 作物の倒伏リスク診断に用いる生育パラメータの測定推奨スポット提示方法、倒伏リスク診断方法、および情報提供装置 |
JP7191785B2 (ja) * | 2019-06-28 | 2022-12-19 | 株式会社クボタ | 農業支援装置 |
JP7229864B2 (ja) * | 2019-06-28 | 2023-02-28 | 株式会社日立製作所 | リモートセンシング画像取得時期決定システム、および、作物生育状況分析方法 |
JP7263200B2 (ja) * | 2019-10-08 | 2023-04-24 | ヤンマーパワーテクノロジー株式会社 | 農業ハウスの環境制御システム |
WO2021157032A1 (ja) * | 2020-02-06 | 2021-08-12 | 株式会社ナイルワークス | 生育診断システム、生育診断サーバ及び生育診断方法 |
WO2024018559A1 (ja) * | 2022-07-20 | 2024-01-25 | 株式会社ファームシップ | 出荷量予測方法及び出荷量予測システム |
CN117575111B (zh) * | 2024-01-16 | 2024-04-12 | 安徽农业大学 | 一种基于迁移学习的农业遥感图像时空序列预测方法 |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001256289A (ja) * | 2000-03-14 | 2001-09-21 | Shigenao Daimatsu | 農業および水産業用生産管理システム |
JP2003006274A (ja) * | 2001-06-25 | 2003-01-10 | National Agricultural Research Organization | 生育段階毎の気候的類似性に基づく導入適作物・適品種、栽培適地、適作期判定システム |
JP2005085059A (ja) * | 2003-09-10 | 2005-03-31 | Sec:Kk | 農作業決定支援用予測システム |
JP4140052B2 (ja) * | 2002-05-28 | 2008-08-27 | 株式会社サタケ | 作物の生育診断方法における画像の補正方法 |
JP2010166851A (ja) * | 2009-01-22 | 2010-08-05 | Chiharu Hongo | 作物の収量予測方法及びその装置 |
JP4873545B2 (ja) * | 2006-05-16 | 2012-02-08 | 株式会社日立ソリューションズ | 圃場管理支援方法及びシステム |
-
2013
- 2013-06-18 JP JP2013127094A patent/JP6147579B2/ja active Active
-
2014
- 2014-05-19 WO PCT/JP2014/063150 patent/WO2014203664A1/ja active Application Filing
- 2014-05-19 BR BR112015031241A patent/BR112015031241B8/pt not_active IP Right Cessation
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001256289A (ja) * | 2000-03-14 | 2001-09-21 | Shigenao Daimatsu | 農業および水産業用生産管理システム |
JP2003006274A (ja) * | 2001-06-25 | 2003-01-10 | National Agricultural Research Organization | 生育段階毎の気候的類似性に基づく導入適作物・適品種、栽培適地、適作期判定システム |
JP4140052B2 (ja) * | 2002-05-28 | 2008-08-27 | 株式会社サタケ | 作物の生育診断方法における画像の補正方法 |
JP2005085059A (ja) * | 2003-09-10 | 2005-03-31 | Sec:Kk | 農作業決定支援用予測システム |
JP4873545B2 (ja) * | 2006-05-16 | 2012-02-08 | 株式会社日立ソリューションズ | 圃場管理支援方法及びシステム |
JP2010166851A (ja) * | 2009-01-22 | 2010-08-05 | Chiharu Hongo | 作物の収量予測方法及びその装置 |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10028426B2 (en) | 2015-04-17 | 2018-07-24 | 360 Yield Center, Llc | Agronomic systems, methods and apparatuses |
CN104914828A (zh) * | 2015-04-22 | 2015-09-16 | 柳州易农科技有限公司 | 用于农业的智能系统 |
JP2017163956A (ja) * | 2016-03-18 | 2017-09-21 | 国立大学法人名古屋大学 | 予測方法、予測プログラム、及び、予測装置 |
JP2017169511A (ja) * | 2016-03-24 | 2017-09-28 | 株式会社日立ソリューションズ東日本 | 農作物の正常株率推定装置、農作物の収穫量予測装置および農作物の正常株率推定方法 |
WO2018151196A1 (ja) * | 2017-02-16 | 2018-08-23 | Necソリューションイノベータ株式会社 | 作物生育の推定装置、作物生育の推定方法、プログラム、および記録媒体 |
JPWO2018151196A1 (ja) * | 2017-02-16 | 2019-11-07 | Necソリューションイノベータ株式会社 | 作物生育の推定装置、作物生育の推定方法、プログラム、および記録媒体 |
JP7007018B2 (ja) | 2017-02-16 | 2022-01-24 | Necソリューションイノベータ株式会社 | 作物生育の推定装置、作物生育の推定方法、プログラム、および記録媒体 |
CN110892440A (zh) * | 2017-07-18 | 2020-03-17 | 株式会社植物生命系统 | 信息处理装置、信息处理方法和程序 |
US10891482B2 (en) | 2018-07-10 | 2021-01-12 | Adroit Robotics | Systems, devices, and methods for in-field diagnosis of growth stage and crop yield estimation in a plant area |
US11580731B2 (en) | 2018-07-10 | 2023-02-14 | Adroit Robotics | Systems, devices, and methods for in-field diagnosis of growth stage and crop yield estimation in a plant area |
WO2020160642A1 (en) * | 2019-02-05 | 2020-08-13 | Farmers Edge Inc. | Harvest confirmation system and method |
US11222206B2 (en) | 2019-02-05 | 2022-01-11 | Farmers Edge Inc. | Harvest confirmation system and method |
JP7474031B2 (ja) | 2019-03-29 | 2024-04-24 | 株式会社トプコン | 圃場情報管理装置、圃場情報管理システム、圃場情報管理装置の制御方法及び圃場情報管理プログラム |
CN110443420A (zh) * | 2019-08-05 | 2019-11-12 | 山东农业大学 | 一种基于机器学习的作物产量预测方法 |
CN112348229A (zh) * | 2020-10-12 | 2021-02-09 | 上海微亿智造科技有限公司 | 基于工业大数据的产品良品率预测方法和预测系统 |
Also Published As
Publication number | Publication date |
---|---|
JP2015000049A (ja) | 2015-01-05 |
JP6147579B2 (ja) | 2017-06-14 |
BR112015031241B1 (pt) | 2020-07-07 |
BR112015031241B8 (pt) | 2020-07-28 |
BR112015031241A2 (pt) | 2017-07-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6147579B2 (ja) | 収量予測システムおよび収量予測装置 | |
Friedli et al. | Terrestrial 3D laser scanning to track the increase in canopy height of both monocot and dicot crop species under field conditions | |
Chappell et al. | Improving ground cover monitoring for wind erosion assessment using MODIS BRDF parameters | |
Tilly et al. | Multitemporal crop surface models: accurate plant height measurement and biomass estimation with terrestrial laser scanning in paddy rice | |
Wardlow et al. | Using USDA crop progress data for the evaluation of greenup onset date calculated from MODIS 250-meter data | |
Hopkinson et al. | The uncertainty in conifer plantation growth prediction from multi-temporal lidar datasets | |
Möller et al. | Coupling of phenological information and simulated vegetation index time series: Limitations and potentials for the assessment and monitoring of soil erosion risk | |
Zimbres et al. | Savanna vegetation structure in the Brazilian Cerrado allows for the accurate estimation of aboveground biomass using terrestrial laser scanning | |
Ren et al. | Winter wheat planted area monitoring and yield modeling using MODIS data in the Huang-Huai-Hai Plain, China | |
JP2011167163A (ja) | 水稲収量予測モデル生成方法、及び水稲収量予測方法 | |
Hou et al. | Use of ALS, Airborne CIR and ALOS AVNIR-2 data for estimating tropical forest attributes in Lao PDR | |
Liu et al. | Estimating potato above-ground biomass by using integrated unmanned aerial system-based optical, structural, and textural canopy measurements | |
Jin et al. | Assimilating remote sensing data into a crop model improves winter wheat yield estimation based on regional irrigation data | |
Lamqadem et al. | Analysis of Landsat-derived multitemporal vegetation cover to understand drivers of oasis agroecosystems change | |
Araújo et al. | Mapping of summer crops in the State of Paraná, Brazil, through the 10-day spot vegetation NDVI composites | |
Cheng et al. | Evaluation of UAV-based drought indices for crop water conditions monitoring: A case study of summer maize | |
Mashonganyika et al. | Mapping of winter wheat using sentinel-2 NDVI data. a case of Mashonaland central province in Zimbabwe | |
CN116579521B (zh) | 产量预测时间窗口确定方法、装置、设备及可读存储介质 | |
CN109960972B (zh) | 一种基于中高分辨率时序遥感数据的农林作物识别方法 | |
Jeppesen et al. | Identification of high-variation fields based on open satellite imagery | |
Kandel | Estimation of above ground forest biomass and carbon stock by integrating LiDAR, satellite image and field measurement in Nepal | |
Akila et al. | Automation in plant growth monitoring using high-precision image classification and virtual height measurement techniques | |
Ghiani et al. | Use of UAVs and Canopy Height Model Applied on a Time Scale in the Vineyard | |
IL289666A (en) | Assesses variety for crop yield | |
Naveed et al. | Cotton cultivated area detection and yield monitoring combining remote sensing with field data in lower Indus River basin, Pakistan |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 14814635 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
REG | Reference to national code |
Ref country code: BR Ref legal event code: B01A Ref document number: 112015031241 Country of ref document: BR |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 14814635 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 112015031241 Country of ref document: BR Kind code of ref document: A2 Effective date: 20151214 |