Rice cultivation management device, rice cultivation management method, and rice cultivation management system

Abstract

Provided are a rice cultivation management device, a rice cultivation management method, and a rice cultivation management system that acquire one or more types of satellite data of a paddy field, calculate a water index for each day of each of a plurality of water management sections included in the paddy field on the basis of the satellite data, and classify the status for each day of each of the plurality of water management sections on the basis of the water index. It is possible to classify the daily state of each of the plurality of water management sections on the basis of the water index, store the daily state data in a calendar (time-series database), detect the water management technique that is implemented using data, and check the contents of a log book of the farmer. Furthermore, it is possible to replace the log book with the calendar in the future by using data from a water level sensor in combination. It is also possible to more accurately estimate the daily greenhouse gas emissions from individual paddy fields for the purpose of creating carbon credits.

Description

Rice cultivation management device, rice cultivation management method, and rice cultivation management system

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

[0002] Carbon credits have been gaining popularity in recent years. Under the carbon credit system, businesses (credit creators) create credits by implementing projects that reduce greenhouse gas emissions or increase absorption. Meanwhile, companies, local governments, and individuals (credit buyers) purchase the created credits and use them for carbon offsetting, compliance with laws and regulations, and other purposes.

[0003] Carbon credits are a major economic tool to promote an incentive to shift from non-AWD rice farming to AWD rice farming. For carbon credits to be traded effectively, they must be highly reliable and have market value. To issue carbon credits with market value (such as J-Credits), it is essential to accurately calculate greenhouse gas emissions.

[0004] Meanwhile, a method is known for determining the start time of water intake from paddy fields using satellite data (Non-Patent Document 1).

[0005] Masato Fukumoto, "Determining the start time of water withdrawal in paddy fields using Sentinel-2 satellite data," [online], March 5, 2019, Systems Agriculture (J.JASS), 35(2): 15-23, 2019, [Retrieved September 21, 2023], Internet <URL: https: / / www.jstage.jst.go.jp / article / jass / 35 / 2 / 35_15 / _pdf / -char / ja>

[0006] According to Non-Patent Document 1, satellite data observed during clear weather from April to June is used to determine the start of water withdrawal from paddy fields every week or so (Figure 6), and the actual state of agricultural water use is understood. However, since Non-Patent Document 1 uses satellite data observed during clear weather, it does not teach how to determine the water withdrawal status in more specific and shorter increments, such as one day (page 22, "5. Conclusion"). Furthermore, its purpose is to understand the actual state of agricultural water use (page 15, "1. Introduction"). Because it is not possible to determine the water withdrawal status in more specific and shorter increments, such as one day, it is not possible to estimate greenhouse gas emissions from paddy fields, for example, on a daily basis.

[0007] In view of the above circumstances, the purpose of this disclosure is to estimate greenhouse gas emissions from rice paddies for the purpose of creating carbon credits.

[0008] A rice cultivation management device, a rice cultivation management method, and a rice cultivation management system according to one embodiment of the present disclosure acquire one or more types of satellite data of a paddy field, calculate a daily water index for each of a plurality of water management sections included in the paddy field based on the satellite data, and classify the daily status of each of the plurality of water management sections based on the water index.

[0009] According to the present disclosure, it is possible to estimate greenhouse gas emissions from rice paddies for the purpose of creating carbon credits.

[0010] The effects described here are not necessarily limited to those described herein, and may be any of the effects described in this disclosure.

[0011] 1 shows the configuration of a rice cultivation management system according to an embodiment of the present disclosure; 2 shows the operation flow of a rice cultivation management system; 3 shows a machine learning model; 4 shows the algorithm of a status classification unit; 5 shows calendars for each of multiple water management sections; 6 shows an example in which the drainage period (intermittent irrigation) analyzed by the analysis unit is superimposed on a calendar; 7 shows an example of the drainage period (intermittent irrigation) analyzed by the analysis unit; 8 shows that daily status can be classified more accurately based on satellite data from multiple types of satellites; 9 shows a field water tube.

[0012] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0013] 1. Background of the present embodiment

[0014] Rice is the staple food for more than half of the world's population and is also a major source of greenhouse gas (GHG) emissions. Rice cultivation is said to be responsible for 12% of global anthropogenic methane emissions (https: / / www.bloomberg.com / news / articles / 2019-06-03 / your-bowl-of-rice-is-hurting-the-climate-too#xj4y7vzkg). Given the current state of global warming, it is urgent to widely establish sustainable farming methods in rice-producing regions around the world as a strategy to reduce global greenhouse gas emissions.

[0015] Alternate wetting and drying (AWD) is a rice cultivation method known as intermittent irrigation. AWD contrasts with conventional rice cultivation methods, in which paddies remain flooded throughout the rice growing season. AWD is a water management technique that temporarily drains water from paddies and exposes the soil to air. Two typical AWD techniques are known: intermediate drainage, in which paddies are drained once during the growing season (so-called mid-season drainage), and intermittent irrigation, in which paddies are drained and re-flooded alternately in multiple cycles throughout the growing season. Of these, mid-season drainage (i.e., draining the paddy field once before heading and allowing the field surface to dry during the rice growing season) can prevent excessive tillering (branching from the base of the rice plant) and control rice growth.

[0016] In both AWD methods, exposing the soil to oxygen stops the anaerobic decomposition of soil organic matter, reducing methane emissions. Therefore, extending the drainage period (for example, the period of mid-season drainage) beyond the conventional period reduces methane emissions from the soil. In this way, adjusting water management can reduce greenhouse gas emissions from rice paddies.

[0017] Carbon credits have been gaining popularity in recent years. Under the carbon credit system, businesses (credit creators) create credits by implementing projects that reduce greenhouse gas emissions or increase absorption. Meanwhile, companies, local governments, and individuals (credit buyers) purchase the created credits and use them for carbon offsetting and other purposes.

[0018] Carbon credits are a major economic tool to promote an incentive to shift from non-AWD rice farming to AWD rice farming. For carbon credits to be traded effectively, they must be highly reliable and have market value. To issue carbon credits with market value (such as J-Credits), it is essential to accurately calculate greenhouse gas emissions.

[0019] In Japan, there is a national credit certification system called "J-Credit." Under the J-Credit system, credits are certified to business owners (e.g., rice farmers) when certain conditions are met for extending the mid-drainage period in paddy rice cultivation (https: / / japancredit.go.jp / pdf / methodology / AG-005_v1.0.pdf). Specifically, if the mid-drainage period in paddy rice cultivation is extended by seven days or more from the average number of days implemented over the past two or more years prior to the project in the project-implemented paddy fields, credit application can be certified. If credit application is certified, the value of the credit is calculated based on the difference between baseline emissions (emissions expected if the drainage period is not extended) and emissions after the project is implemented (emissions expected if the drainage period is extended).

[0020] To prove that greenhouse gas emissions have been reduced or that the reduction amount has increased compared to baseline emissions, first, it is necessary to prove that the drainage period extension was not implemented (or that no drainage period was set) before the start of the project (baseline period), and second, that the drainage period extension was implemented during the project period. In other words, to be certified for credits, it is necessary to accurately prove the drainage period over several years.

[0021] Furthermore, accurate data on the drainage period (the period during which the paddy field is waterless) allows for accurate calculation of greenhouse gas emissions. For example, when a farmer artificially drains and re-floods a paddy field, if rain falls during the drainage period and water accumulates in the paddy field, greenhouse gas emissions will decrease. Because the value of the credit depends on the amount of greenhouse gas emissions, accurate calculation of the credit value requires accurate understanding of not only the artificial drainage period due to artificial drainage and re-flooding, but also the presence or absence of water and the amount of water in the paddy field, which depend on factors such as the presence or absence of rainfall, the drainage properties of the paddy field, and water management for the paddy field.

[0022] Typically, farmers keep logbooks (diary records) to record the dates of flooding and draining their paddy fields as a means of proving the duration of drainage. However, relying solely on logbooks to receive carbon credits poses problems. First, when a new project begins, logbooks from the baseline period (several years ago) may not be available. Second, financial incentives may lead to doubts about the accuracy and authenticity of the logbooks. Third, even if farmers accurately record their logbooks, while their own actions (e.g., the dates of opening and closing flood gates and drainage gates) are generally recorded, the presence of water due to rainfall, for example, may not be recorded in the logbook because it is not based on the farmer's own actions. However, to accurately calculate greenhouse gas emissions, it is necessary to objectively determine the actual state of the paddy field (the presence or absence of water, and the amount of water) at a specific time.

[0023] The accuracy of greenhouse gas emission calculations is a particularly important issue. For example, doubts have been raised about the accuracy of the rice cultivation methodology of the United Nations Framework Convention on Climate Change (UNFCCC) Clean Development Mechanism (CDM) (https: / / cdm.unfccc.int / methodologies / DB / D14KAKRJEW4OTHEA4YJICOHM26M6BM) (https: / / verra.org / verra-inactivates-unfccc-cdm-rice-cultivation-methodology / ).

[0024] Methods for estimating greenhouse gas emissions from rice paddies include process-based computer models, such as the DeNitrification-DeComposition (DNDC) model developed at the University of New Hampshire. This model is commonly used to simulate complex biogeochemical interactions between crops, soil, irrigation, fertilizer, etc. This model has already been adopted in rice paddy methane reduction methods for carbon credits, but its accuracy is limited because it is generally based on rough estimates and seasonal averages.

[0025] In view of the above circumstances, according to an embodiment of the present disclosure, the waterless periods of paddy fields that are not affected by artificial drainage or rainfall are analyzed on a daily basis over a period of several years or more, thereby enabling farmers to receive appropriate carbon credit certification, thereby reliably proving the drainage periods that have been carried out over a period of several years or more.

[0026] 2. Structure of the rice cultivation management system

[0027] FIG. 1 shows the configuration of a rice cultivation management system according to an embodiment of the present disclosure.

[0028] The rice cultivation management system 1 acquires satellite data of paddy fields acquired by multiple types of artificial satellites 210 from the satellite database 200 via a network such as the Internet. The rice cultivation management system 1 performs calculations based on the satellite data. The rice cultivation management system 1 outputs the calculation results via a network such as the Internet.

[0029] The rice cultivation management system 1 has a satellite data acquisition unit 110, a water index calculation unit 120, a data processing unit 140, a calendar generation unit 150, an analysis unit 160, an output unit 170, and a discharge amount estimation unit 180. In a single information processing device (rice cultivation management device 100) functioning as a server, a processor may load an information processing program recorded in a ROM into a RAM and execute the program, thereby functioning as the satellite data acquisition unit 110, the water index calculation unit 120, the status classification unit 140, the calendar generation unit 150, the analysis unit 160, and the output unit 170. Alternatively, a plurality of distributed information processing devices may cooperate to function as the satellite data acquisition unit 110, the water index calculation unit 120, the data processing unit 140, the calendar generation unit 150, the analysis unit 160, the output unit 170, and the discharge amount estimation unit 180.

[0030] The multiple types of satellites 210 have different detection characteristics and acquire satellite data with different characteristics, and the multiple types of satellites 210 may be, for example, Sentinel-2 (referred to as S2), LandSat-9 (referred to as LS9), Sentinel-1 (referred to as S1), and PlanetScope (referred to as PS).

[0031] The wavelength bands that S2 can detect are visible light, near infrared (NIR), and shortwave infrared (SWIR). The recurrence period of S2 is 5 days on the equator, and the resolution is 10m to 60m.

[0032] The LS9 can detect wavelength bands in visible light, near infrared (NIR), and shortwave infrared (SWIR). The recurrence period of LS9 is 16 days, and the resolution is 30 m.

[0033] S1 has a synthetic aperture radar (SAR) that irradiates microwaves (electromagnetic waves) on a target and receives the reflected (backscattered) signal. S1's satellite data (SAR backscattered) includes VV polarization (vertical output, vertical reception) and VH polarization (vertical output, horizontal reception). S1's recurrence period is 12 days, and its resolution is 10m.

[0034] The detectable wavelength band of PS is visible light and near-infrared (NIR), but does not include short-wave infrared (SWIR). The multiple types of satellites 210 are not limited to those described above; for example, other satellites 210 may also be used. Also, some of the above may not be used (for example, PS may not be used). In this case, the third water index calculation unit 123, which will be described later, may not be provided.

[0035] 3. Operation of the rice cultivation management system

[0036] FIG. 2 shows the operational flow of the rice cultivation management system.

[0037] The satellite data acquisition unit 110 acquires satellite data on paddy fields acquired by multiple types of artificial satellites 210 (S2, S1, LS9) from the satellite database 200 via a network such as the Internet (step S101).

[0038] The water index calculation unit 120 calculates a daily water index for each of multiple water management compartments included in a paddy field based on satellite data (optical band data) from multiple types of artificial satellites 210 (S2, S1, LS9). In this specification, a water management compartment refers to a compartment where appropriate water management is possible. A water management compartment may be, for example, a field plot (the largest compartment where appropriate water management is possible). The water index calculation unit 120 has a first water index calculation unit 121, a second water index calculation unit 122, and a third water index calculation unit 123.

[0039] The first water index calculation unit 121 calculates two types of water indices for each of the multiple water management sections included in the rice paddy based on the reflected light band data (visible light, near-infrared (NIR), and short-wave infrared (SWIR)) of the satellite data S2 (step S102). The two types of water indices are the modified normalized difference water index (MNDWI) and the surface water index (LSWI). The MNDWI is calculated by MNDWI = (Green - SWIR) / (Green + SWIR) (Green is the green band). The LSWI is calculated by LSWI = (nir - swir) / (nir + swir). The MNDWI is effective for detecting surface water, but it is difficult to detect surface water when rice plants grow above the water surface. On the other hand, the LSWI can more appropriately detect water indices (by quantifying the presence of water in vegetation) during periods when rice plants are above the water surface, such as the latter half of the rice-growing season.

[0040] The first water index calculation unit 121 calculates the water index (MNDWI and LSWI) for each of the multiple water management sections included in the rice paddy field based on the reflected light optical band data (visible light, near infrared (NIR), shortwave infrared (SWIR)) which is satellite data from LS9, in the same manner as described above.

[0041] The third water index calculation unit 123 calculates the water index for each of the multiple water management compartments included in the paddy field based on the reflected light band data (visible light and near-infrared (NIR)) that is satellite data from the PS. The detectable wavelength band of the PS is visible light and near-infrared (NIR), but does not include shortwave infrared (SWIR). Therefore, it is not possible to calculate the MNDWI and LSWI based solely on the satellite data from the PS. In this case, the third water index calculation unit 123 may use machine learning to identify the correlation between the visible light and near-infrared (NIR) data and the water indices (MNDWI and LSWI). For example, the third water index calculation unit 123 may use a neural network trained using, for example, a logbook as training data to calculate the correlation between the visible light band and the predicted waterlogging / draining values. This allows the water indices (MNDWI and LSWI) to be detected based on the satellite data from the PS. The training data may be data from the field water tube 300 and / or the digital water level sensor 301 described below.

[0042] The second water index calculation unit 122 calculates the water index for each of the water management sections included in the paddy field based on the synthetic aperture radar data, which is satellite data from S1 (step S103). Specifically, the second water index calculation unit 122 includes a machine learning model 130 for calculating the LSWI from the synthetic aperture radar data from S1. The satellite data from S1 (SAR backscatter) includes VV polarization (vertical output, vertical reception) and VH polarization (vertical output, horizontal reception). The VV polarization is used to detect surface shape. The VH polarization is used to detect soil moisture content. Radar does not rely on visible light and can pass through clouds. Therefore, unlike satellite data from S2, LS9, and PS (optical band data), radar data from S1 can reinforce the MNDWI and LSWI based on satellite data from S2, LS9, and PS (optical band data).

[0043] FIG. 3 shows the machine learning model.

[0044] The machine learning model 130 is a model obtained by training the VV backscattering coefficient and VH backscattering coefficient based on synthetic aperture radar data, the VV / VH ratio, the date (Julian day), and the LSWI value to obtain a representative index value of the radar data through a simple neural network. The XGBoost machine learning algorithm can be used to improve the correlation accuracy of the machine learning model 130. XGBoost is a decision tree-based supervised machine learning algorithm for problems of return date and classification.

[0045] The second water index calculation unit 122 inputs the backscattering coefficients of VV and VH based on the synthetic aperture radar data, the ratio VV / VH, and the date (Julian day) to the machine learning model 130. The machine learning model 130 outputs the LSWI.

[0046] FIG. 4 shows the algorithm of the status classifier.

[0047] The data processing unit 140 has a status classification unit 141 and a water level estimation unit 142. The status classification unit 141 classifies the daily status of each of the multiple water management sections based on the MNDWI and LSWI calculated by the first water index calculation unit 121 and the third water index calculation unit 123, and the LSWI calculated by the second water index calculation unit 122 (step S104).

[0048] There are five types of status for each day: Dry, Wet, Cloud-covered, Snow, or No satellite observation. The algorithm for detecting snow (the third step below) does not need to be executed depending on the climate of the region where the paddy field is located.

[0049] The status classification unit 141 performs threshold processing on the MNDWI and LSWI calculated by the first water index calculation unit 121 and the third water index calculation unit 123. If MNDWI<0.0 and LSWI<0.17, the status is likely to be flooded. If MNDWI<0.0 and LSWI>0.17 or if MNDWI>0.0, the status is likely to be flooded.

[0050] The status classification unit 141 performs threshold processing on the LSWI calculated by the second water index calculation unit 122. If LSWI<0.17, the status is likely to be flooded. If LSWI>0.17, the status is likely to be flooded.

[0051] The status classification unit 141 determines whether the status is flooded or drained based on the threshold processing of the MNDWI and LSWI calculated by the first water index calculation unit 121 and the third water index calculation unit 123 and the threshold processing of the LSWI calculated by the second water index calculation unit 122. It is possible that the results of the threshold processing of the first water index calculation unit 121 and the third water index calculation unit 123 may differ from the results of the threshold processing of the second water index calculation unit 122. Therefore, it is preferable for the status classification unit 141 to perform machine learning in advance and determine whether the status is flooded or drained based on the threshold processing of the first water index calculation unit 121, the second water index calculation unit 122, and the third water index calculation unit 123. Furthermore, each of the multiple types of artificial satellites 210 has a recurrence period (5 days for S2, 16 days for LS9). Therefore, there may be days when only one artificial satellite 210 can obtain satellite data on paddy fields, while the other cannot. In such a case, the status classification unit 141 may determine whether the area is flooded or has fallen into water based on the satellite data obtained.

[0052] Thus, as a first step, the status classification unit 141 uses a multi-level threshold algorithm that utilizes two types of water indices (MNDWI and LSWI) to distinguish between waterlogged and flooded. The threshold for MNDWI (0.0) and the threshold for LSWI (0.17) will be described. A common threshold used to detect water for MNDWI is 0. An MNDWI greater than 0 indicates the presence of water, while an MNDWI less than 0 indicates the absence of water. On the other hand, there is no such predefined threshold for LSWI. Therefore, when training the algorithm, farmer logbook data can be used as training data to determine the LSWI threshold that returns the most accurate results.

[0053] In addition to farmers' logbooks, digital water level sensors 301 (IoT sensors) installed in rice paddies can also be used as training data for machine learning.

[0054] Techniques exist for digitally measuring water levels in rice paddies frequently (e.g., daily or more frequently) using digital water level sensors 301 that are more accurate than farm logbooks. Such sensors are typically mounted above the maximum water level in the paddy field and measure water depth using the reflection of a laser or electronic eye sensor. Collected data may be transmitted wirelessly to a central database or stored locally, requiring periodic data downloads.

[0055] However, deploying digital water level sensors 301 to monitor every individual rice field in a project can be impractical and expensive. In contrast, in this embodiment, deploying digital water level sensors 301 in a representative sample of rice fields can generate training data for a neural network to detect water levels from satellites, allowing for more accurate and cost-effective data collection.

[0056] FIG. 9 shows a field water tube.

[0057] A field water tube 300 may be used to monitor water depth in rice paddies. The field water tube 300 is a hollow, cylindrical tube that is buried vertically in the soil without any soil inside. The lower portion of the tube has multiple punched holes that allow water to flow in and out. When water is above the soil surface, the upper portion of the field water tube 300 (the portion without punched holes) collects water, allowing a positive water level to be measured. On the other hand, when water is below the soil surface, water in the soil flows into the field water tube 300 through the punched holes, collecting water in the lower portion of the field water tube 300 (the portion with punched holes), allowing a negative water level to be measured. In this way, the field water tube 300 can be used to confirm both positive and negative water levels. In AWD practice, where the soil is drained to a depth of -15 cm, the field water tube 300 is important in determining when to re-flood the field. Combining a digital water level sensor 301 with a field water tube 300 allows for digital measurements of both positive values ​​(when the water is above ground) and negative values ​​(when the water is below ground).

[0058] Digital datasets of positive and negative water levels from a representative sample of rice fields can be used as training data (instead of farm logbooks) to estimate rice field water levels with greater accuracy. In this scenario, neural networks can be run on all bands collected from each satellite source to generate individual correlations for each set of source data, without using water indices such as MNDWI or LSWI to predict water levels accurately to the centimeter.

[0059] The use of the digital water level sensor 301 provides a new and reliable source of satellite data for training machine learning algorithms. Furthermore, the algorithm can be used to generate an accurate water level calendar for each rice field. This also creates the ability to see below the soil surface and detect negative water levels. Because intermittent irrigation techniques stipulate draining water 15 cm below the soil surface, machine learning based on training data from the digital sensor data can potentially reinforce that drainage periods are being implemented correctly.

[0060] In the second step, the status classification unit 141, based on the satellite data from S1, S2, LS9, and PS, excludes clouds that account for 80% or more from the median calculation process and determines the status as cloudy. The status classification unit 141 has already determined whether the area is flooded or has fallen in the first step, but if the area is cloudy or has fallen in the second step, it simply determines the status as cloudy. This eliminates the possibility of misidentifying the area as flooded or has fallen in the third place on a cloudy day.

[0061] In the third step, the status classification unit 141 determines the status as "snow accumulation" if snow accumulation accounts for more than 50%. The status classification unit 141 has already determined whether the status is flooded or waterlogged in the first step, but if snow accumulation accounts for more than 50%, it simply determines the status as "snow accumulation." This eliminates the possibility of misidentifying a snowy day as waterlogged.

[0062] In this way, the status classification unit 141 determines whether the status for each day is flooded or falling water in the first step. If the status classification unit 141 determines cloudy or snowfall in the second and third steps, it updates the flooded or falling water status in the first step to cloudy or snowfall.

[0063] Furthermore, as mentioned above, each of the multiple types of satellites 210 has its own recurrence period (5 days for S2, 16 days for LS9, and 12 days for S1). Therefore, there may be days when none of the satellites 210 (S2, LS9, PS, S1) can observe the rice paddies. In such cases, the status classification unit 141 determines the status for each day as no observation.

[0064] Meanwhile, the water level estimation unit 142 uses the data collected from the field water tube 300 and the digital water level sensor 301 as a learning dataset for the neural network and combines it with the satellite data. As a result, the water level estimation unit 142 estimates the positive and negative water levels for each day (step S109). Through machine learning, the water level estimation unit 142 can estimate the positive and negative water levels (cm) for each day in the paddy field using only the satellite data, even if the water level is below the soil surface (for example, -15 cm below).

[0065] FIG. 5 shows a calendar for each of a number of water management zones.

[0066] On the other hand, the calendar generating unit 150 generates calendars C1 to C4 that distinguish and display the daily status (five types: water fall, flooded, cloudy, snow accumulation, and no observation) of each of the multiple water management sections 1 to 4 classified by the status classifying unit 141 (step S105). The calendar can also be said to be a daily time series database.

[0067] A calendar is a two-dimensional matrix consisting of multiple grid-like squares. Each square on the calendar represents a day. The squares are continuous from the first day to the last day without missing a single day. One axis of the two-dimensional matrix that makes up the calendar displays weeks, while the other axis displays consecutive months, without separating them into separate rows (arrangements), for a continuous year. In this example, the vertical axis has seven squares, each corresponding to a day, representing a week from Monday to Sunday from top to bottom along the vertical axis. The horizontal axis displays consecutive months without separating them (arrangements without separation or separation), continuously displaying the year from January 1 to December 31 from left to right along the horizontal axis. In a calendar, the squares for consecutive months are arranged consecutively without any gaps. However, as in this example, adjacent months may be separated by different line types (thick lines in this example) for calendar functionality (to make the dates easier to understand). In addition, if the calendar already has functionality (where the date is clearly indicated) by entering the date in numbers within the grid, it is not necessary to separate adjacent months with different line types.

[0068] The calendar displays different statuses by filling in (painting) the cells in different ways (for example, different colors, different hatching, etc.), thereby visually clearly distinguishing the status for each day. There are, for example, five types of statuses: Dry, Wet, Cloud-covered, Snow, or No satellite observation. Depending on the climate of the region where the rice paddies are located, there may be four types of statuses excluding Snow.

[0069] FIG. 6 shows an example in which the drainage period (intermediate drainage) analyzed by the analysis unit is superimposed on a calendar.

[0070] The analysis unit 160 determines to be a drainage period the longest period during a specific period (e.g., the rice growing season from April to September) that includes a predetermined number of consecutive days or more, does not include any flooded days, the proportion of drained days during the period is equal to or greater than a threshold, and both the first and last days are drained days (step S106). In this example, the analysis unit 160 determines to be a drainage period (mid-season drying period) the longest period during a growing season (e.g., from April to September) that includes a predetermined number of consecutive days or more (e.g., five days), does not include any flooded days, the proportion of drained days during the period (11 / 19 days) is equal to or greater than a threshold (e.g., 50%), and both the first day (Sunday, May 28) and the last day (Thursday, June 16) are drained days.

[0071] The threshold value (e.g., 50%) may be determined based on the cycle or number of days when none of the satellites 210 can observe the rice paddy field due to the recurrent periods of the multiple types of satellites 210 (5 days for S2, 16 days for LS9, and 12 days for S1). The reason why both the first and last days must be drainage days is that, for example, the status for Friday, June 17th is "no observation," so there is a possibility that the paddy field will actually be drained, but the status for Saturday, June 18th is "flooded," so there is a possibility that the paddy field will also be flooded on Friday, June 17th. For this reason, it is more reliable to exclude Friday, June 17th from the drainage period and determine the last day of the drainage period as Thursday, June 16th, which is a water-drainage day.

[0072] In this way, using a daily calendar of flooding and drainage data, the start and end dates of mid-season drainage (a drainage period during the midseason) can be determined. Mid-season drainage is usually performed from June to July. Therefore, the analysis unit 160 searches for consecutive drainage periods between flooding periods within June and July. The analysis unit 160 determines the first observed drainage day of the mid-season drainage period as the start date of the mid-season drainage. The analysis unit 160 determines the last drainage day before the consecutive flooding days following the mid-season drainage as the end date of the mid-season drainage.

[0073] The calendar generating unit 150 displays the drainage period (interim drying period) analyzed by the analysis unit 160 by superimposing it on a two-dimensional matrix calendar. For example, the calendar generating unit 150 clearly displays the drainage period (interim drying period) on the calendar, clearly indicating the first day (Sunday, May 28th) and the last day (Thursday, June 16th). In this specific example, the calendar generating unit 150 simply circles the square for the first day (Sunday, May 28th) and the square for the last day (Thursday, June 16th), and surrounds the period including the possession value and the last day with a roughly rectangular thick frame.

[0074] FIG. 7 shows an example of a drainage period (intermittent irrigation) analyzed by the analysis unit.

[0075] The analysis unit 160 determines, as a drainage period (intermittent irrigation period), a period that includes a predetermined number of consecutive days or more during the growing season (e.g., April to September), does not include any flooded days, the percentage of drainage days during the period is equal to or exceeds a threshold, and both the first and last days are drainage days. The analysis unit 160 may analyze whether continuous flooding, intermediate drainage (intermittent drainage), or intermittent irrigation is being implemented. The analysis unit 160 may also analyze whether the farmer is following an intermittent irrigation technique (i.e., intermediate drainage (intermittent drainage) or intermittent irrigation). As shown in (b), intermittent irrigation involves alternating multiple cycles of draining and re-flooding a paddy field throughout the growing season, resulting in multiple short drainage periods. The predetermined number of consecutive days and the threshold for the percentage of drainage days may be determined based on a logbook, the orbital periods of multiple types of artificial satellites 210, or the like. Note that (a) shows a comparative example of a calendar in which no drainage period (intermittent irrigation, mid-season drainage) is performed.

[0076] Thus, intermittent irrigation is a rice cultivation method that uses intermittent irrigation rather than mid-season drainage periods. Intermittent irrigation involves flooding the field to 5 cm above soil level and then allowing the soil to dry out until the water reaches 15 cm below soil level. When intermittent irrigation is used, the analysis unit 160 may use a daily calendar of flooding and drainage data to detect alternating multi-day patterns of flooding and drainage of the rice field.

[0077] The calendar generating unit 150 may superimpose on a two-dimensional matrix calendar the drainage periods (multiple periods of intermittent irrigation) analyzed by the analyzing unit 160. For example, the calendar generating unit 150 may explicitly display on the calendar the drainage periods (multiple periods of intermittent irrigation) with the first and last days clearly indicated.

[0078] The analysis unit 160 analyzes the drainage periods for several consecutive years (which may or may not include the current year) (step S107). For example, the analysis unit 160 analyzes the length of the drainage periods for several consecutive years, the start date, whether or not there was rainfall before or after the drainage period, etc. As mentioned above, under the J-Credit Scheme, if the mid-drying period is extended by seven days or more from the average number of days implemented over the past two or more years, credit application may be approved. For this reason, records from the past two years as a baseline and analysis results of the baseline and the current situation are required.

[0079] For example, the analysis unit 160 may analyze the start date of the recommended drainage period this year before the drainage period begins. The analysis unit 160 may analyze, before the drainage period begins, how many days this year's drainage period should be implemented to extend the drainage period by at least seven days from the average number of days implemented over the past two or more years (the recommended number of days to implement the drainage period). The analysis unit 160 may analyze, during the drainage period, how many days this year's drainage period should continue to extend the drainage period by at least seven days from the average number of days implemented over the past two or more years (the recommended number of remaining days for the ongoing drainage period). The analysis unit 160 may analyze, after the end of the drainage period, whether the drainage period this year has been extended by at least seven days from the average number of days implemented over the past two or more years (whether the goal after the drainage period has been achieved).

[0080] The emission estimation unit 180 further estimates the amount of greenhouse gas emissions every day based on the positive and negative water levels measured by the water level estimation unit 142 (step S110). ４ ) daily emissions, expected nitrous oxide (N ２ O) daily emissions, expected daily CO ２ The emissions estimator 180 may also analyze the daily greenhouse gas emissions by combining environmental data (e.g., soil type, temperature) in addition to water levels and using a process-based computer model such as DNDC. The emissions estimator 180 may also apply standard coefficients to estimate the daily greenhouse gas emissions. ４ and N ２ O to CO ２ Equivalent tonnes (tCO ２ e) and summing the daily emissions over the entire growing season to estimate the emissions for the entire season. The actual emissions during the project can be subtracted from the baseline emissions to obtain tCO ２ The amount of e-reduction is calculated and the corresponding carbon credits are issued.

[0081] The output unit 170 outputs and records the calendar, analysis results, and emission amounts to a large-capacity nonvolatile storage device. The output unit 170 may output the calendar, analysis results, and emission amounts in a format that can be displayed on a display device (step S108). The output unit 170 may also output the analysis results to the terminal device 300 as recommendations.

[0082] The displayed calendar allows users to understand past drainage periods (prior to last year), which makes it easier to determine when to start the drainage period for the future (this year) and how many days it should last, from the perspective of being less susceptible to the effects of rainfall and generating credits based on a baseline.

[0083] 4. Conclusion

[0084] FIG. 8 shows that daily status can be more accurately classified based on satellite data from multiple types of satellites.

[0085] Water in rice paddies can be detected via satellite using a variety of techniques, depending on the specific data bands collected. Satellite coverage has improved, and data providers such as PlanetLabs can provide near-daily imagery of the Earth's land mass. However, to create a daily dataset with minimal missing days, data from multiple satellite sources still needs to be combined and layered into a single calendar. As shown in (d) of the same figure, as more types of satellite data are integrated, the resulting calendar has fewer missing days. (a) shows only S2; (b) shows two types: S2 and S1; (c) shows three types: S2, S1, and LS9; and (d) shows four types: S2, S1, LS9, and PlanetLabs.

[0086] Typically, the use of remote sensing in surface water mapping in rice paddies is limited to detecting water early in the season when the rice plants are usually low enough to observe the water directly. Later in the season, when the rice plants are above the water level, optical band satellites, and even radar satellites, cannot penetrate the rice vegetation to observe the water directly.

[0087] In contrast, this embodiment has the advantages of (1) detecting water under rice using laser data from S1, (2) enabling periodic and high-frequency detection by combining data from multiple satellites, and (3) enabling highly reliable detection by combining optical-band satellites and radar satellites. This embodiment enables the detection of surface water throughout the year at the rice growth stage with a time resolution of one day by integrating various satellite data sources.

[0088] This embodiment allows a baseline to be established based on satellite data without the need to obtain farm logbooks for the period before the project began to extend the drainage period. Using this approach, a daily calendar can be constructed from historical satellite data from several years before the project began, identifying whether a drainage period was not implemented or was implemented, and estimating the corresponding greenhouse gas emissions.

[0089] Records of drainage periods for rice cultivation still largely depend on farmer logbook records, and considering the challenges of verifying logbooks and collecting past logbooks, this embodiment can also be used to verify and reinforce farmer logbooks.

[0090] According to this embodiment, a process-based computer model, such as the DeNitrification-Decomposition (DNDC) model, is coupled with continuous digital measurements, more specifically daily frequency satellite data representing water depth in each rice field. This embodiment may be used to verify the contents of a farmer's logbook. A calendar may be used instead of the logbook as a reliable source of water volume data. By incorporating daily measurements, this embodiment can produce more accurate estimates of greenhouse gas reductions compared to existing techniques.

[0091] According to this embodiment, the daily status of each of multiple water management plots is classified based on water indices, and the daily status data is stored in a calendar (time-series database). The data can be used to detect the water management techniques being implemented and to check the contents of the farmer's logbook. Furthermore, by combining this data with data from water level sensors, the logbook can be replaced with a calendar in the future. It can also more accurately estimate daily greenhouse gas emissions from individual rice paddies for the purpose of generating carbon credits.

[0092] According to this embodiment, by analyzing the waterless periods of paddy fields that are not affected by artificial drainage or rainfall on a daily basis over a period of several years or more, it is possible to reliably prove the drainage periods that have been carried out over a period of several years or more, so that farmers can properly receive carbon credit certification. This embodiment can directly contribute to the convenience of farmers who actually grow rice (recording and analysis over the past several years is possible without creating a logbook) and ultimately to their profits (creation of valuable credits based on accurate greenhouse gas reductions). As a result, this embodiment can achieve the following effects, for example.

[0093] - When farmers want to start a drainage period extension project, they can know the history of baseline drainage periods for the past few years. This allows them to start the project promptly. - After starting a drainage period extension project, farmers can easily determine the start and end dates and length of this year's drainage period appropriately based on the history of baseline drainage periods for the past few years. Furthermore, this year's actual drainage period will be an appropriate start and end date and length, taking into account the history of past drainage periods. - By reliably proving drainage periods over multiple years, greenhouse gas emissions over multiple years can be calculated more accurately. As a result, the value of carbon credits created by farmers can be calculated more accurately. - This will directly contribute to the convenience and profits of farmers who actually grow rice.

[0094] Although the embodiments and modified examples of the present technology have been described above, the present technology is not limited to the above-described embodiments, and it goes without saying that various modifications can be made within the scope of the gist of the present technology.

[0095] 1 Rice cultivation management system 100 Rice cultivation management device 110 Satellite data acquisition unit 120 Water index calculation unit 121 First water index calculation unit 122 Second water index calculation unit 123 Third water index calculation unit 130 Machine learning model 140 Data processing unit 150 Calendar generation unit 160 Analysis unit 170 Output unit 180 Discharge amount estimation unit 200 Satellite database 210 Artificial satellite

Claims

1. A rice cultivation management device comprising: a satellite data acquisition unit that acquires one or more types of satellite data on paddy fields; a water index calculation unit that calculates a daily water index for each of a plurality of water management compartments included in the paddy field based on the satellite data; and a status classification unit that classifies a daily status for each of the plurality of water management compartments based on the water index.

2. A rice cultivation management device as described in claim 1, wherein the water index calculation unit has: a first water index calculation unit and / or a third water index calculation unit that calculates a Modified Normalized Difference Water Index (MNDWI) and a Surface Water Index (LSWI) based on optical band data, which is the satellite data; and a second water index calculation unit that calculates an LSWI based on radar data, which is the satellite data; and the status classification unit classifies the status based on the MNDWI and the LSWI calculated by the first water index calculation unit and / or the third water index calculation unit and the LSWI calculated by the second water index calculation unit.

3. A rice cultivation management device as described in claim 2, wherein the first water index calculation unit calculates the MNDWI and the LSWI based on visible light, near infrared and shortwave infrared light band data, and the third water index calculation unit calculates the MNDWI and the LSWI using a neural network based on the visible light and near infrared light band data.

4. A rice cultivation management device according to claim 1, further comprising a calendar generating unit that generates a calendar that distinguishes and displays the daily status of each of the plurality of water management sections.

5. A rice cultivation management device as described in claim 4, wherein the status classification unit classifies the status into flooded, water falling, cloudy, snow cover or no observation based on the satellite data, and the calendar displays days with flooded, water falling, cloudy, snow cover or no observation in a manner that makes each day distinct.

6. A rice cultivation management device as described in claim 4, wherein the calendar is a two-dimensional matrix of a plurality of successive squares arranged in a grid pattern, each square representing one day, the plurality of squares continuing from the first day to the last day without missing a single day, one axis of the two-dimensional matrix displaying weeks, and the other axis displaying a continuous year with a plurality of consecutive months not separated by any separation in the arrangement.

7. A rice cultivation management device according to claim 6, wherein the calendar displays different statuses by filling up squares in different ways, thereby distinguishing between the statuses.

8. A rice cultivation management device as described in claim 6, further comprising an analysis unit that determines to be a drainage period the longest period that is a period of a predetermined number of consecutive days or more during a specific time period, does not include waterlogged days, has a ratio of water-drained days within the period equal to or greater than a threshold, and has both the first and last days as water-drained days.

9. A rice cultivation management device according to claim 8, wherein the calendar generation unit further displays the drainage period by superimposing it on the two-dimensional matrix calendar.

10. A rice cultivation management device according to claim 8, wherein the analysis unit analyzes whether the current irrigation method is continuous flooding, intermediate drainage (intermediate drainage), or intermittent irrigation.

11. A rice cultivation management device as described in claim 8, wherein the analysis unit analyzes drainage periods over multiple consecutive years to determine a recommended start date for a future drainage period, the recommended number of days for which a drainage period is to be implemented, the recommended remaining number of days for a drainage period currently being implemented, and / or whether or not a goal has been achieved after the implementation of a drainage period.

12. A rice cultivation management device as described in claim 1, further comprising a water level estimation unit that estimates daily positive and negative water levels based on the satellite data through machine learning by learning the satellite data using water level measurement results from an IoT sensor and / or water tube installed in the rice field as training data.

13. The rice cultivation management device according to claim 12, further comprising an emission amount estimation unit that estimates daily greenhouse gas emissions based on the daily positive and negative water levels.

14. A rice cultivation management method comprising: acquiring one or more types of satellite data on a paddy field; calculating a daily water index for each of a plurality of water management compartments contained in the paddy field based on the satellite data; and classifying a daily status of each of the plurality of water management compartments based on the water index.

15. A rice cultivation management system comprising a plurality of information processing devices which function as: a satellite data acquisition unit which acquires one or more types of satellite data on paddy fields; a water index calculation unit which calculates a daily water index for each of a plurality of water management compartments included in the paddy fields based on the satellite data; and a status classification unit which classifies the daily status of each of the plurality of water management compartments based on the water index.

Images