JP7169022B2 - Harvest forecast system, harvest forecast method, harvest forecast program, and harvest time forecast system - Google Patents

Description

The present invention relates to a harvest amount prediction system, a harvest amount prediction method, a harvest amount prediction program, and a harvest time prediction system.

There is a need for a system that predicts the yield of growing crops. For example, a method has been proposed for estimating the amount of nitrogen absorbed by photographing crops in a field from satellites, drones, or cameras installed in the sky, and analyzing the spectrum, especially in the near-infrared region ( Patent document 1). However, it is difficult to accurately predict the harvest volume at locations where crops bearing crops, such as rice grains, are not sufficiently captured within the camera range and orientation of the crops. there were. More specifically, when the paddy is in the blind spot of the image, it was difficult to predict the yield per plant. Therefore, there is a need for a harvest amount prediction system that can accurately predict the harvest amount even when the grain is not sufficiently captured in the image.

Patent Literature 2 discloses an information processing device that acquires image data of a plant and determines whether or not the plant needs to be watered. Patent Document 3 discloses a plant necessity determination control device that determines whether plants are good or bad based on the color of seedlings planted in a plug tray, the number of leaves, the leaf area, the length of leaves, and the like. It is Patent Literature 4 discloses a crop image processing program that detects ridges and the top of crops using a camera fixed on land and calculates the height of crops.

Patent publication publication JP 2003-339238 Patent Publication JP 2018-108041 Japanese patent publication JP-A-6-138041 Patent publication publication JP 2012-198688

Provided is a yield prediction system that accurately predicts the yield of agricultural crops.

In order to achieve the above object, a yield prediction system according to one aspect of the present invention includes an imaging unit that acquires an image of a growing crop, a crop length of the crop based on the image, and a curvature of the crop in the length direction. and a harvest amount prediction unit that predicts a harvest amount from the crop based on the crop length and the curvature.

The harvest amount prediction unit may predict that the harvest amount increases as the crop length increases and the curvature increases.

The crop shape acquisition unit further acquires the thickness of the stem of the crop based on the image,
The yield prediction unit may predict the yield based on the crop length, the curvature, and the stem thickness.

A strong wind detection unit that detects wind blowing on the crops may be further provided, and the harvest amount prediction unit may vary the harvest amount prediction result depending on whether the strong wind detection unit detects a strong wind. .

The harvest amount prediction unit may downwardly correct the prediction value at the time when the strong wind is detected by the strong wind detection unit.

When the strong wind is detected by the strong wind detection unit, the harvest amount prediction unit may stop the prediction of the harvest amount at a point where the strong wind is detected.

The apparatus may further include a re-capturing unit that re-obtains an image of the crop at a point where the strong wind is detected when the strong wind is not detected when there is a history of the strong wind being detected.

The crop is rice that bears husks in the upper part in the length direction, and further includes a husk counting unit that counts the number of husks. Based on this, the harvest amount may be predicted.

The harvest amount prediction unit calculates an assumed range of at least one of the crop length and the curvature in which the harvest amount is harvested from the harvest amount predicted based on the number of the paddy, and obtains the crop shape. When the acquired value of at least one of the crop length and the curvature acquired by the unit is not included in the assumed range, the harvest amount may be corrected based on the acquired value.

When the obtained values of the crop length and the curvature are larger than the assumed range, the harvest amount may be corrected upward.

When the acquired values of the crop length and the curvature are smaller than the assumed range, the harvest amount may be corrected downward.

The harvest amount prediction unit calculates a first harvest amount predicted based on the length and the curvature, and a second harvest amount predicted based on the number of the unhulled grains, and calculates the second harvest amount. Based on the value, it may be determined whether or not to adopt the first yield as the prediction result.

When the second harvested amount is larger than the first harvested amount by a predetermined amount or more, the harvested amount prediction unit may adopt the second harvested amount as the prediction result.

At least the imaging unit may be mounted on an aircraft.

In order to achieve the above object, a yield prediction system according to another aspect of the present invention includes an imaging unit that acquires an image of a growing crop, a crop length of the crop based on the image, and a curvature of the crop in the length direction. a crop shape acquisition unit that acquires the crop shape acquisition unit, a paddy counting unit that counts the number of grains that bear fruit on the crop, and a yield prediction unit that predicts the yield from the crop based on the number of grains, The harvest amount prediction unit calculates an assumed range of at least one of the crop length and the curvature in which the harvest amount is harvested from the harvest amount predicted based on the number of the paddy, and obtains the crop shape. When the acquired value of at least one of the crop length and the curvature acquired by the unit is not included in the assumed range, the harvested amount is corrected based on the acquired value.

In order to achieve the above object, a yield prediction method according to still another aspect of the present invention comprises the steps of: obtaining an image of a growing crop; and predicting yield from the crop based on the crop length and the curvature.

To achieve the above object, a yield prediction program according to still another aspect of the present invention includes a command to obtain an image of a growing crop, and a crop length and a curvature of the crop in the longitudinal direction based on the image. and predicting yield from the crop based on the crop length and the curvature.
The computer program can be provided by downloading via a network such as the Internet, or can be provided by being recorded on various computer-readable recording media such as a CD-ROM.

It is possible to accurately predict the yield of agricultural products.

1 is a plan view of a drone according to the present invention; FIG. It is a front view of the said drone. It is a right view of the said drone. It is a rear view of the said drone. It is a perspective view of the said drone. 1 is an overall conceptual diagram of a chemical spraying system that the drone has. FIG. FIG. 4 is an overall conceptual diagram showing another example of a chemical spraying system that the drone has. FIG. 11 is an overall conceptual diagram showing still another example of a chemical spraying system that the drone has. It is a schematic diagram showing the control function of the said drone. It is a functional block diagram which the said drone has. It is a flow chart in which the drone predicts the yield. Fig. 10 is a flow chart showing a second embodiment in which the drone predicts yield. FIG. 11 is a flow chart showing a third embodiment in which the drone predicts harvest yield. FIG. It is a flowchart which shows a state when the said drone detects a strong wind.

Embodiments for carrying out the present invention will be described below with reference to the drawings. All figures are illustrative. In the following detailed description, for purposes of explanation, certain details are set forth in order to provide a thorough understanding of the disclosed embodiments. However, embodiments are not limited to these specific details. Also, well-known structures and devices are schematically shown to simplify the drawings.

First, the configuration of the drone according to the present invention will be described. In the specification of the present application, a drone refers to a power means (electric power, prime mover, etc.), a control method (wireless or wired, and whether it is an autonomous flight type or a manually operated type, etc.), It refers to an aircraft in general that has a plurality of rotor blades.

As shown in FIGS. 1 to 5, the rotor blades 101-1a, 101-1b, 101-2a, 101-2b, 101-3a, 101-3b, 101-4a, 101-4b (also called rotors) are: It is a means to fly the drone 100, and considering the balance of flight stability, aircraft size, and power consumption, it is equipped with 8 aircraft (4 sets of two-stage rotor blades). Each rotor 101 is arranged on the four sides of the main body 110 of the drone 100 by means of arms protruding from the main body 110 of the drone 100 . That is, rotor blades 101-1a and 101-1b are on the left rear in the traveling direction, rotor blades 101-2a and 101-2b are on the left front, rotor blades 101-3a and 101-3b are on the right rear, and rotor blades 101- on the right front. 4a and 101-4b are arranged respectively. In addition, the drone 100 makes the downward direction of the paper surface in FIG. 1 the advancing direction. Rod-shaped legs 107-1, 107-2, 107-3, and 107-4 extend downward from the rotating shaft of rotor blade 101, respectively.

Motors 102-1a, 102-1b, 102-2a, 102-2b, 102-3a, 102-3b, 102-4a, 102-4b are equipped with rotors 101-1a, 101-1b, 101-2a, 101- Means for rotating 2b, 101-3a, 101-3b, 101-4a, 101-4b (typically an electric motor, but may be a motor, etc.), one machine provided for one rotor blade It is Motor 102 is an example of a propeller. The upper and lower rotors in one set (e.g. 101-1a and 101-1b) and their corresponding motors (e.g. 102-1a and 102-1b) are used for drone flight stability etc. The axes are collinear and rotate in opposite directions. As shown in FIGS. 2 and 3, the radial members for supporting the propeller guard provided to prevent the rotor from interfering with foreign objects are not horizontal but have a scaffold-like structure. This is to prevent the member from interfering with the rotor by promoting the buckling of the member to the outside of the rotor blade at the time of collision.

Four drug nozzles 103-1, 103-2, 103-3, and 103-4 are provided and are means for spraying the drug downward. In the present specification, chemicals generally refer to liquids or powders such as pesticides, herbicides, liquid fertilizers, insecticides, seeds, and water that are applied to fields.

The drug tank 104 is a tank for storing the sprayed drug, and is provided at a position close to and lower than the center of gravity of the drone 100 from the viewpoint of weight balance. The drug hoses 105-1, 105-2, 105-3, 105-4 are means for connecting the drug tank 104 and the drug nozzles 103-1, 103-2, 103-3, 103-4, and are hard material, and may also serve to support the drug nozzle. A pump 106 is means for ejecting a drug from a nozzle.

FIG. 6 shows an overall conceptual diagram of a system using an embodiment of the drone 100 for chemical spraying according to the present invention. This figure is a schematic diagram and not to scale. In the figure, a drone 100, a controller 401, a small portable terminal 401a, and a base station 404 are connected to a farming cloud 405, respectively. These connections may be wireless communication using Wi-Fi, a mobile communication system, or the like, or may be partially or wholly wired.

The operation device 401 is a means for transmitting commands to the drone 100 by the operation of the user 402 and displaying information received from the drone 100 (for example, position, drug amount, remaining battery level, camera image, etc.). Yes, and may be implemented by a portable information device such as a general tablet terminal that runs a computer program. The drone 100 according to the present invention is controlled to fly autonomously, but manual operation may be performed during basic operations such as takeoff and return, and in an emergency. In addition to the portable information device, an emergency operator (not shown) having a dedicated emergency stop function may be used. The emergency operation device may be a dedicated device equipped with a large emergency stop button or the like so that a quick response can be taken in case of emergency. Furthermore, apart from the operation device 401, the system may include a small portable terminal 401a, such as a smart phone, capable of displaying part or all of the information displayed on the operation device 401. FIG. Further, it may have a function of changing the operation of the drone 100 based on information input from the small portable terminal 401a. The small portable terminal 401a is connected to a base station 404, for example, and can receive information and the like from the farming cloud 405 via the base station 404. FIG.

A farm field 403 is a rice field, a field, or the like to which the drone 100 sprays chemicals. In reality, the topography of the farm field 403 is complicated, and there are cases where a topographic map cannot be obtained in advance, or there are cases where the topographic map differs from the situation of the field. Fields 403 are usually adjacent to houses, hospitals, schools, other crop fields, roads, railroads, and the like. Moreover, there may be intruders such as buildings and electric wires in the field 403 .

The base station 404 is a device that provides a Wi-Fi communication base unit function, etc., and may also function as an RTK-GPS base station and provide an accurate position of the drone 100 (Wi-Fi Fi communication master unit function and RTK-GPS base station may be independent devices). Also, the base station 404 may be able to communicate with the farming cloud 405 using a mobile communication system such as 3G, 4G, and LTE.

The farming cloud 405 is typically a group of computers and related software operated on a cloud service, and may be wirelessly connected to the operation device 401 via a mobile phone line or the like. The farming cloud 405 may analyze the image of the field 403 captured by the drone 100, grasp the growth status of crops, and perform processing for determining the flight route. In addition, the drone 100 may be provided with topographical information and the like of the field 403 that has been saved. In addition, a history of flight and captured images of the drone 100 may be accumulated and various analysis processes may be performed.

The small portable terminal 401a is, for example, a smart phone. The display unit of the small portable terminal 401a displays information about the expected operation of the drone 100, more specifically, the scheduled time for the drone 100 to return to the departure/arrival point 406, and the work to be done by the user 402 when returning. Information such as the content is displayed as appropriate. Also, the operations of the drone 100 and the moving object 406a may be changed based on the input from the small portable terminal 401a. The small portable terminal 401a can receive information from both the drone 100 and the moving object 406a. Information from the drone 100 may also be transmitted to the small portable terminal 401a via the mobile object 406a.

Normally, the drone 100 takes off from the departure/arrival point 406 outside the field 403 and returns to the departure/arrival point 406 after spraying the chemical on the field 403 or when replenishment of the chemical, charging, or the like is required. The flight path (entry path) from the departure/arrival point 406 to the target field 403 may be stored in advance in the farming cloud 405 or the like, or may be input by the user 402 before starting takeoff. Landing point 406 may be a virtual point defined by coordinates stored in drone 100, or may be a physical landing pad.

Note that the drone 100, the operation device 401, the small portable terminal 401a, and the farming cloud 405 may each be connected to the base station 404, as in the example shown in FIG.

Further, as in the example shown in FIG. 8, the drone 100, the controller 401, and the small portable terminal 401a are each connected to the base station 404, and only the controller 401 is connected to the farming cloud 405. may

The drone 100 flies over the field 403 and performs work in the field. In this embodiment, one drone 100 flies and works in one farm field 403 (example of work area), but a plurality of drones may fly and work in one farm field 403 . The drone 100 flies along a driving route planned in advance in the farm field 403 to spray chemicals and photograph the inside of the farm field 403 . The driving route is, for example, a route that flies back and forth in the farm field 403 so as to scan, but may be any route.

FIG. 9 shows a block diagram showing control functions of an embodiment of the chemical spraying drone according to the present invention. The flight controller 501 is a component that controls the entire drone, and specifically may be an embedded computer including a CPU, memory, related software, and the like. Flight controller 501 controls motors 102-1a and 102-1b via control means such as ESC (Electronic Speed Control) based on input information received from operation device 401 and input information obtained from various sensors described later. , 102-2a, 102-2b, 102-3a, 102-3b, 104-a, and 104-b, the flight of the drone 100 is controlled. The actual rotation speeds of motors 102-1a, 102-1b, 102-2a, 102-2b, 102-3a, 102-3b, 104-a, and 104-b are fed back to flight controller 501 to ensure normal rotation. It is configured to be able to monitor whether or not Alternatively, an optical sensor or the like may be provided on the rotor blade 101 and the rotation of the rotor blade 101 may be fed back to the flight controller 501 .

The software used by the flight controller 501 is rewritable through a storage medium or the like, or through communication means such as Wi-Fi communication or USB, in order to extend/change functions or correct problems. In this case, protection is provided by encryption, checksum, electronic signature, virus check software, etc. to prevent rewriting by unauthorized software. Also, part of the calculation processing used for control by the flight controller 501 may be executed by another computer existing on the operation device 401, on the farming cloud 405, or at another location. Due to the high importance of flight controller 501, some or all of its components may be duplicated.

The flight controller 501 communicates with the controller 401 via the Wi-Fi slave device function 503 and via the base station 404, receives necessary commands from the controller 401, and sends necessary information to the controller. You can send to 401. In this case, the communication may be encrypted to prevent fraudulent acts such as interception, impersonation, and device hijacking. The base station 404 also has a function of an RTK-GPS base station in addition to a Wi-Fi communication function. By combining the signals from the RTK base station and the signals from the GPS positioning satellite, the flight controller 501 can measure the absolute position of the drone 100 with an accuracy of several centimeters. Due to the high importance of the flight controller 501, it may be duplicated or multiplexed, and each redundant flight controller 501 should use a different satellite in order to cope with the failure of a particular GPS satellite. may be controlled.

The 6-axis gyro sensor 505 is means for measuring the acceleration of the drone body in three mutually orthogonal directions, and is means for calculating the velocity by integrating the acceleration. The 6-axis gyro sensor 505 is means for measuring changes in the attitude angle of the drone body in the three directions described above, that is, angular velocity. The geomagnetic sensor 506 is a means of determining the direction of the drone body by measuring geomagnetism. The air pressure sensor 507 is a means of measuring air pressure, and can also indirectly measure the altitude of the drone. The laser sensor 508 is means for measuring the distance between the drone body and the ground surface using reflection of laser light, and may be an IR (infrared) laser. Sonar 509 is a means of measuring the distance between the drone body and the ground using the reflection of sound waves such as ultrasonic waves. These sensors may be selected according to the drone's cost targets and performance requirements. Also, a gyro sensor (angular velocity sensor) for measuring the inclination of the airframe, a wind sensor for measuring wind force, etc. may be added. Also, these sensors may be duplicated or multiplexed. If there are multiple sensors for the same purpose, the flight controller 501 may use only one of them and switch to an alternative sensor if it fails. Alternatively, multiple sensors may be used at the same time and a failure is assumed to occur if their measurements do not match.

Flow rate sensors 510 are means for measuring the flow rate of the drug, and are provided at multiple locations along the path from drug tank 104 to drug nozzle 103 . A liquid shortage sensor 511 is a sensor that detects when the amount of medicine has fallen below a predetermined amount. Multispectral camera 512 is a means of photographing field 403 and acquiring data for image analysis. The intruder detection camera 513 is a camera for detecting drone intruders, and is a separate device from the multispectral camera 512 because its image characteristics and lens orientation are different from those of the multispectral camera 512 . Switch 514 is a means for user 402 of drone 100 to make various settings. The intruder contact sensor 515 is a sensor for detecting that the drone 100, especially its rotor or propeller guard portion, has come into contact with an intruder such as an electric wire, building, human body, standing tree, bird, or other drone. . Note that the intruder contact sensor 515 may be replaced by the 6-axis gyro sensor 505. The cover sensor 516 is a sensor that detects that the operation panel of the drone 100 and the internal maintenance cover are open. A drug inlet sensor 517 is a sensor that detects that the inlet of the drug tank 104 is open. These sensors may be selected, duplicated or multiplexed depending on the drone's cost targets and performance requirements. Also, a sensor may be provided outside the drone 100 at the base station 404, the controller 401, or at another location to transmit the read information to the drone. For example, a wind sensor may be provided in the base station 404 to transmit information on wind power and wind direction to the drone 100 via Wi-Fi communication.

The flight controller 501 transmits a control signal to the pump 106 to adjust the medicine ejection amount and stop the medicine ejection. The current status of the pump 106 (eg, number of revolutions, etc.) is configured to be fed back to the flight controller 501 .

The LED 107 is display means for informing the operator of the drone of the state of the drone. Display means such as a liquid crystal display may be used in place of or in addition to LEDs. The buzzer 518 is an output means for informing the state of the drone (especially error state) by means of an audio signal. A Wi-Fi slave device function 519 is an optional component for communicating with an external computer or the like for transferring software, for example, separately from the operation device 401 . In place of or in addition to the Wi-Fi slave unit function, infrared communication, Bluetooth (registered trademark), ZigBee (registered trademark), other wireless communication means such as NFC, or wired communication means such as USB connection may be used. Also, instead of the Wi-Fi slave device function, it may be possible to communicate with each other by mobile communication systems such as 3G, 4G, and LTE. The speaker 520 is output means for notifying the state of the drone (especially error state) by means of recorded human voice, synthesized voice, or the like. Weather conditions can make it difficult to see the visual display of the drone 100 in flight, and in such cases, audible status communication is effective. A warning light 521 is a display means such as a strobe light that indicates the state of the drone (especially an error state). These input/output means may be selected, duplicated or multiplexed according to the drone's cost targets and performance requirements.

●Overview of Harvest Volume Prediction System As shown in FIG. The drone 100 and the yield prediction device 300 may be wirelessly or wiredly connected, capable of transmitting and receiving information, and connected via a network. Note that the configuration of the yield prediction device 300 may be incorporated in the drone 100. FIG.

●Configuration of drone 100 The drone 100 has a flight control unit 21, an imaging unit 22, a strong wind detection unit 23, and a re-imaging unit .

The flight control unit 21 is a functional unit that operates the motor 102 of the drone 100 and controls flight, takeoff and landing of the drone 100 . The flight control unit 21 is implemented by the functions of the flight controller 501, for example.

The imaging unit 22 is a functional unit that acquires images of growing crops, and is realized by a camera such as the multispectral camera 512, for example. The photographing unit 22 acquires an image that can be used to analyze the growing conditions of crops. The photographing unit 22 may further include an irradiating unit that irradiates the field 403 with a light beam of a specific wavelength, and may be capable of receiving reflected light of the light beam from the farm field 403 . Light rays of specific wavelengths may be, for example, red light (wavelength about 650 nm) and near-infrared light (wavelength about 774 nm). By analyzing the reflected light of the light beam, it is possible to estimate the amount of nitrogen absorbed by the crop, and thus to analyze the growth state of the crop.

In addition, the photographing unit 22 can acquire an image from which the shape of the crop can be analyzed.

The strong wind detection unit 23 is a functional unit that detects wind that blows against crops and that satisfies a predetermined condition. When crops are lodged by the wind due to temporary strong winds, it becomes difficult to compare with previous diagnoses, and the accuracy of diagnosis decreases depending on the direction of lodging. Therefore, the strong wind detection unit 23 detects a wind that prevents shooting as planned, and stores the detection point.

The strong wind detection unit 23 includes a wind speed measurement unit 231 , a wind direction measurement unit 232 , a strong wind determination unit 233 and a detection point storage unit 234 .

The wind speed measurement unit 231 is a functional unit that measures the wind speed of the wind that blows around the drone 100 . The wind speed measurement unit 231 may record the measured wind speed along with time.

The wind speed measurement unit 231 is a measurement unit that calculates the wind speed by measuring the stress generated by the wind using, for example, a contact detector. Further, the anemometer 231 may have an anemometer such as a cup type or windmill type. The wind speed measurement unit 231 may have a separate sensor that directly detects the wind speed. The wind speed measuring unit 231 may calculate the wind speed based on the difference between the current attitude angle and the attitude angle in the windless state.

The wind speed measurement unit 231 is configured to be able to measure wind speeds of wind blowing against the drone 100 from all directions. In addition, the wind speed measurement unit 231 may be configured to be able to measure the wind speed in the front-rear direction and the left-right direction, particularly in the normal flight state of the drone 100 .

The wind speed measurement unit 231 may obtain the wind speed in the traveling direction that blows against the drone 100 by subtracting the ground speed from the air speed. Ground speed is the actual realized speed of the drone 100 relative to the ground. The airspeed is the speed when the operating force exerted by the propulsion device of the drone 100 in consideration of the effect of the wind is converted into the speed in a no-wind condition in order to achieve a predetermined ground speed. Since the airspeed in the direction perpendicular to the direction of travel of the drone 100 is 0, the wind speed perpendicular to the direction of travel can be obtained by obtaining the ground speed. The wind speed measurement unit 231 can determine the direction of the wind blowing against the drone 100 by calculating the ground speed and the air speed as a vector with the direction taken into account.

The wind speed measurement unit 231 includes a weight estimation unit 231-1, a ground speed calculation unit 231-2 that calculates the ground speed, and an air speed calculation unit 231-3 that calculates the air speed.

The weight estimation unit 231 - 1 is a functional unit that estimates the total weight m of the drone 100 . The weight estimating unit 231-1 may estimate the total weight m of the drone 100 including the load weight of the load, or estimate the load weight of the load that can change, and then configure a configuration in which the weight does not change, such as By adding the weights of the flight controller 501, rotors 101, motors 102 and other accessories of the drone 100, the total weight m of the drone 100 including payload may be estimated. The variable weight payload is a drug in this embodiment.

The weight estimating unit 231-1 estimates the total weight m of the drone 100 including the load weight of the cargo based on the thrust T in the height direction exerted by the propulsion device when the altitude of the drone 100 does not change. good. This is because the thrust T in the height direction exerted by the propulsion device of the drone 100 is balanced with the gravitational acceleration g that the drone 100 receives when the altitude of the drone 100 does not change.

Weight estimator 231-1 obtains the amount of medicine ejected by integrating the discharge flow rate from medicine tank 104 measured by flow rate sensor 510, and subtracts the amount of medicine ejected from the amount of initially loaded medicine. 104 weight may be estimated. According to this configuration, the weight of the medicine tank 104 can be estimated regardless of the flight state of the drone 100 . Also, the weight estimation unit 231-1 may have a function of estimating the liquid level in the medicine tank 104, for example. The weight estimator 231-1 may estimate the weight using a liquid level gauge, a water pressure sensor, or the like placed inside the drug tank 104. FIG.

The ground speed calculator 231-2 can calculate the ground speed by obtaining the absolute speed in space from the GPS module 504. FIG. Also, the ground speed measurement unit 242-1 can be determined by the GPS modules RTK504-1 and 504-2 that the drone 100 has. Further, the ground speed measuring unit 242-2 can also be obtained by integrating the acceleration of the drone 100 acquired by the 6-axis gyro sensor 505. That is, according to this configuration, the wind speed of the wind blowing against the drone 100 can be obtained with a simple configuration without installing a separate wind speed measurement means on the drone 100 .

（４）
ｇは、重力加速度である。このように、ドローン100の姿勢角θおよび重量ｍに基づいて、ドローン100の対気速度v_aを求めることができる。The airspeed calculator 231-3 can obtain the airspeed based on the attitude angle θ and the weight of the drone 100. FIG. The displacement amount D between the drug dropping point when the drone 100 is flying at an altitude L from the ground and an attitude angle of 0 degrees and the drug dropping point when flying at an attitude angle θ is as follows: Desired.
D=L×tanθ (1)
While the drone 100 is moving at a constant speed or hovering, the following equation holds for the drag force Fd due to air resistance and the airspeed v _a .
Fd=(1/2)×ρv _a 2S×Cd ( ² )
Note that the air density ρ and the air resistance coefficient Cd. The representative area S such as the front projected area is a value obtained in advance based on the size and shape of the drone 100 .
In addition, the following formula holds between the posture angle θ and the drag force Fd.
Fd=mg tanθ (3)
Note that m is the weight of the drone 100 . While the drone 100 is moving at a constant speed or hovering, the airspeed v _a can be obtained by the following formula by solving the formulas (1) and (2).

(4)
g is the gravitational acceleration. Thus, the airspeed v _a of the drone 100 can be obtained based on the attitude angle θ and the weight m of the drone 100 .

The wind direction measurement unit 232 is a functional unit that measures the wind direction of the wind blowing on the drone 100 and its surroundings. The wind direction measurement unit 232 may record the measured wind direction along with time.

The strong wind determination unit 233 is a functional unit that determines a strong wind based on wind speed and wind direction. The strong wind determination unit 233 may determine a temporary strong wind in real time or after the fact. The strong wind determination unit 233 determines that there is a strong wind when the measured wind speed is equal to or higher than a predetermined wind speed. The strong wind determination unit 233 determines that normal shooting is possible when strong wind is not detected. When high winds are detected, the drone 100 may suspend flight and take shelter action. The evacuation action may be, for example, a return to the departure/arrival point 406, or an emergency landing action to land on the spot.

The detection point storage unit 234 is a functional unit that stores detection points where strong winds are detected. The detected points are stored, for example, by three-dimensional coordinates.

The re-capturing unit 24 is a functional unit that flies again to the point where the strong wind is detected by the strong wind determination unit 233 and performs re-capturing. Each configuration of the complementary photographing unit 241 is the same as that of the photographing unit 221. FIG. Complementary photographing unit 241 performs re-photographing when strong wind determination unit 233 does not detect strong wind, in other words, when normal photographing is possible.

●Configuration of yield prediction device 300 The yield prediction device 300 includes a crop shape acquisition unit 31 and a yield prediction unit 32 .

The crop shape acquisition unit 31 is a functional unit that analyzes the image acquired by the imaging unit 221 and acquires information about the shape of the crop. More specifically, the crop shape acquisition unit 31 includes a paddy counting unit 311 , a crop length acquisition unit 312 and a curvature acquisition unit 313 .

The paddy counting unit 311 is a functional unit that counts the number of fruiting paddy based on the image acquired by the imaging unit 221 . The paddy counting unit 311 counts the number of paddy by analyzing chlorophyll by, for example, NDVI analysis based on the ratio of reflected light of red light (about 650 nm wavelength) and near-infrared light (about 774 nm wavelength). The paddy counting unit 311 may perform a calculation process for converting the counted number of paddy into the actual number of paddy bearing on the stalk.

The crop length acquisition unit 312 is a functional unit that acquires the length of the crop from the ground surface. The crop length acquisition unit 312 distinguishes between soil and crops by image analysis, for example, and calculates the crop length.

The curvature acquisition unit 313 is a functional unit that acquires the curvature of the crop in the length direction.

In addition to the above, the crop shape acquisition unit 31 acquires the thickness of the stem, and the yield prediction unit 32, which will be described later, predicts the yield based on the length, curvature, and thickness of the stem. good too. Acquiring the thickness of the stem improves the accuracy of pattern matching for identifying crops with sagging rice, and improves the accuracy of crop yield prediction based on shape.

The yield prediction unit 32 is a functional unit that predicts the yield from the crop based on the information acquired by the crop shape acquisition unit 31, ie, the length, curvature and number of grains. Since crops grow upward from the ground as they grow, the longer the crop, the higher the probability that it is growing. Crops, especially rice, bear seeds at the top of the stem in the lengthwise direction. Therefore, when the paddy swells, the rice stem bends under the weight of the paddy. In addition, the greater the number of rice seeds that bear fruit on one stem, the greater the deflection of the rice stem. Therefore, the amount of paddy harvested from the crop can be predicted by obtaining the curvature of the rice plant in the longitudinal direction. For example, the yield prediction unit 32 predicts that the longer the length of the crop and the greater the curvature, the greater the yield. According to this configuration, the amount of rice can be predicted based on the shape of the stalk, even if the captured image does not show enough rice.

Moreover, if only the length and curvature are used to predict the yield, it is difficult to accurately predict the yield for strains whose stems are bent due to factors other than the weight of the paddy. According to the configuration in which the harvest amount prediction unit 32 predicts the harvest amount based on the length, curvature, and number of grains, the harvest amount can be predicted based on indicators from a plurality of viewpoints. Quantity prediction can be realized.

In addition, after the crop shape acquisition unit 31 acquires any one of the crop length, the curvature in the length direction of the crop, and the thickness of the stem, the yield prediction unit 32 acquires The yield of the crop may be predicted based on any one of the crop length, the curvature of the crop in the longitudinal direction, and the thickness of the stem. In addition, the crop shape acquisition unit 31 may acquire any two of the crop length, the curvature in the longitudinal direction of the crop, and the stem thickness. Alternatively, the thickness and curvature of the stem may be obtained. Then, the yield prediction unit 32 may predict the yield from the crop based on the crop length and stem thickness, or the stem thickness and curvature acquired by the crop shape acquisition unit 31 .

The harvest amount prediction unit 32 makes the harvest amount prediction result different depending on whether or not the strong wind detection unit 23 detects a strong wind at the point. If there is a strong wind blowing in the field 403, the crop will be lodged and will be in a state different from the no wind or low wind conditions. Therefore, the harvest amount prediction unit 32 can more accurately predict the harvest amount by correcting the predicted value when the strong wind is detected. More specifically, when the wind is strong, the curvature of crops increases due to the wind, so the yield prediction unit 32 corrects downward the prediction value obtained based on the curvature.

Moreover, the yield prediction unit 32 may stop predicting the yield at the point when a strong wind is detected. At this time, the user 402 may be notified through the operation device 401 or the like that the prediction has been stopped. The operation device 401 may superimpose and display the unpredictable points on the map of the farm field 403 . The history information about the abortion of harvest amount prediction is stored by the detection point storage unit 234 together with the position of the point.

The harvest amount prediction unit 32 may determine whether to correct the predicted value of the harvest amount or stop the prediction, depending on the degree of the wind. The harvest amount prediction unit 32 corrects the predicted value when the wind speed is greater than or equal to a first threshold that is a strong wind threshold and less than a second threshold that is greater than the first threshold, and corrects the predicted value when the wind speed is greater than or equal to the second threshold. Prediction may be stopped at some point.

The yield prediction unit 32 may calculate a second yield predicted based on the number of grains and correct the second yield based on the length and curvature. For example, the yield prediction unit 32 holds a table in which the second yield and the expected range of the length and curvature of the strain from which the second yield is harvested are associated with each other. The yield prediction unit 32 calculates the assumed crop length and curvature range of the strain based on the second yield. The harvest amount prediction unit 32 determines whether the crop length acquired by the crop length acquisition unit 312 and the curvature acquired by the curvature acquisition unit 313 are included in the ranges of the expected crop length and curvature.

If the obtained values of the crop length and curvature are included in the assumed ranges of the crop length and curvature, it is determined that the second harvest amount has been correctly predicted, and the harvest amount prediction unit 32 calculates the value of the second harvest amount is adopted as the prediction result. If the obtained crop length and curvature values are not within the expected range of crop length and curvature, the second crop is corrected based on the obtained length and curvature values. Specifically, when the obtained values of the length and curvature are larger than the assumed ranges, the yield prediction unit 32 corrects the second yield upward. If the obtained values of the length and curvature are smaller than the assumed range, the yield prediction unit 32 corrects the second yield downward.

The harvest amount prediction unit 32 calculates a first harvest amount predicted based on the length and curvature and a second harvest amount predicted based on the number of unhulled grains, and based on the value of the second harvest amount , whether or not to adopt the first yield as the prediction result. For example, when the difference between the first harvested amount and the second harvested amount is less than a predetermined amount, it is determined that the second harvested amount has been accurately predicted, and the second harvested amount is adopted as the prediction result. When the difference between the first harvested amount and the second harvested amount is large, if the first harvested amount is larger than the second harvested amount, it is judged that the paddy was not seen because it was shadowed by other objects, and the first harvested amount is reduced. Adopted for prediction results. In addition, if the second harvest is extremely small compared to the first harvest, it is judged that there is a high probability that the rice is warped due to factors other than the weight of the paddy, and the second harvest is adopted as the prediction result. may If the second harvested amount is larger than the first harvested amount by a predetermined amount or more, the second harvested amount is adopted as the prediction result. The predicted results of the first yield and the second yield may be provided to the user 402 through the operator 401 or the like. In addition, the user 402 may be notified of information on which harvest amount was adopted as the prediction result for each point.

●Flowchart for drone prediction of yield (1)
As shown in FIG. 11, first, the drone 100 flies in the field 403 and photographs crops (S11). Next, the crop length and curvature are calculated based on the captured image (S12). The yield is then predicted (S13) based on the length and curvature of the crop.

●Flowchart for drone prediction of yield (2)
As shown in FIG. 12, first, the drone 100 flies in the field 403 and photographs crops (S21). Next, the number of unhulled rice is counted to calculate the second yield (S22). Based on the second yield, the assumed crop length and curvature range are calculated (S23). This calculation is performed, for example, by referring to a pre-stored table of the second harvest amount and assumed crop length and curvature ranges. Also, the crop length and curvature are calculated from the state of the stem reflected in the image (S24). Steps S22 and S23 and step S24 may be performed in any order and may be performed simultaneously.

Next, it is determined whether or not the calculated crop length and curvature are within the expected range based on the second harvest (S25). If it is within the expected range, the second yield is adopted as the prediction result (S26). If not within the assumed range, it is determined whether or not the calculated crop length and curvature are greater than the assumed range (S27). If it is larger than the assumed range, the second harvest amount is corrected upward (S28). If it is smaller than the assumed range, the second harvest amount is corrected downward (S29). The correction amount may be constant, or may vary according to the difference between the calculated crop length and curvature and the assumed range.

●Flowchart for drone prediction of yield (3)
As shown in FIG. 13, first, the drone 100 flies in the field 403 and photographs crops (S31). Next, the length and curvature are acquired to calculate the first yield (S32). The number of paddy is counted to calculate the second yield (S33). The order of steps S32 and S33 is random and may be performed simultaneously. It is determined whether or not the difference between the first harvested amount and the second harvested amount is equal to or greater than a predetermined value (S34). When the difference between the first harvested amount and the second harvested amount is less than the predetermined amount, the second harvested amount is adopted as the prediction result (S35). If the difference between the first harvested amount and the second harvested amount is greater than or equal to the predetermined amount and the first harvested amount is greater than the second harvested amount, the first harvested amount is adopted as the prediction result (S37). If the second harvested amount is greater than the first harvested amount, the second harvested amount is adopted as the prediction result (S38).

Flowchart for Detecting Strong Wind and Re-flying As shown in FIG. 14, first, the drone 100 flies over the field 403 and takes an image of the field (S41). When a strong wind is detected while flying over the field 403 (S42), the strong wind detection point is stored (S43). The strong wind detection unit 23 continuously or periodically detects whether there is a strong wind, and determines whether the strong wind has stopped (S44). When it is determined that the strong wind has stopped, the aircraft flies over the detection point again and takes pictures again (S45). Step S44 may be performed at any time during the flight within the farm field 403, or may be performed after the flight within the farm field 403 is completed.

In this explanation, the yield prediction of rice growing in fields has been mainly explained, but this system is not limited to rice and can be applied to other crops. In particular, the present invention can be applied to various crops in which fruits or buds are formed in the upper part of the crop in the longitudinal direction, and the weight of the fruits or buds causes the stem or trunk to bend.

Harvest time prediction system A harvest time prediction unit that predicts the harvest time of the crop based on at least one of crop length, curvature, and stem thickness, and includes an imaging unit 211 and a crop shape acquisition unit. A harvest time prediction system may be configured together with 31. By analyzing the extent of growth, such as crop length, curvature, and stem thickness, it is possible to predict when the plant will be ready for harvest. For example, the harvest time prediction unit may predict the harvest time by referring to a table in which crop lengths, curvatures, or stem thicknesses are associated with predicted harvest times.
The harvest time prediction system may predict the harvest time based on any combination of crop length and stem thickness, stem thickness and curvature, and crop length and curvature. The harvest time prediction system may also predict harvest time based on crop length, curvature, and stem thickness.

In addition, in this description, an agricultural drone has been described as an example, but the technical idea of the present invention is not limited to this, and even if it is a device that photographs and analyzes a field from land, water, or water good. Also, the imaging unit is not limited to being mounted on a moving object, and may be one that does not move.

(Technically Remarkable Effects of the Present Invention)
The drone according to the present invention can accurately predict crop yields.

Claims (21)

an imaging unit that acquires images of growing crops;
a crop shape obtaining unit that obtains a crop length of the crop and a curvature in the length direction of the crop based on the image;
a yield prediction unit that predicts a yield from the crop based on the crop length and the curvature;
a strong wind detection unit that detects the wind blowing on the crops;
with
The harvest amount prediction unit varies the harvest amount prediction result depending on whether the strong wind detection unit detects a strong wind.
Yield prediction system. The harvest amount prediction unit predicts that the longer the crop length and the larger the curvature, the larger the harvest amount.
The yield prediction system according to claim 1. The crop shape acquisition unit further acquires the thickness of the stem of the crop based on the image,
The yield prediction unit predicts the yield based on the crop length, the curvature, and the stem thickness.
The yield prediction system according to claim 1 or 2. The harvest amount prediction unit downwardly corrects the predicted value at the time when the strong wind is detected by the strong wind detection unit.
A yield prediction system according to any one of claims 1 to 3 . When the strong wind is detected by the strong wind detection unit, the harvest amount prediction unit stops predicting the harvest amount at a point where the strong wind is detected.
A yield prediction system according to any one of claims 1 to 4 . further comprising a re-capturing unit that re-acquires an image of the crop at a point where the strong wind is detected when the strong wind is not detected when there is a history of the strong wind being detected;
A yield prediction system according to any one of claims 1 to 5 . The crop is a rice plant bearing husks in the upper part in the longitudinal direction,
Further comprising a paddy counting unit for counting the number of paddy,
The harvest amount prediction unit predicts the harvest amount based on the crop length, the curvature, and the number of the grains.
A yield prediction system according to any one of claims 1 to 6 . The harvest amount prediction unit calculates an assumed range of at least one of the crop length and the curvature in which the harvest amount is harvested from the harvest amount predicted based on the number of the paddy, and obtains the crop shape. correcting the harvest amount based on the obtained value when the obtained value of at least one of the crop length and the curvature obtained by the unit is not included in the assumed range;
The yield prediction system according to claim 7 . an imaging unit that acquires images of growing crops;
a crop shape obtaining unit that obtains a crop length of the crop and a curvature in the length direction of the crop based on the image;
a paddy counting unit that counts the number of paddies that bear fruit on the crop;
a yield prediction unit that predicts the yield from the crop based on the number of the unhulled rice;
with
The harvest amount prediction unit calculates an assumed range of at least one of the crop length and the curvature in which the harvest amount is harvested from the harvest amount predicted based on the number of the paddy, and obtains the crop shape. correcting the harvest amount based on the obtained value when the obtained value of at least one of the crop length and the curvature obtained by the unit is not included in the assumed range;
Yield prediction system. correcting the harvest amount upward when the acquired value of the crop length or the curvature is greater than the assumed range;
The yield prediction system according to claim 8 or 9 . correcting the harvest amount downward when the acquired value of the crop length or the curvature is smaller than the assumed range;
A yield prediction system according to any one of claims 8 to 10 . The harvest amount prediction unit calculates a first harvest amount predicted based on the crop length and the curvature, and a second harvest amount predicted based on the number of the unhulled grains, and calculates the second harvest amount. determining whether to adopt the first yield as a prediction result based on the value;
A yield prediction system according to any one of claims 8 to 11 . an imaging unit that acquires images of growing crops;
a crop shape obtaining unit that obtains a crop length of the crop and a curvature in the length direction of the crop based on the image;
a yield prediction unit that predicts a yield from the crop based on the crop length and the curvature;
a paddy counting unit that counts the number of paddies that bear fruit on the crop;
with
The harvest amount prediction unit calculates a first harvest amount predicted based on the crop length and the curvature, and a second harvest amount predicted based on the number of the unhulled grains, and calculates the second harvest amount. determining whether to adopt the first yield as a prediction result based on the value;
Yield prediction system. When the second harvested amount is greater than the first harvested amount by a predetermined amount or more, the harvested amount prediction unit adopts the second harvested amount as a prediction result.
The yield prediction system according to claim 12 or 13 . At least the imaging unit is mounted on an aircraft,
A yield prediction system according to any one of claims 1 to 14 . obtaining an image of the growing crop;
a crop shape obtaining step of obtaining a crop length of the crop and a curvature in the length direction of the crop based on the image;
predicting yield from the crop based on the crop length and the curvature;
a strong wind detection step of detecting the wind blowing on the crops;
with
In the step of predicting the amount of harvest, the result of predicting the amount of harvest varies depending on whether or not strong wind is detected in the step of detecting strong wind.
Yield prediction method. a photographing step of acquiring an image of the growing crop;
a crop shape obtaining step of obtaining a crop length of the crop and a curvature in the length direction of the crop based on the image;
a paddy counting step of counting the number of paddies that bear fruit on the crop;
a yield prediction step of predicting the yield from the crop based on the number of the paddy;
with
In the harvest amount prediction step, an assumed range of at least one of the crop length and the curvature in which the harvest amount is harvested is calculated from the harvest amount predicted based on the number of the paddy, and the crop shape is acquired. when the acquired value of at least one of the crop length and the curvature acquired in the step is not included in the assumed range, correcting the harvested amount based on the acquired value;
Yield prediction method. a photographing step of acquiring an image of the growing crop;
a crop shape obtaining step of obtaining a crop length of the crop and a curvature in the length direction of the crop based on the image;
a yield prediction step of predicting a yield from the crop based on the crop length and the curvature;
a paddy counting step of counting the number of paddies that bear fruit on the crop;
with
The yield prediction step calculates a first yield predicted based on the crop length and the curvature, and a second yield predicted based on the number of unhulled grains, and calculates the second yield predicted based on the number of grains. determining whether to adopt the first yield as a prediction result based on the value;
Yield prediction method. an instruction to obtain an image of a growing crop;
obtaining a crop length of the crop and a curvature along the length of the crop based on the image;
instructions for predicting yield from said crop based on said crop length and said curvature;
a strong wind detection command for detecting the wind blowing on the crops;
on the computer, and
In the command for predicting the harvest amount, the result of predicting the harvest amount differs depending on whether or not the strong wind is detected by the strong wind detection instruction.
Yield forecasting program. a photographing command for acquiring an image of the growing crop;
a crop shape acquisition command for acquiring a crop length of the crop and a curvature in the length direction of the crop based on the image;
a paddy counting instruction for counting the number of paddies that bear fruit on the crop;
a yield prediction instruction for predicting a yield from the crop based on the number of the paddy;
on the computer, and
In the harvest amount prediction command, an assumed range of at least one of the crop length and the curvature in which the harvest amount is harvested is calculated from the harvest amount predicted based on the number of the paddy, and the crop shape acquisition is performed. correcting the harvest amount based on the obtained value when the obtained value of at least one of the crop length and the curvature obtained by the command is not included in the assumed range;
Yield forecasting program. a photographing command for acquiring an image of the growing crop;
a crop shape acquisition command for acquiring a crop length of the crop and a curvature in the length direction of the crop based on the image;
a yield prediction instruction for predicting a yield from the crop based on the crop length and the curvature;
a paddy counting instruction for counting the number of paddies that bear fruit on the crop;
on the computer, and
The yield prediction instruction calculates a first yield predicted based on the crop length and the curvature, and a second yield predicted based on the number of the unhulled grains. determining whether to adopt the first yield as a prediction result based on the value;
Yield forecasting program.

Images