JP2021108586A - Harvest prediction device and harvest prediction method - Google Patents

Abstract

To provide a harvest prediction device and a harvest prediction method which are capable of improving the prediction accuracy of a harvest index in a farm product.SOLUTION: A harvest prediction device 20 comprises a control part 30 connected to an imaging part, in which the control part 30 specifies a growth state by using an image of a farm product acquired from the imaging part, calculates a residual growth amount required for fulfilling a harvest standard of the farm product by using the growth state, acquires weather information of a cultivation area of the farm product, and calculates a harvest index of the crop by using the weather information and the residual growth amount. The harvest index includes harvest time, yield and harvest probability, which makes it possible to make a harvest plan on the basis of harvest index predicted by the harvest prediction device 20.SELECTED DRAWING: Figure 2

Description

The present invention relates to a harvest prediction device and a harvest prediction method.

Fruits and vegetables reach the harvest time when the integrated environmental value, which is the integrated value of the environmental information, exceeds a certain value. Therefore, a technique for predicting the harvest time of an agricultural product by estimating the current integrated environmental value and the remaining environmental value required for the agricultural product until harvest is disclosed (for example, Patent Document 1).

JP-A-2019-165655

By the way, harvest indicators such as the yield and harvest time of agricultural products are predicted based on the integrated temperature and many years of experience by the harvester. However, in the prediction based on the integrated temperature, a difference is likely to occur between the integrated temperature used for the prediction and the temperature actually integrated in the agricultural product. Forecasts based on years of experience by harvesters are also prone to differences between harvesters with skill and harvesters without that skill. Therefore, it is required to improve the prediction accuracy of the harvest index in agricultural products by using the harvest prediction device.

The purpose of this study is to provide a harvest prediction device and a harvest prediction method that can improve the prediction accuracy of the harvest index in agricultural products.

The harvest prediction device for solving the above problems is a harvest prediction device provided with a control unit connected to the image pickup unit, and the control unit uses an image of the agricultural product acquired from the image pickup unit to determine the growing state. Identify and use the growing condition to calculate the residual growth amount required to satisfy the harvesting standard of the agricultural product, acquire the weather information of the cultivation area of the agricultural product, and obtain the weather information and the residual growth amount. It is characterized in that it is used to calculate a harvest index of the crop.

According to the above configuration, the image of the current agricultural product is reflected in the growing condition, and the harvest index of the agricultural product is used by using the growing condition, the residual growth amount according to the harvest standard, and the weather information from the present to the harvest time. Is expected. For example, when the harvest index is the yield, the current growth rate of the agricultural product is reflected in the forecast of the yield, and the remaining growth amount, which is the number of days from the present to the harvest time, and the current to the harvest time. Weather information is reflected in the yield forecast. In addition, when the harvest index is the harvest time, the current degree of growth of agricultural products is reflected in the prediction of the harvest time, and the residual growth amount required to meet the weight and size that is the harvest standard is calculated from the present. The harvest time is predicted after considering the weather information up to the harvest time. As a result, it is possible to improve the prediction accuracy of the harvest index of agricultural products.

In the harvest prediction device, the control unit specifies the color of the agricultural product contained in the image, and the yield of the agricultural product satisfying the harvest standard at the harvest time is specified as the harvest index according to the color.

According to the above configuration, it is possible to suppress forcing the harvester to harvest agricultural products that do not meet the harvesting standards. In addition, since it is easy to cultivate agricultural products until they meet the harvesting standards, it is possible to improve the accuracy of the yield of agricultural products having high quality.

In the harvest prediction device, the control unit specifies at least one of the weight of the agricultural product predicted by using the image and the size of the agricultural product contained in the image, and the growing state and the amount of photosynthesis of the agricultural product. The residual growth amount required to satisfy the harvest standard is calculated based on the growth amount of the agricultural product with respect to the above, and the harvest time satisfying the harvest standard is specified as the harvest index based on the residual growth amount.

According to the above configuration, it is possible to calculate the residual growth amount required to satisfy the harvest standard based on the growth state and the growth amount of the agricultural product with respect to the photosynthetic amount of the agricultural product. Then, it becomes possible to predict the harvest time, which is a harvest index, from the residual growth amount and the weather information.

The harvest prediction method for solving the above problems is a harvest prediction method using a harvest prediction device provided with a control unit connected to the image pickup unit, and the control unit obtains an image of an agricultural product from the image pickup unit. The growth state is specified using, the residual growth amount required to satisfy the harvest standard is calculated using the growth state, the weather information of the cultivation area of the agricultural product is acquired, and the weather information and the residual growth are obtained. It is characterized in that the harvest index of the agricultural product is calculated using the amount.

According to the above method, the growth state is specified using the image of the agricultural product acquired by the imaging unit, and the residual growth amount is calculated using the growth state, and the weather from the present to the harvest time in the cultivation area of the agricultural product. Using the information and residual growth, the yield index of agricultural products is predicted. As a result, as described above, it is possible to improve the prediction accuracy of the harvest index of agricultural products.

A system configuration diagram showing a configuration of a harvest prediction system for the first embodiment of the harvest prediction system. The block diagram which shows the functional configuration example of the harvest prediction apparatus. A system configuration diagram showing a configuration of a harvest prediction system in the second embodiment of the harvest prediction system. The block diagram which shows the functional configuration example of the harvest prediction apparatus.

[First Embodiment]
The first embodiment of the harvest prediction system including the harvest prediction device and the harvest prediction method executed by the harvest prediction device will be described with reference to FIGS. 1 and 2.

The harvest prediction system of the present embodiment is used to calculate the harvest index of agricultural products. The harvest index of an agricultural product is a value related to the harvest of an agricultural product satisfying the shipping standard, and is, for example, any one of the yield amount, the harvest time, and the harvest probability.

(Harvest index)
The yield, which is an example of the harvest index, is at least one of the weight of the agricultural product to be harvested, the size such as the size in a predetermined direction, and the volume.

The harvest time, which is an example of a harvest index, is determined by, for example, the color of the produce, the integrated temperature, and at least one of any preset periods. The color of agricultural products includes at least one of hue, lightness and saturation. The integrated temperature is the temperature integrated from a certain starting point such as flowering to the time when the agricultural product is harvested. For any preset period, the harvest time is determined by the number of days counted from a certain starting point such as flowering. When the harvest time is determined by the color of the agricultural product, the agricultural product whose color meets the shipping standard or the harvesting standard can be harvested. When the harvest time is determined by the integrated temperature, the agricultural product matures when the integrated temperature exceeds a certain value, and the agricultural product becomes suitable for harvesting.

The harvest probability, which is an example of the harvest index, is, for example, a probability based on a predicted value of the harvest amount, and is a probability that the harvest amount exceeds a certain amount within a specified period. Further, the harvest probability, which is an example of the harvest index, is, for example, a probability based on a predicted value of the harvest time, and is a probability of being harvested within a designated period.

Predicting the yield makes it possible to predict the amount of work required for harvesting, and it is possible to secure the workers required for the harvesting work and to accurately carry out the shipping plan. Predicting the harvest time makes it possible to predict when the agricultural products can be harvested and to make a harvest plan according to the harvest time. Predicting the harvest probability makes it possible to numerically evaluate the reliability of the harvest forecast.

(Agricultural products)
Agricultural products may be vegetables whose harvest site can be photographed in the pre-harvest state, for example, fruit vegetables including tomatoes, eggplants and cucumbers, fruit vegetables including strawberries, melons and squid, cabbage, spinach and lettuce. It is a leafy vegetable including.

Agricultural products are particularly preferably of a type in which the integrated value of environmental information from the starting point such as flowering to the optimum harvest index is constant in the growth process. It is possible to predict the harvest index of agricultural products by the integrated environmental value, and it is possible to predict the harvest index by considering not only the harvest standard but also the maturity level.

Agricultural products are provided with shipping standards that enable them to be shipped or that classify the quality and standards at the time of shipping. Shipping standards are divided into grades and classes, and grades relate to quality, shape, and appearance, such as shape, color, and luster. Class is size and weight.

In addition, for agricultural products, a harvest standard, which is a standard for determining the harvest time of agricultural products, is established. Harvesting standards are standards that meet the grade and class of agricultural products at the time of harvesting. It is desirable that agricultural products that meet the harvesting standards are harvested. The harvest standard may be the same as the shipping standard, or may be set at a value lower than the shipping standard in consideration of the process from harvesting to shipping. In the configuration for predicting the yield as a harvest index or the configuration for predicting the harvest probability based on the harvest, a predetermined harvest time may be included in the harvest standard.

(Cultivation plant)
Agricultural products are cultivated in farms. The growing environment of the plantation may or may not be controlled. The growing environment is a factor that influences the photosynthesis of agricultural products, for example, temperature, carbon dioxide concentration, humidity, light conditions, watering amount, fertilizer, air volume and wind direction. At least one of the growing environments is managed in the cultivation plant where the growing environment is managed.

The plantation may be open, sealed or semi-sealed. Cultivations where the growing environment is controlled are preferably sealed or semi-sealed. A cultivated plant that is sealed or semi-sealed can suppress changes in the cultivated environment in the cultivated environment that are received from the external environment, and it becomes easier to manage the cultivated environment.

The light condition is composed of a long day, a light wavelength, and a light intensity. The long day is the time of the light period in one day, and is the time when the lights are turned on. The daytime is set in consideration of the photoperiodism of agricultural products. Since the light wavelength affects the nutrients, taste, and texture of the agricultural product, it may be selected according to the characteristics of the agricultural product.

The light intensity is the intensity of light radiated to agricultural products, and is adjusted by a solar radiation shielding member that blocks sunlight. It is desirable that the light intensity is equal to or higher than the light compensation point and lower than the light saturation point. When the light intensity is equal to or higher than the light compensation point, it is possible to promote the growth of agricultural products. When the light intensity is below the light saturation point, the cost of the light source is suppressed and the deterioration of the quality of agricultural products caused by excessive photosynthesis is suppressed.

A plantation in which the growing environment is controlled may be configured to control the growing environment of agricultural products based on at least one of a harvest index and a residual growth amount.
For example, in a configuration for predicting the yield as a harvest index, the plantation may include a control panel, a sensor, and a growing environment control device. The control panel controls the growth environment control unit by using the residual growth amount up to a predetermined harvest time and the environmental information transmitted from the environment information unit 50. The growth environment control unit is a device capable of controlling the growth environment in the cultivation plant, and is, for example, an air conditioner, a carbon dioxide generator, a light source, a blower, a humidifier, and a solar radiation shielding member. The sensor measures the growing environment in the plantation and sends it to the control panel. The control panel controls the growth environment control unit so as to form a growth environment suitable for the growth of agricultural products in consideration of the measurement result of the growth environment transmitted from the sensor.

For example, in a configuration for predicting the harvest time as a harvest index, the plantation may further be equipped with photosynthesis optimum software. The photosynthesis optimum software calculates the current optimum photosynthesis rate by using at least one of the residual growth amount and the harvest index transmitted from the harvest prediction device 20 and the environmental information transmitted from the environmental information unit 50. The photosynthesis optimum software transmits the calculated current optimum photosynthesis rate to the control panel. The control panel controls the growth environment control unit based on the current optimum photosynthesis rate received from the photosynthesis optimum software. The control panel controls the growing environment so that the optimum photosynthetic rate can be obtained in consideration of the measurement result of the growing environment transmitted from the sensor.

For example, if the measurement result of the growth environment received from the sensor is a growth environment that can maintain the optimum photosynthesis rate calculated by the photosynthesis optimum software, the growth environment control unit so as to maintain the growth environment. Controlled by. On the other hand, if the measurement result of the growing environment received by the sensor differs from the growing environment in which the optimum photosynthetic rate can be maintained, the growing environment is adjusted by the growing environment control unit so as to approximate the optimum growing environment. Control.

(Structure of harvest prediction system)
The configuration of the harvest prediction system will be described with reference to FIG. The harvest prediction system includes an imaging unit 10, a harvest prediction device 20, and a terminal 40.

The image pickup unit 10, the terminal 40, and the harvest prediction device 20 mutually transmit and receive data through a network. Further, the imaging unit 10 and the terminal 40 may mutually transmit and receive data through the network. Each of the networks used for communication between the image pickup unit 10, the harvest prediction device 20, and the terminal 40 may be a general-purpose communication line such as the Internet, or may be a dedicated communication line for communication of each device. good. Further, the network may include a plurality of independent networks, or may be a common network.

(Image pickup unit)
The image pickup unit 10 takes an image of the agricultural product and transmits the taken image to the harvest prediction device 20. The imaging unit 10 may be fixed or movable. Further, the imaging unit 10 may take a bird's-eye view of the agricultural products, or may take a close-up shot of one or a bunch of agricultural products. The imaging unit 10 may be capable of capturing a moving image. The imaging unit 10 may take pictures at a regular or irregular timing.

(Terminal)
The terminal 40 is a computer terminal such as a smartphone, a tablet terminal, or a personal computer. By inputting the harvest standard, the terminal 40 can transmit the data of the harvest standard to the harvest prediction device 20. The input harvesting standard may be any value and is set as a value that the agricultural product meets at the time of harvesting. Harvesting standards may be entered as multiple parameters.

The terminal 40 displays the harvest index predicted by the harvest prediction device 20. The terminal 40 may be able to view not only the harvest index but also the image of the agricultural product, the growing state, the location of the agricultural product in the cultivation plant, and the data stored in the harvest prediction device 20. Further, the terminal 40 may transmit the data received from the imaging unit 10 and the harvest prediction device 20 to another terminal.

The harvest index displayed on the terminal 40 is, for example, a predicted value of the harvest index or an index expressing a predetermined numerical range including the predicted value by colors, figures, pictures, characters, and the like. In addition, the terminal 40 may be able to display a code indicating the amount, place, and place of agricultural products that can be harvested in the entire cultivation plant during the designated period. The amount of agricultural products that can be harvested may be the yield of the entire plantation or a part of the plantation.

(Harvest prediction device)
The configuration of the harvest prediction device 20 will be described with reference to FIG.
The harvest prediction device 20 is provided with a memory such as a CPU, RAM, and ROM, acquires weather information of the cultivation area of the agricultural product, and calculates the harvest index of the agricultural product using the weather information and the residual growth amount. And. The harvest prediction device 20 has a communication unit 21, a control unit 30, and a calculation model 22 connected to the control unit 30. The harvest index predicted by the harvest prediction device 20 may be one or a plurality.

The communication unit 21 receives and transmits information from the imaging unit 10, the terminal 40, other terminals, and an external information providing center through the network. The communication unit 21 may receive data periodically or irregularly. The communication unit 21 may receive or transmit data each time the harvest index is calculated. Further, the communication unit 21 may select a target for receiving data or data to be received based on an instruction from the control unit 30.

In addition, the communication unit 21 acquires weather information. Information provision of weather information is performed by, for example, an external information provision center, and is, for example, a server installed by the Japan Meteorological Agency or a specific corporation or individual. The weather information is weather information in a cultivation area that is an area within a predetermined range including a cultivation plant where agricultural products are cultivated.

To acquire the weather information of the cultivation area, the weather information may be collectively received on a regular basis, or the weather information may be updated hourly or daily. You may also receive it by contacting an external information provision center.

The weather information may be, for example, a weekly weather forecast or an hourly weather forecast, but an hourly weather forecast is desirable. Hourly weather forecasts can improve the accuracy of predicting the amount of solar radiation per day.

The weather information may be only the weather information, but it is desirable that the weather information has one or more of the sunshine duration, the average temperature, and the probability of precipitation. In particular, the sunshine duration can improve the prediction accuracy of the amount of solar radiation of agricultural products in one day, and the average temperature can improve the prediction accuracy of the integrated temperature. Further, the weather information may be in a table format or a map format. The map format is, for example, a weather map or a rain cloud radar.

The control unit 30 controls the information received by the communication unit 21, and has an image analysis unit 31, a residual growth amount calculation unit 32, a harvest index calculation unit 33, and an information providing unit 34. The control unit 30 is connected to the image pickup unit 10 through the communication unit 21.

The image analysis unit 31 includes, for example, a processor such as a GPU (Graphics Processing Unit) or an FPGA (Field-Programmable Gate Array) configured for image processing. The image analysis unit 31 identifies the growing state of the agricultural product by analyzing the image received from the image pickup unit 10. The growing condition is the color, luster, size, weight, shape, and large amount of fertilizer of the current agricultural products.

The image analysis unit 31 identifies the color of an agricultural product, which is an example of a growing state, with reference to a reference such as a color chart. At this time, the color of the agricultural product may be an average value calculated from the color of the entire agricultural product, or may be the color at a specific part. Alternatively, the image analysis unit 31 may store, for example, a mathematical model constructed based on actual measurement values or a trained model, and specify the growth state of the agricultural product by image analysis using the model.

The image analysis unit 31 may detect individual recognition information when distinguishing individuals of agricultural products. The individual recognition information is information that can distinguish agricultural products individually, and is associated with, for example, the position of the agricultural product in the cultivation plant or the position in the agricultural product. The individual recognition information may be added to the agricultural product by a tag, a color code, a barcode, or a QR code (registered trademark). The individual recognition information makes it possible to recognize an individual even if it is a close-up agricultural product, and it is possible to record the growth process of one individual agricultural product.

The image analysis unit 31 may detect agricultural products that cannot be shipped due to rust fruit, cracked fruit, disease, pest damage, feeding damage, etc. from the image of the agricultural product. The user is notified through the terminal 40 that the agricultural product that cannot be shipped has been detected. The notified user can take appropriate action based on the detection of unshippable agricultural products. Detecting agricultural products that cannot be shipped from the images of agricultural products and notifying the user is, for example, if the agricultural products are damaged by a disease, the damaged agricultural products are removed early, and secondary damage to other agricultural products is caused. It is possible to suppress.

The image analysis unit 31 may create a 3D model of the agricultural product using the images taken by the imaging unit 10 from a plurality of angles, and calculate the volume and weight of the agricultural product from the created 3D model. A mathematical model may be used or a trained model may be used when calculating the volume and weight of the agricultural product from the 3D model of the agricultural product.

(Residual growth amount calculation unit)
The residual growth amount calculation unit 32 executes a process of calculating the residual growth amount required to satisfy the harvest standard of the agricultural product by using the growth state. The residual growth amount calculation unit 32 is provided by the image analysis unit 31 with the growth state of the current agricultural product as an input value. The harvest standard is preset by the provision from the terminal 40 by the user.

The process of calculating the residual growth amount differs depending on the type of harvest index.
For example, in the configuration of predicting the yield as a harvest index or the configuration of predicting the harvest probability based on the yield, the residual growth amount calculation unit 32 calculates the number of days from the present to a predetermined harvest time as the residual growth amount.

For example, in the configuration of predicting the harvest time as a harvest index or the configuration of predicting the harvest probability based on the harvest time, the residual growth amount calculation unit 32 first calculates the residual standard value required to satisfy the harvest standard. The remaining standard value is the weight, size, color, etc. of the agricultural product required to meet the harvest standard. The remaining standard value is calculated by comparing the current growth state of the agricultural product calculated by the image analysis unit 31 with the harvest standard. For example, if the color harvesting standards for produce are Red: 255, Green: 22, Blue: 22, and the current growing conditions for produce are Red: 255, Green: 69, Blue: 22, Green: 47. Is the remaining standard value. For example, when the weight harvest standard for agricultural products is 150 g and the current growing state of agricultural products is 100 g, 50 g is the residual standard value.

Next, the residual growth amount calculation unit 32 calculates the residual growth amount required for the agricultural product to satisfy the harvest standard based on the residual standard value. The residual growth amount is the growth amount of agricultural products required to meet the harvest standard. The residual growth amount is, for example, the amount of afterglow synthesis, which is the amount of photosynthesis required to satisfy the harvest standard, or the residual integrated temperature, which is the integrated temperature required to satisfy the harvest standard. The integrated temperature is the integrated daily average temperature in the area including the cultivation plant. The residual growth amount calculation unit 32 uses the previously calculated residual standard value and the model included in the calculation model 22 when calculating the residual growth amount.

(Harvest index calculation department)
The harvest index calculation unit 33 executes a process of calculating the harvest index by using the residual growth amount calculated by the residual growth amount calculation unit 32 and the weather information of the cultivation area acquired by the communication unit 21.

When predicting the harvest index, the harvest index calculation unit 33 predicts, for example, the daily average temperature in the area from the present to the harvest time from the weather information of the cultivation area from the present to the harvest time. At this time, when the weather information includes the average temperature, the harvest index calculation unit 33 may use the average temperature as the daily average temperature. When the weather information includes the maximum temperature and the minimum temperature, the average value of these may be used as the daily average temperature. If the weather information does not include the temperature, the daily average temperature may be predicted from the weather information, and the monthly average temperature for each weather in the previous year is treated as the weather information, and this monthly average temperature is used. The daily average temperature may be predicted from.

When predicting the harvest index, the harvest index calculation unit 33 predicts, for example, the amount of solar radiation in the area from the present to the harvest time from the weather information of the cultivation area from the present to the harvest time. At this time, the harvest index calculation unit 33 calculates the amount of solar radiation by multiplying the amount of solar radiation in fine weather by the cloudiness coefficient in the forecasted weather. The amount of solar radiation in fine weather is the amount of sunlight radiated to the area including the cultivation plant in fine weather, and is calculated by multiplying the amount of solar radiation directly reaching the normal outside the atmosphere by the atmospheric transmittance in fine weather. .. The atmospheric transmittance in fine weather may be set for each month, for example. The cloudiness coefficient is the cloudiness coefficient, which is the ratio of the cloud-covered portion to the entire sky. For example, the monthly average of cloudy weather, cloudy weather, and rainy weather. The amount of fogging is represented by an integer value in 13 steps of 0+, 0 to 9, 10-10, and the amount of fogging may be indicated by a numerical value in the range of 0 to 1. In the weather information, the cloudiness coefficient is determined by the displayed weather. Further, the harvest index calculation unit 33 may use the average value of the cloudiness coefficient when a plurality of weathers are included in the daily weather information. For example, when the daily weather information is sunny and then cloudy, the harvest index calculation unit 33 calculates the average value of the cloudiness coefficient of sunny and cloudy as the amount of solar radiation. In calculating the harvest index, for example, the process from harvest to shipment may be taken into consideration, and the date from harvest to shipment, movement conditions, etc. are input from the terminal 40 and reflected in the prediction of harvest time. May be done. The harvest index calculation unit 33 may use a bunch of agricultural products as a prediction unit of the harvest index, or may use one agricultural product as a prediction unit of the harvest index.

The process of predicting the harvest index differs depending on the type of harvest index.
For example, in a configuration for predicting the yield as a harvest index or a configuration for predicting the harvest probability based on the harvest, the harvest index calculation unit 33 uses the image analysis unit 31 to calculate the growth state and residual growth amount of the current agricultural product. The harvest index is calculated using the number of days which is the residual growth amount calculated by the calculation unit 32 and the weather information from the present to the predetermined harvest time. At this time, the harvest index calculation unit 33, as an example, uses the weight of the agricultural product at a predetermined harvest time from the daily average temperature predicted from the weather information and the growth amount of the agricultural product predicted from the number of days remaining. Calculate a certain yield. The amount of growth is the amount by which an agricultural product grows by photosynthesis, and is at least one of the weight of the agricultural product, the size such as the size in a predetermined direction, and the volume. The harvest index calculation unit 33 may use the model included in the calculation model 22 when calculating the harvest index.

For example, in a configuration for predicting the harvest time as a harvest index or a configuration for predicting the harvest probability based on the harvest time, the harvest index calculation unit 33 determines the residual growth amount calculated by the residual growth amount calculation unit 32 and the present amount. The harvest index is calculated using the weather information up to the harvest time of. At this time, the harvest index calculation unit 33 outputs, as an example, the time when the integrated temperature predicted from the weather information reaches the remaining integrated temperature as the harvest time. Further, as another example, the harvest index calculation unit 33 outputs the time when the integrated photosynthesis amount obtained from the amount of solar radiation predicted from the weather information reaches the afterglow synthesis amount as the harvest time. The harvest index calculation unit 33 may use the model included in the calculation model 22 when calculating the harvest index. Further, when the weather information includes information on the sunshine duration, the amount of solar radiation, and the light intensity, the harvest index calculation unit 33 may use them for calculating the integrated photosynthesis amount.

The calculation model 22 stores one or more models for calculating the harvest index. The model stored in the calculation model 22 may be a mathematical model calculated based on the actually measured values, or may be a trained model in which machine learning is performed using the teacher data.

For example, a model for calculating the amount of afterglow synthesis as the amount of residual growth correlates the amount of photosynthesis with at least one residual standard value selected from the group consisting of the color, size, weight, and amount of growth of agricultural products. It is a model showing. The model for calculating the afterglow synthesis amount as the afterglow amount is constructed so that, for example, the afterglow synthesis amount required to satisfy the harvest standard is output by using the afterglow standard value Green: 47, which is the amount of change in hue, as an input value. Has been done.

For example, a model for calculating the integrated temperature as the residual growth amount is a correlation between the integrated temperature and at least one residual standard value selected from the group consisting of the color, size, weight, and growth amount of agricultural products. It is a model showing. The model for calculating the residual integrated temperature as the residual growth amount is constructed so that, for example, the residual standard value Green: 47, which is the amount of change in hue, is used as an input value, and the residual integrated temperature required to satisfy the harvest standard is output. Has been done.

For example, in a configuration that predicts the yield as a harvest index, or a configuration that predicts the harvest probability based on the yield, the model for calculating the harvest index is the color, luster, size, weight, shape, and the current agricultural product. In addition, the growth state such as a large amount of fertilizer, the number of days remaining to grow, and the daily average temperature based on the weather information from the present to the predetermined harvest time are input. The model for calculating the harvest index is constructed to output the yield by using the correlation between the growth state, the number of days, and the weather information and the yield, which is the weight of the agricultural product.

For example, in the configuration of predicting the harvest time as a harvest index or the configuration of predicting the harvest probability based on the harvest time, the model for calculating the harvest index is the residual growth amount such as the afterglow synthesis amount and the integrated temperature, and the present. Enter the amount of solar radiation based on the weather information up to the specified harvest time. The model for calculating the harvest index is constructed to output the harvest time by using the correlation between the residual growth amount and the weather information and the harvest time.

The residual growth amount calculation unit 32 and the harvest index calculation unit 33 may use the mathematical model properly according to the maturity of the agricultural product based on the integrated environmental value and the color of the agricultural product. For agricultural products in which the growth rate of agricultural products changes depending on the maturity of agricultural products with respect to the amount of photosynthesis, the growth amount of agricultural products is calculated more accurately by using the calculation model 22 that matches the maturity of agricultural products. It becomes possible.

The information providing unit 34 creates information to be displayed on the terminal 40. The information providing unit 34 generates information including the harvest index calculated by the harvest index calculation unit 33. The information including the harvest index may include, for example, an image of the agricultural product, a growth state, the amount of afterglow synthesis, the location of the agricultural product in the cultivation plant, and information for distinguishing the individual of the agricultural product. Displaying the harvest amount and the harvest time on the terminal 40 makes it possible to grasp the amount of work required for harvesting and facilitates the making of a harvest plan. The fact that the harvest probability is displayed on the terminal 40 makes it possible to grasp the credibility of the harvest time.

The harvest prediction device 20 may transmit data such as residual growth amount and weather information used for predicting the harvest index, and the prediction result of the harvest index to the plantation. In addition, the farm may control the growing environment of the agricultural product based on the data transmitted from the harvest prediction device 20.

Controlling the growing environment of agricultural products based on the data transmitted from the harvest prediction device 20 suppresses a significant change in the harvest index due to the climate. It is also possible to control the growing environment so that the harvest index becomes an arbitrary value. Controlling the growing environment of agricultural products based on the data transmitted from the harvest prediction device 20 makes it possible to easily make a harvest plan.

(Forecasting method of harvest index)
The method of predicting the harvest index executed by the harvest prediction device 20 will be described.
The method of predicting the harvest index is a harvest prediction method using a harvest prediction device provided with a control unit connected to the image pickup unit, and the control unit uses an image of an agricultural product acquired from the image pickup unit to grow. Is specified, the residual growth amount required to satisfy the harvest standard is calculated using the growth state, the weather information of the cultivation area of the agricultural product is acquired, and the weather information and the residual growth amount are used. It is characterized in that the harvest index of the agricultural product is calculated.

Identifying the growth state using the image of the agricultural product acquired from the imaging unit 10 means that the image analysis unit 31 analyzes the image of the agricultural product taken by the imaging unit 10 to determine the growth state of the agricultural product. To identify.

The image pickup unit 10 takes an image of an agricultural product at the cultivation plant. The imaging unit 10 transmits the captured image of the agricultural product to the harvest prediction device 20. The harvest prediction device 20 recognizes the agricultural product from the image of the agricultural product in the image analysis unit 31 and performs image analysis. The image analysis unit 31 detects a bunch or a single agricultural product. The image analysis unit 31 identifies the growing state of the agricultural product by analyzing the image of the detected agricultural product.

To calculate the residual growth amount required to meet the harvest standard using the growth state, the residual growth amount calculation unit 32 calculates the residual growth amount required to meet the harvest standard using the growth state. That is.

The harvest index calculation unit 33 calculates the harvest index of the agricultural product by acquiring the weather information of the cultivated area of the agricultural product and using the weather information and the residual growth amount. The harvest index calculation unit 33 acquires the weather information of the cultivation area of the agricultural product in the communication unit 21, and calculates the harvest index of the agricultural product by using the weather information and the residual growth amount.

In the configuration of predicting the yield as a harvest index or the configuration of predicting the harvest probability based on the yield, the residual growth amount calculation unit 32 calculates the number of days from the present to a predetermined harvest time as the residual growth amount.

On the other hand, in the configuration of predicting the harvest time as a harvest index or the configuration of predicting the harvest probability based on the harvest time, the residual growth amount calculation unit 32 first first determines the residual standard values such as color and weight required to satisfy the harvest standard. Is calculated. Next, the residual growth amount calculation unit 32 calculates the residual growth amount such as the photosynthetic amount and the integrated temperature required for the agricultural product to satisfy the harvest standard.

The predicted harvest index is processed and converted by the information providing unit 34 so that it can be displayed on the terminal 40. The data that can be displayed on the terminal 40 by the information providing unit 34 is transmitted to the terminal 40. The terminal 40 that has received the data provides the prediction result of the harvest index by the harvest prediction device 20.

If the harvest time depends on the temperature, the harvest time may be calculated by predicting the temperature from the weather information.
After specifying the harvest time by the color of the agricultural product, the yield that meets the harvest standard may be predicted.

As described above, according to the above embodiment, the following effects can be obtained.
(1) The image analysis unit 31 can analyze the growth state of the agricultural product from the image taken by the image pickup unit 10. This makes it possible to reduce the burden of manually measuring the growth state of agricultural products.

(2) The harvest prediction device 20 predicts the harvest index by using the growth state, the residual growth amount, and the weather information from the present to the harvest time in the current agricultural product. The fact that the harvest index is predictable by the harvest index calculation unit 33 makes it possible to easily grasp the situation until the agricultural product is harvested. Therefore, it becomes easier to make a harvest plan.

(3) Then, when the harvest index is the harvest amount, the current growth degree of the agricultural product is reflected in the forecast of the harvest amount, and the remaining growth amount which is the number of days from the present to the harvest time and the harvest from the present. The temperature up to the time is reflected in the forecast of yield. In addition, when the harvest index is the harvest time, the current degree of growth of agricultural products is reflected in the prediction of the harvest time, and the residual growth amount required to meet the weight and size that is the harvest standard is calculated from the present. The harvest time is predicted after considering the temperature up to the harvest time. As a result, it is possible to improve the prediction accuracy of the harvest index of agricultural products.

(4) The fact that the amount of afterglow synthesis can be calculated by the residual growth amount calculation unit 32 means that the amount of photosynthesis required by the harvest time of the agricultural product can be grasped, and the degree of growth of the agricultural product can be grasped numerically. Is possible. Therefore, the provision of an index for grasping the growth of agricultural products makes it easy to grasp the growth status of agricultural products.

(5) The fact that the harvesting standard is at least one of color, weight, and size can improve the shipping amount shipped in a state where the shipping standard is satisfied. In addition, it is possible to suppress the harvesting of agricultural products whose color, weight, and size do not meet the harvesting regulations at the time of harvesting, and to improve the yield and quality of the agricultural products at the time of harvesting. In addition, since it is easy to cultivate agricultural products until they meet the shipping standards, it is of high quality that the harvesting standard for specifying the residual growth amount is at least one of color, weight, and size. It is also possible to improve the accuracy of the yield or the harvest time of the produced agricultural products.

(6) Based on the harvesting standard input on the terminal 40, the time when the agricultural product satisfying the harvesting standard can be harvested is calculated. Therefore, the agricultural products predicted to be harvestable by the harvest prediction device 20 increase the probability of being in a state of satisfying the harvest standard.

[Second Embodiment]
The second embodiment of the harvest prediction system including the harvest prediction device will be described with reference to FIGS. 3 and 4, focusing on the differences from the first embodiment. The harvest prediction system according to the second embodiment has an environmental information unit, a log control unit, a learning unit, a log storage unit, a learning information storage unit, and a learning result storage unit, and the harvest time is determined by the learning unit. It differs from the first embodiment in that the prediction of is learned. In FIGS. 3 and 4, substantially the same configurations as those described in the first embodiment are designated by the same reference numerals, and duplicate description will be omitted.

(Environmental Information Department)
As shown in FIG. 3, the harvest prediction system further includes an environmental information unit 50. The environmental information unit 50 generates environmental information including the current temperature, carbon dioxide concentration, and light conditions in the area including the plantation. The environmental information unit 50 may acquire the environmental information on a regular basis or may acquire the environmental information irregularly. Further, in conjunction with the imaging unit 10, the imaging unit 10 may acquire environmental information at the time of harvesting the agricultural product. It is possible to associate the image captured by the imaging unit 10 with the environmental information.

The environmental information unit 50 transmits the actual environmental information in the area including the generated cultivation plant to the harvest prediction device 20 through the network. Further, the environment information unit 50 may transmit and receive data to and from the imaging unit 10 and the terminal 40. Each of the networks used for communication between the environment information unit 50, the imaging unit 10, the harvest prediction device 20, and the terminal 40 may be a general-purpose communication line such as the Internet, or a dedicated communication for communication of each device. It may be a line. Further, the network may include a plurality of independent networks, or may be a common network.

(Harvest prediction device)
As shown in FIG. 4, the harvest prediction device 20 further includes a log control unit 35 and a learning unit 36 in the control unit 30. Further, in addition to the calculation model 22, the log storage unit 23, the learning information storage unit 24, and the learning result storage unit 25 are connected to the control unit 30.

The log control unit 35 controls the storage of information related to agricultural products generated by the imaging unit 10, the environmental information unit 50, and the harvest prediction device 20. The log control unit 35 associates the data transmitted from the imaging unit 10, the environment information unit 50, and the harvest prediction device 20 with the date and time, and stores the data in the log storage unit 23.

The learning unit 36 performs machine learning using the learning information stored in the learning information storage unit 24. The learning unit 36 acquires a learning result for obtaining an optimum harvest time value from a plurality of values included in the learning information. Machine learning may be supervised learning or unsupervised learning. Further, the supervised learner may be an arithmetic unit that performs linear regression, a support vector machine, a decision tree, a random forest, or Ada. It may be a Boost or a neural network.

The learning unit 36 stores the learning result in the learning result storage unit 25, and corrects the models used in the image analysis unit 31, the residual growth amount calculation unit 32, and the harvest index calculation unit 33 based on the learning result. ..

The log storage unit 23 stores a log in which the information about the agricultural products acquired by the imaging unit 10, the environmental information unit 50, and the harvest prediction device 20 is associated with the date and time. The log storage unit 23 stores, for example, a record in which at least one of the environmental information, the captured image, and the calculation model is associated with the date and time information.

The learning information storage unit 24 stores learning information. The learning information is data used when the learning unit 36 performs a learning process. The learning information is information in which the information and model used for predicting the harvest index, the predicted harvest index, and the evaluation information for measuring the appropriateness of the harvest index are associated with each other. The learning information may include environmental information in the area including the agricultural product, an image of the agricultural product, and information in which the environmental information and the date and time when the image was acquired are associated with each other.

The learning result storage unit 25 stores information indicating the learning result obtained by the learning process of the learning unit 36. The learning result stored by the learning result storage unit 25 is used by the image analysis unit 31, the residual growth amount calculation unit 32, and the harvest index calculation unit 33.

As described above, according to the present embodiment, the above-mentioned effects (1) to (6) can be obtained, and the following effects can be further obtained.
(7) The environmental information unit 50 generates environmental information in the cultivation plant. Therefore, the harvest prediction device 20 can acquire information on the growing environment in which the agricultural product grows, and can associate the growing environment with the agricultural product. As a result, the harvest prediction device 20 can cause the learning unit 36 to perform the learning process in consideration of the relationship between the current growing state of the agricultural product and the growing environment up to the present.

(8) The learning unit 36 calculates a learning result from the learning information, and stores the learning result in the learning result storage unit 25. The image analysis unit 31, the residual growth amount calculation unit 32, and the harvest index calculation unit 33 can predict the harvest index more accurately by correcting the calculation model 22 based on the learning result.

[Other Embodiments]
The above embodiment can be modified as follows.
-The growing environment controlled in the cultivation plant may be one or more of temperature, carbon dioxide concentration, and light conditions, or all of them may not be controlled. The fact that the growing environment is not controlled in the farm or the number of controlled growing environments is small means that it is not necessary to have a complicated growing environment management system for growing agricultural products in the farm. In addition, it is possible to reduce the cost incurred in managing the growing environment.

-The harvest prediction device 20 may calculate the temperature in the optimum growing environment for the growth of agricultural products from the amount of solar radiation calculated from the weather information. The harvest prediction device 20 may calculate the amount of photosynthesis of agricultural products from the amount of solar radiation and the optimum temperature, and calculate the harvest index based on the amount of photosynthesis.

-In the cultivation plant, the growing environment may be controlled to a certain value or regularly. In a cultivation plant where the growing environment is controlled to a certain value or regularly, the growing environment can be easily managed.
-The growth state is not limited to the information generated from the image, but may be an actually measured value measured using the current agricultural product.

10 ... Imaging unit 20 ... Harvest prediction device 21 ... Communication unit 22 ... Calculation model 23 ... Log storage unit 24 ... Learning information storage unit 25 ... Learning result storage unit 30 ... Control unit 31 ... Image analysis unit 32 ... Residual growth amount calculation Department 33 ... Harvest index calculation department 34 ... Information provision department 35 ... Log control department 36 ... Learning department 40 ... Terminal 50 ... Environmental information department

Claims (4)

It is a harvest prediction device equipped with a control unit connected to the image pickup unit.
The control unit
The growth state was identified using the image of the agricultural product acquired from the imaging unit, and the growth state was specified.
Using the growing condition, the residual growing amount required to meet the harvesting standard of the agricultural product was calculated.
A harvest prediction apparatus characterized in that the weather information of the cultivation area of the agricultural product is acquired, and the harvest index of the agricultural product is calculated by using the weather information and the residual growth amount. The control unit
Identify the color of the agricultural product contained in the image and
The harvest prediction device according to claim 1, wherein the yield of the agricultural product satisfying the harvest standard at the time of harvest is specified as the harvest index according to the color. The control unit
At least one of the weight of the produce predicted using the image and the size of the produce contained in the image is specified.
Based on the growing state and the growing amount of the agricultural product with respect to the photosynthetic amount of the agricultural product, the residual growing amount required to satisfy the harvesting standard was calculated.
The harvest prediction device according to claim 1, wherein a harvest time satisfying the harvest standard is specified as the harvest index based on the residual growth amount. It is a harvest prediction method using a harvest prediction device equipped with a control unit connected to an imaging unit.
The control unit
The growth state was identified using the image of the agricultural product acquired from the imaging unit, and the growth state was specified.
Using the above growing conditions, the amount of residual growth required to meet the harvesting standards was calculated.
A harvest prediction method characterized in that the weather information of the cultivation area of the agricultural product is acquired, and the harvest index of the agricultural product is calculated by using the weather information and the residual growth amount.

Images