CN115803607A

CN115803607A - Method for obtaining protein content of grain

Info

Publication number: CN115803607A
Application number: CN202080102262.2A
Authority: CN
Inventors: 渡桥启介; 崔明达; 河野元信
Original assignee: Satake Manufacturing Suzhou Co Ltd; Satake Corp
Current assignee: Satake Manufacturing Suzhou Co Ltd; Satake Corp
Priority date: 2020-09-29
Filing date: 2020-09-29
Publication date: 2023-03-14
Also published as: WO2022070232A1; JPWO2022070232A1

Abstract

The present invention is characterized by having: a protein content prediction step for obtaining a predicted value of the protein content of grain in a mature period by remote sensing; a protein content actual measurement step of acquiring a protein content actual measurement value of the grain from a sample of the grain to be harvested by ground truth; and a protein content correction step of correcting the predicted protein content value using the measured protein content value to obtain a corrected predicted protein content value of the grain.

Description

Method for obtaining protein content of grain

Technical Field

The present invention relates to a method for obtaining the protein content of grains.

Background

Conventionally, as a method for predicting the protein content of rice, there is a method for predicting the protein content of rice before the time when rice is actually harvested (see patent document 1). That is, rice in a field before harvesting is photographed from the ground to obtain a predicted value of the protein content of rice, and the rice is classified based on a predetermined protein content as a threshold value and harvested.

As a method for classifying and drying harvested rice received by a grain processing facility based on a protein content, there is known a method in which a sample is collected from the harvested and received rice, the protein content is measured with an optical analysis device, and the classified rice is stored after being dried (see patent document 2).

Documents of the prior art

Patent document

Patent document 1: japanese patent laid-open publication No. 2006-101768

Patent document 2: japanese patent laid-open publication No. 2018-179436

Disclosure of Invention

Problems to be solved by the invention

However, in the method disclosed in patent document 1, since the field before the actual harvesting period is photographed by remote sensing and the protein content of the rice is predicted, there is a possibility that the protein content of the rice may change due to the influence of growth, weather conditions, and the like until the rice is harvested.

In the method disclosed in patent document 2, the protein content is measured from the accepted rice, and there is a possibility that the acceptance congestion occurs in the grain processing facility because there is a time loss until the measurement result is obtained.

Therefore, an object of the present invention is to provide a method for obtaining a grain protein content, which can predict the protein content of a grain with high accuracy.

Means for solving the problems

The invention of scheme 1 is a method for obtaining the protein content of a grain, characterized in that: the acquisition method comprises: a protein content prediction step for obtaining a predicted value of the protein content of grains in a maturation stage by remote sensing; a protein content actual measurement step of acquiring an actual measurement value of the protein content of the grain from a sample of the grain to be harvested by ground truth; and a protein content correction step of correcting the predicted protein content value on the basis of the measured protein content value to obtain a corrected predicted protein content value of the grain.

The method for obtaining the protein content of grain according to the embodiment 1, the invention of the embodiment 2 is characterized in that: the protein content prediction step includes the steps of: a grid consisting of a plurality of sections is covered with image data of a field imaged by remote sensing, the predicted protein content value is acquired for each of the plurality of sections, and a section having the predicted protein content value with the smallest deviation from the peripheral section is determined as a reference field for collecting a sample of the grain.

The method for obtaining the protein content of the grain according to the

embodiment

1 or 2, the invention of the embodiment 3 is characterized in that: the remote sensing device includes an unmanned aerial vehicle capable of autonomous flight, and an imaging unit is mounted on the unmanned aerial vehicle.

The method for obtaining the protein content of the grain according to any one of the aspects 1 to 3, the invention of the aspect 4 is characterized in that: the method for obtaining the protein content correction prediction value includes a step of setting a predetermined threshold value for the protein content correction prediction value.

The method for obtaining the protein content of the grain according to any one of the aspects 1 to 4, wherein the invention of the aspect 5 is characterized in that: the acquisition method includes a distribution map display step of plotting and displaying a distribution range of the protein content corrected predicted values in the field.

The method for obtaining the protein content of grain according to claim 5, the invention of claim 6 is characterized in that: the method includes a fertilization performance recording step of recording fertilization performance in the field, and the distribution map display step can display the mapping data of the field by synthesizing the fertilization performance.

Effects of the invention

According to the invention of claim 1, the predicted protein content value of the grain at the maturation stage obtained by remote sensing can be corrected using the measured protein content value of the grain obtained from the grain sample immediately before harvesting, and the corrected predicted protein content value of the grain can be obtained. With such a configuration, the protein content of the grain to be harvested in the field can be predicted with higher accuracy than in the conventional case. Furthermore, the protein content can be predicted with high accuracy at the time of harvesting. This enables, for example, rice with a low protein content and a good taste to be classified and harvested, thereby promoting improvement and stabilization of yield and quality. Further, since the grains can be classified and harvested according to the protein content, the burden of classification in the grain treatment facility is reduced, and the benefit can be extended without investing in equipment related to the classification device.

According to the invention of claim 2, in the protein content prediction step, a grid composed of a plurality of sections is overlaid on the captured data of the captured field, predicted protein content values in the respective sections are acquired, and a section having a predicted protein content value with the smallest deviation from the peripheral section is determined as a reference field in which the grain sample is collected in the ground truth. With such a configuration, even if there is a slight positional error due to GPS or the like when a grain sample is collected, it is possible to reach an accurate position in the reference field. This makes it possible to correct the predicted protein content value with high accuracy.

According to the invention of claim 3, low-altitude flight and planned route flight can be easily performed by using an unmanned aerial vehicle capable of autonomous flight in remote sensing. In addition, in the next year and later, when the field is again photographed, the photographed data of the field can be acquired along the same flight path.

According to the invention of the embodiment 4, a predetermined threshold value is set for the protein content correction predicted value. This enables high-quality grains, such as rice with a low protein content and a good taste, to be classified and harvested.

According to the invention of claim 5, the harvesting plan in the field can be efficiently prepared by plotting and displaying the distribution range of the protein content correction prediction values in the field.

According to the invention of claim 6, the results of fertilization in the field can be recorded, and the results of fertilization can be combined with the mapping display and displayed as an output. Thus, for example, based on the protein content of grains harvested in the previous year and the actual results of fertilization in the previous year, fertilization can be performed appropriately in the next year, and grains of higher quality can be harvested.

Drawings

Fig. 1 is a perspective view showing a remote sensing system according to an embodiment of the present invention.

Fig. 2 is a plan view of a field in which field information in each section is calculated and displayed on captured field data according to one embodiment of the present invention.

Fig. 3 is an enlarged top view of a portion a in fig. 2.

FIG. 4 is a flow chart of a first step of an embodiment of the present invention.

FIG. 5 is a flow chart of a second step of an embodiment of the present invention.

FIG. 6 is a flow chart of a third step of an embodiment of the present invention.

Fig. 7A is a graph showing the relationship between the yield and the protein content before and after correction of the protein content in one embodiment of the present invention, and is a graph when the content of protein after correction is predicted to be lower than the content of protein before correction.

Fig. 7B is a graph showing the relationship between the yield and the protein content before and after correction of the protein content in one embodiment of the present invention, and is a graph when the content of protein after correction is predicted to be higher than the content of protein before correction.

Fig. 8 shows a map display of a field according to an embodiment of the present invention.

Detailed Description

Hereinafter, an embodiment of a method for obtaining the protein content of the grain of the present invention will be described based on the drawings. By using this embodiment, the image data 2 of the field 1 can be acquired by remote sensing, and the protein content of the grain at the time of harvesting can be predicted and acquired with high accuracy.

As shown in fig. 1, in the present embodiment, the unmanned aerial vehicle 10 capable of autonomous flight, which is equipped with a camera 11 as imaging means, captures an image from above a field 1 in which rice is cultivated, and acquires imaging data 2 of the field 1. The unmanned aerial vehicle 10 is equipped with GPS equipment, an electronic compass, an acceleration sensor, and the like, and is configured to be able to easily fly at low altitude or along a planned flight route according to a program.

According to the configuration of the unmanned aerial vehicle 10 as described above, when the field 1 is imaged again in the next year or later, for example, the imaging data 2 of the field 1 can be acquired on the basis of data such as flight history along the same flight route as in the previous year. The captured data 2 of the field 1 may be recorded in a storage medium mounted on the unmanned aerial vehicle 10, but the captured data 2 of the field 1 may be wirelessly transmitted to a computer, not shown, serving as a remote sensing processing device in real time.

The camera 11 in the present embodiment includes a CCD having a predetermined resolution, and can capture near-infrared images through a band-pass filter in addition to visible images. Further, the remote sensing processing device performs image analysis of the captured data 2, and can input/output field information 4 in the field 1 shown in fig. 2.

As shown in fig. 2, the remote sensing processing device performs processing for covering the grid 3 formed of a plurality of segments 31 with the captured data 2 of the captured field 1, and calculates field information 4 in each segment 31. Fig. 2 shows a plan view of the field 1 in which the field information 4 in each section 31 is displayed in the imaging data 2 of the field 1, and a value obtained by converting the Ratio Vegetation Index (RVI) obtained by remote sensing into the leaf nitrogen content is output as the field information 4.

In the present embodiment, in order to improve the accuracy of predicting the leaf nitrogen content of each segment 31 obtained by remote sensing, as shown in fig. 3, a reference field 32 is determined from a plurality of segments (a and b are about 3m each) in the field 1. In fig. 3, the leaf nitrogen content value corresponding to the predicted value of the protein content of rice is shown for each segment 31. When the reference field 32 is determined, 9 sections having the smallest variation in leaf nitrogen content of the sections 31 are searched for by computer processing, and the section 31 at the center thereof is determined as the reference field 32. When the 9 sections are confirmed at a plurality of places in the field 1, the section 31 having the smallest leaf nitrogen content in the section 31 at the center of each of the 9 sections is determined as the reference field 32.

According to the above-described method of determining the reference field 32, when the rice leaf or rice sample is collected in the determined reference field 32, even if the positional error due to the GPS or the like is about 2 meters, the rice leaf or rice sample can be collected by reaching a certain position of the reference field 32 with a substantial accuracy.

The leaf nitrogen content and the protein content of the rice leaf samples collected in the reference field 32 can be obtained by measuring the leaf and rice samples with a dedicated optical analyzer. For example, a taste meter (model: RPTA11B, manufactured by 124699, 12465798. However, since the protein content is different between the state of the rice and the state of the treated brown rice, the protein content of the brown rice can be obtained by performing conversion processing based on the actual result data. The leaf nitrogen content of the rice leaf can be measured by a known leaf nitrogen content measuring instrument, and the leaf nitrogen content is in a correlation with the protein content of the rice or the brown rice, so that the protein content of the rice or the brown rice can be predicted according to the leaf nitrogen content obtained by remote sensing.

Next, an example of each execution step will be described with respect to a method for obtaining the protein content of the grain in the present embodiment. The embodiment described below is an embodiment for obtaining the protein content of rice.

(first step)

Fig. 4 shows the process of fertilizer application diagnosis at the growth stage. Leaf nitrogen content information of rice plants in the field 1 was obtained by remote sensing from a young ear formation period 2 to 3 weeks before ear emergence to about a meiosis period (S100). At the same time as the execution timing of the remote sensing, a sample of the rice leaf is collected in the reference field 32 of the field 1 by the ground truth, and the leaf nitrogen content information of the rice leaf is acquired by the leaf nitrogen content measuring instrument (S101).

Next, leaf nitrogen content information of rice plants in field 1 obtained by remote sensing is corrected based on leaf nitrogen content information of a sample of rice leaves in reference field 32 obtained by ground truth (S102). That is, by correcting leaf nitrogen content information of a rice plant obtained by analyzing the imaging data 2 of the field 1 based on leaf nitrogen content information actually measured from a sample of rice leaves, it is possible to improve the prediction accuracy of leaf nitrogen content information of a rice plant obtained by remote sensing.

In the present embodiment, the fertilizer application performance recording step of recording the fertilizer application performance in the field 1 is provided, and the past fertilizer spraying performance is recorded and accumulated. Therefore, in the present embodiment, a fertilization diagnosis corresponding to the target quality is performed based on the fertilization performance data accumulated so far and the quality data of the harvested rice (S103). For example, a fertilization diagnosis for harvesting more tasty rice having a protein content of 6.5% or less is performed. That is, the first step is a fertilization diagnosis step as follows: accurately controlling the leaf nitrogen content of the rice plant from the young ear forming period to the meiosis period, and controlling the amount of the fertilizer sprayed to obtain more rice with good taste.

(second step)

Fig. 5 shows a process of the step of predicting the protein content of rice in the mature period. In the maturity stage 2 to 3 weeks before harvest, leaf nitrogen content information of the rice plants in the field 1 was obtained by remote sensing (S200). At the same time as the execution timing of the remote sensing, a sample of the rice leaf is collected in the reference field 32 of the field 1 by the ground truth, and the leaf nitrogen content information of the rice leaf is acquired (S201).

Next, leaf nitrogen content information of rice plants in the field 1 obtained by remote sensing is corrected based on leaf nitrogen content information of a sample of rice leaves in the reference field 32 obtained by live ground conditions (S202). Then, based on the leaf nitrogen content information of the corrected rice plant obtained in S202, a predicted protein content value of rice in the mature period is calculated (S203). Based on the predicted values of protein content of the rice in field 1 thus obtained, a harvesting plan for classifying and harvesting rice that is good-tasting rice is prepared in advance as a primary classification (S204). For example, the predicted value of the protein content of the rice that becomes the tasteless rice is set as a threshold value, and the section 31 of the field 1 in which the rice that becomes the tasteless rice is distributed is identified to make a harvest plan. Thus, the rice grains can be classified with high accuracy to obtain rice with good taste.

(third step)

Fig. 6 shows a process of correcting the predicted value of the protein content of rice calculated in the second step to the protein content of rice to be harvested. A sample of rice is collected in the reference field 32 in the field 1 about 2 to 3 days before the expected harvest date immediately before harvesting, and a measured protein content value of the rice is obtained using a taste meter (registered trademark) (S300). Then, the difference between the actual measured value of the protein content of the rice obtained in the actual measurement and the predicted value of the protein content of the rice in the reference field 32 obtained in the second step is obtained. Then, the difference is uniformly added and subtracted for each of the plurality of sections 31, thereby calculating a corrected predicted protein content value (S301).

The predicted value is corrected based on the protein content of the paddy in field 1 obtained as described above, and an actual harvest schedule is prepared as a secondary classification (S302). For example, the corrected and predicted value of the protein content of the rice that becomes the delicious rice can be set as a threshold value, and the section 31 of the field 1 in which the rice that becomes the delicious rice is distributed can be identified to make a harvest plan. Thus, the rice grains can be classified with high accuracy to be harvested as rice with good taste.

Fig. 7A and 7B show the relationship between the yield of rice and the protein content P before and after correction from the measured value of the protein content of rice obtained immediately before harvesting. In fig. 7A and 7B, the distribution range of the protein content corrected predicted values of the rice in field 1 after correction is indicated by a solid line, and the distribution range of the protein content predicted values of the rice in field 1 before correction is indicated by a broken line.

As shown in the corrected predicted distribution of fig. 7A, it is found that more rice is distributed in the range of rice as good taste than the predicted distribution before correction. In this way, the protein content of the rice collected as a sample is actually measured 2 to 3 days before the expected harvesting date of harvesting, and the corrected predicted protein content value is obtained, whereby the protein content of the rice closer to the actual protein content can be predicted. This enables a large amount of rice with good taste to be reliably classified and harvested.

Fig. 7B shows an example in which the distribution range of rice, which is good-tasting rice, is reduced compared to the distribution range before correction. Even in such a case, it can be grasped that the yield of rice as the rice having a good taste is reduced contrary to the initial prediction. Moreover, the sections 31 of the field 1 in which the paddy as the tasteful rice is distributed can be accurately specified, and the paddy as the tasteful rice can be reliably classified and harvested. Further, by accumulating such quality data together with the actual fertilization results, the fertilization schedule of the next year and later can be re-evaluated.

The method for obtaining the protein content of rice according to the present embodiment is explained above. By the above method, the protein content of the rice can be predicted with high accuracy, and a harvest schedule corresponding to the quality of the rice can be prepared, thereby promoting improvement and stabilization of yield and quality. Further, after the rice is received by the grain processing facility as in the related art, the loss in time due to the measurement and classification of the protein content of the rice can be reduced. In addition, it is not necessary to add storage facilities for storing received rice in order to cope with the reception congestion in the grain processing facility, and it is not necessary to invest additional facilities in the grain processing facility, and it is possible to expand the profit.

Fig. 8 shows an example of a distribution display mode of the corrected and predicted value of protein content of rice according to the present embodiment. That is, the obtained corrected and predicted value of the protein content of the rice is put into the map data of the field 1. Further, by performing color discrimination and mapping on the distribution range of the protein content corrected predicted value, the distribution state can be visually expressed. By adding such a distribution map display step, it is possible to efficiently create a harvesting plan for classifying and harvesting rice that is good-tasting rice.

In the present embodiment, as shown in fig. 8, the fertilization performance information 5 can be displayed in addition to the distribution display of the protein content correction prediction value of rice. That is, in the distribution map display step, the fertilization performance results can be displayed by being combined with the plotted data of the field 1 in which the distribution range of the protein content correction prediction value is displayed in different colors. In the illustrated example, the amount of fertilizer sprayed and the number of days sprayed at that amount are shown. With this configuration, the relationship between the actual result of fertilization and the distribution of the protein content of the rice can be effectively grasped. By accumulating such information, the accumulated information can be effectively used in fertilizer application diagnosis at the next year and later growth stages, and a large amount of rice with good taste can be harvested.

(other embodiments)

One embodiment of the method for obtaining the protein content of the grain of the present invention is described above. However, the present invention is not necessarily limited to the above embodiment, and includes modifications as described below, for example.

For example, in the above embodiment, the processes of step 1 to step 3 are performed, but the process of step 1 is not necessarily required. That is, by the prediction processing of the protein content of the rice 2 to 3 weeks before the harvest due to step 2 (maturation period) and the correction prediction processing of the protein content of the rice just before the harvest (2 to 3 days before the harvest due date) of step 3, the protein content of the harvested rice can be predicted with high accuracy, and the rice with good taste can be efficiently classified and harvested.

In the above embodiment, as shown in fig. 1 and the like, an unmanned aerial vehicle 10 capable of autonomous flight is used, on which a camera 11 as an imaging unit is mounted. However, it is not necessarily limited to the unmanned aerial vehicle 10. For example, the camera 11 as the imaging means may be provided at the top of the column of the field 1 as viewed from above. Further, not only the unmanned aerial vehicle 10, but also an manned aircraft or a fixed wing aircraft can be used.

In the above embodiment, the unmanned aerial vehicle 10 capable of autonomous flight is used, but the present invention is not necessarily limited to such an unmanned aerial vehicle, and the operator may fly the unmanned aerial vehicle by performing remote operation each time.

In the above embodiment, the protein content of the rice is predicted at each stage, but a harvesting plan for classifying and harvesting rice with good taste may be made by converting the protein content into that of brown rice.

While the embodiments and some modifications of the present invention have been described above, the embodiments of the present invention are intended to facilitate understanding of the present invention and are not intended to limit the present invention. The present invention can be modified and improved within a range not departing from the gist thereof, and the present invention also includes equivalents thereof. In addition, in a range in which at least a part of the above problems can be solved or in a range in which at least a part of the effects can be achieved, a combination or omission of the protection range and each component described in the specification can be performed.

Description of the reference numerals

1 field

2 shooting data

3 grid

4 field information

5 fertilizing actual performance information

10 imaging device

11 Camera

31 division

A 32 reference field.

Claims

1. A method for obtaining the protein content of grains,

the acquisition method comprises:

a protein content prediction step for obtaining a predicted value of the protein content of grains in a maturation stage by remote sensing;

a protein content actual measurement step of acquiring an actual measurement value of the protein content of the grain from a sample of the grain to be harvested by ground truth; and

and a protein content correction step of correcting the predicted protein content value on the basis of the measured protein content value to obtain a corrected predicted protein content value of the grain.

2. The method for obtaining the protein content of grain according to claim 1,

the protein content prediction step includes the steps of:

a grid consisting of a plurality of sections is covered with image data of a field imaged by remote sensing, the predicted protein content value is acquired for each of the plurality of sections, and a section having the predicted protein content value with the smallest deviation from the peripheral section is determined as a reference field for collecting a sample of the grain.

3. The method for obtaining the protein content of grain according to claim 1 or 2,

the remote sensing device includes an unmanned aerial vehicle capable of autonomous flight, and an imaging unit is mounted on the unmanned aerial vehicle.

4. The method for obtaining the protein content of cereal grains according to any of claims 1 to 3,

the method for obtaining the protein content correction prediction value includes a step of setting a predetermined threshold value for the protein content correction prediction value.

5. The method for obtaining the protein content of cereal grains according to any of claims 1 to 4,

the acquisition method includes a distribution map display step of plotting and displaying a distribution range of the protein content corrected predicted values in the field.

6. The method for obtaining the protein content of cereal grains according to claim 5,

the method comprises a fertilization performance recording step of recording fertilization performance in the field,

in the distribution map display step, the fertilization performance may be displayed by synthesizing the mapping data of the field.