KR20240007390A - A continuous leaf temperature prediction method using smart farm environmental data. - Google Patents

Abstract

Many sensors are used to measure the growth status of plants in smart farms, etc. In particular, the temperature of plant leaves, that is, leaf temperature, is used as an indicator of the current state of the plant. By comparing the leaf temperature inside the smart farm and the air temperature inside the smart farm, it is possible to measure whether the transpiration of crops (plants) is active. Additionally, it is possible to diagnose high-temperature stress in plants using leaf temperature.
In order to measure leaf temperature, which reflects the growth state of crops (plants), an additional thermal imaging camera or infrared temperature sensor is required. An infrared sensor is installed at the tip of the leaf or only the image of the leaf is captured from the image taken with a thermal imaging camera. There is a problem in that it is not easy to measure leaf temperature because additional work is needed to differentiate. In order to solve the above problem, an internal temperature measurement step of measuring the temperature (intemp.) inside the smart farm; And an external temperature measurement step of measuring the temperature (outtemp.) outside the smart farm; And a solar radiation measurement step of measuring the solar radiation amount (rad.) of sunlight entering the smart farm; And canopy temperature (canopy temp.) is calculated as intemp. × A + outtemp. × B + rad. . (A, B, C, and D are constants depending on the degree to which they contribute to the increase in leaf temperature relative to the measured value being multiplied)
The above-mentioned configuration of the invention of this application has the effect of providing a means of predicting the leaf temperature of plants from the environmental variable values provided inside and outside the smart farm without directly measuring the leaf temperature of the plants grown inside the smart farm. .

Description

Leaf temperature prediction method using smart farm internal environmental data{.}

The present application relates to technology for predicting leaf temperature of plants grown in smart farms. More specifically, it concerns technology that accurately predicts plant leaf temperature without directly measuring it.

Prior art prior to the application of the present invention discloses a device for collecting plant growth environment data by measuring the temperature of plants. This technology is about an image processing system that can recognize the location of plants and a technology that can automatically measure the temperature value of the exact location using this.

Another prior technology is a technology that measures the temperature of leaves in a non-contact manner using color images and thermal images taken of plants. In this technology, the first step is to calculate mapping parameters between color images and thermal images using captured images of the same shot taken through a color camera and a thermal camera, and to target plants photographed through the color camera and thermal camera. A second step of acquiring a color image and a thermal image respectively, a third step of separating a leaf area with the corresponding color based on the color of the plant's leaf in the color image obtained in the second step, calculated in the first step A fourth step of setting an area corresponding to the leaf area in the color image in the thermal image based on the mapping parameters, and calculating the average temperature value for each pixel for the leaf area in the thermal image set in the fourth step. Thus, a technology including a fifth step of obtaining the leaf temperature of the plant is disclosed.

Public Patent Publication No. 10-2013-0035816 Public Patent Publication No. 10-2016-0145319

Many sensors are used to measure the growth status of plants in smart farms, etc. In particular, the temperature of plant leaves, that is, leaf temperature, is used as an indicator of the current state of the plant. By comparing the leaf temperature inside the smart farm and the air temperature inside the smart farm, it is possible to measure whether the transpiration of crops (plants) is active. Additionally, it is possible to diagnose high-temperature stress in plants using leaf temperature.

In order to measure leaf temperature, which reflects the growth state of crops (plants), an additional thermal imaging camera or infrared temperature sensor is required. An infrared sensor is installed at the tip of the leaf or only the image of the leaf is captured from the image taken with a thermal imaging camera. Additional work is needed to differentiate.

In order to solve the above problems, the present application seeks to develop a technology to predict leaf temperature inside a smart farm by obtaining a regression model using the smart farm's internal temperature, external temperature, and solar radiation.

In order to solve the above problems, the following problem solving means are provided.

An internal temperature measurement step of measuring the temperature (intemp.) inside the smart farm; and

An external temperature measurement step of measuring the temperature (outtemp.) outside the smart farm; and

A solar radiation measurement step of measuring the solar radiation (rad.) of sunlight entering the smart farm; and

We provide a canopy temperature prediction method using smart farm internal environment data, wherein canopy temperature is calculated as intemp.×A + outtemp.×B + rad.×C + D (Equation (1)).

(A, B, C, and D are constants depending on the degree to which they contribute to the increase in leaf temperature relative to the measured value being multiplied)

In addition, an internal humidity measurement step of measuring the humidity (inhum.) inside the smart farm; and

An external humidity measurement step of measuring the humidity (outhum.) outside the smart farm; and

A solar radiation measurement step of measuring the solar radiation (rad.) of sunlight entering the smart farm; and

Provides a transpiration prediction method using smart farm internal environment data, where the transpiration amount of crops is calculated as (inhum.- outhum.) do.

In addition, in predicting the leaf temperature, if (inhum.-outhum.) > 0, transpiration is active, so the leaf temperature is predicted using Equation (1), and (inhum.-outhum.) <= 0 In this case, transpiration is not active and the leaf temperature is not cooled by transpiration, so the leaf temperature (canopy temp.) is intemp.×A + outtemp.×B + rad.×C + D + t×E (equation (2) It provides a leaf temperature prediction method using smart farm internal environment data, which is calculated as )). (A, B, C, D, ㄸ are constants depending on the degree to which they contribute to the increase in leaf temperature for the measured value being multiplied, and t refers to the time during which (inhum.-outhum.) <= 0 (1 hour unit))

The above-mentioned configuration of the invention of this application has the effect of providing a means of predicting the leaf temperature of plants from the environmental variable values provided inside and outside the smart farm without directly measuring the leaf temperature of the plants grown inside the smart farm. .

Figure 1 is a data set used to predict leaf temperature of the present invention.
Figure 2 is a simulation result for setting the predictor variable of the present invention
Figure 3 is a simulation result for comparing the linear prediction method of the present invention
Figure 4 is a graph comparing actual leaf temperature measurements and predictions of the present invention (summer)
Figure 5 is a graph comparing actual leaf temperature measurements and predictions of the present invention (winter)

Water stress in crops refers to a phenomenon in which the crop loses more moisture through transpiration than it absorbs from the soil, resulting in a decrease in moisture content in the body, which results in a decrease in crop growth. When crops are subjected to moisture stress, stomata are reduced, transpiration is reduced, and leaf temperature (canopy temperature, Tc) increases.

Therefore, it is important to accurately measure the leaf temperature of the crop to check the current growth status of the crop. However, there is a problem in that measuring the leaf temperature of crops is not physically simple. If you measure once or twice a day, you can easily measure by manpower. However, it is not easy to monitor leaf temperature throughout the day. Because ventilation fans are running and workers are moving inside the smart farm, it is not easy for the leaf temperature sensor to accurately contact the leaves or to measure continuously at the exact location remotely using a thermal imaging camera.

The present application seeks to propose a method of predicting leaf temperature indirectly using indoor and outdoor environmental information, rather than the direct measurement method as described above.

Growth environment data related to leaf temperature are as shown in Figure 1.

In the present invention, the smart farm internal environmental variables are smart farm internal temperature (intemp.), smart farm internal humidity (inhum.), carbon dioxide concentration (co2), and solar radiation (rad), and the smart farm external environmental variables are smart farm external temperature. (outtemp.), smart farm external humidity (outhum.), and smart farm external wind speed (windspeed) were used.

First, the correlation between these environmental variables and leaf temperature was analyzed.

Three variable groups were created and the linear regression fit for each was analyzed to select the optimal leaf temperature predictor.

input_data used for the first group: internal temperature/humidity, CO2, solar radiation, external temperature/humidity, wind speed

input_data used for group 2: inside/outside temperature, CO2, solar radiation, wind speed

input_data used for the third group: inside/outside temperature, solar radiation

am.

As a result of the analysis, the result using the internal and external temperature and solar radiation of the third group's smart farm had the best value with a coefficient of determination of 0.702 and RMSE of 1.753. It turned out the best. The experimental results are shown in Figure 2.

In order to find a better leaf temperature prediction method based on the selected variables of the third group, the accuracy and/or interpretability of the linear model was added to the previously used linear regression method. To improve this, additional leaf temperature prediction was performed using the Ridge and Lasso methods with various regularization functions, and the results were analyzed.

Method 1: Linear Regression

Method 2: Ridge

Method 3: Lasso

As a result, method 3, the Lasso model, had the best value with a coefficient of determination of 0.728 and RMSE of 1.676. Figure 3 shows the analysis results.

At this time, the regression equation of the prediction model is as follows.

Leaf temperature = 9.651772156859273 + 0.46214678

can be predicted.

However, these predictions are limited to the demonstration farm to which the data used in the experiment was applied, and in order to use each of the formulas (Equation 0), it is necessary to modify the weights for changes in the smart farm's internal temperature, external temperature, and solar radiation. Additionally, modification of these weights is related to the structure and location of the smart farm.

The method of setting the weight in relation to the environmental condition variable is as follows.

<Example 1>

Inside the smart farm, there is always enough wind necessary for the transpiration of crops, and the smart farm structure is such that there is a difference of more than 5 degrees between the external and internal temperatures, and the relative humidity does not exceed 70%, so the transpiration function of the plants is maintained. In normal cases, the crop transpires and radiates leaf temperature normally even when solar radiation is strong, making it possible to predict the crop's leaf temperature using the equation (0) above.

Meanwhile, in order to additionally correct for differences in leaf temperature prediction due to differences in smart farm facilities, the internal temperature, external temperature, solar radiation, and leaf temperature can be measured more than once and the coefficients in equation (0) can be corrected or calibrated.

<Example 2>

In cases where the temperature inside the smart farm is higher than the outside temperature, such as in the middle of summer, and there is enough wind for transpiration of crops, but the relative humidity exceeds 90%, the transpiration of plants is not easy, in normal cases, transpiration of crops is possible. Because the operation is not smooth, it is impossible to dissipate the leaf temperature of the crop through transpiration, so the leaf temperature of the crop may be at a higher temperature compared to the temperature predicted by the above equation (0).

For these conditions

In order to check the humidity (inhum.) inside the smart farm, the humidity (outhum.) outside the smart farm, and the midday conditions, the amount of solar radiation (rad.) entering the smart farm is measured, and then the transpiration rate is calculated to determine the leaf temperature. The predicted temperature can be modified.

In other words, the transpiration amount of the crop can be calculated as (inhum.-outhum.)*V when (inhum.-outhum.) > 0 under the condition of solar radiation > 700 w/m2, where V is a weight value. The high humidity inside the smart farm is interpreted as the presence of water vapor generated by transpiration, which means that the crops are transpiration. Therefore, the leaf temperature prediction model of equation (0) can be used as is.

However, under conditions of solar irradiation > 700 w/m2 (inhum. - outhum.) <= 0, the external humidity is high and the crop cannot transpire. In this case, the crop cannot transpire, so an increase in leaf temperature is expected. do. Therefore, by adding -(inhum.-outhum.) However, since this increase in leaf temperature is a cumulative value over time, the leaf temperature must be predicted by accumulating the time when the amount of solar radiation is more than 700 w/m2. Therefore, the formula for predicting leaf temperature by multiplying the time when solar radiation is above 700 w/m2 is as follows.

Here, the unit of time (t) is hour, and the unit of minutes is expressed as a decimal point.

Leaf temperature = 9.651772156859273 + 0.46214678

Considering the variables that may differ depending on the smart farm from the above equation, canopy temperature (canopy temp.) = intemp.×A + outtemp.×B + rad.×C + D (equation (1)) A general equation can be obtained, and if the humidity inside the smart farm is high, (canopy temp.) = intemp.×A + outtemp.×B + rad.×C + D + t×E (equation (2)) considering humidity ) could be generalized.

Figure 4 shows the results of comparing the actual leaf temperature measured value and the leaf temperature predicted by the present invention. These are the results of an experiment from mid-May to early June, and the crop used in the experiment was tomatoes. It can be seen that the temperature of the actual measurement result is slightly higher than the predicted result.

Figure 5 is a graph comparing the actual measurement results and predicted results of leaf temperature from mid-December to the end of February. You can see that the actual leaf temperature is displayed slightly below.

However, the measurement results in the graphs of Figures 4 and 5 are experimental results that meet the predictive power sufficient for actual use compared to the effort and time to measure actual leaf temperature.

However, if we look at the cause of this difference by considering our country's climate, we have a climate environment where humidity is low in winter and humidity is high in summer.

Taking these climate conditions into consideration, in the winter when relative humidity is low, transpiration is smooth, so the leaf temperature can easily drop, and in the summer when the relative humidity is high, because transpiration is not smooth, the predicted temperature is lower than the actual temperature. There was. To solve this problem, we attempted to correct this by adding the internal humidity and external humidity of the smart farm to the formula. This is what equation (2) contains. In other words, if the humidity inside the smart farm is high during times of high solar radiation, symptoms are difficult, and the leaf temperature rises due to solar radiation, so the leaf temperature was corrected using the concept of cumulative time during times of high solar radiation. By making the above corrections, we were able to obtain more accurate leaf temperature prediction results.

The structure of the invention to demonstrate the effects of the invention as described above is as follows.

An internal temperature measurement step of measuring the temperature (intemp.) inside the smart farm; and

An external temperature measurement step of measuring the temperature (outtemp.) outside the smart farm; and

A solar radiation measurement step of measuring the solar radiation (rad.) of sunlight entering the smart farm; and

We provide a canopy temperature prediction method using smart farm internal environment data, wherein canopy temperature is calculated as intemp.×A + outtemp.×B + rad.×C + D (Equation (1)).

(A, B, C, and D are constants depending on the degree to which they contribute to the increase in leaf temperature relative to the measured value being multiplied)

In addition, an internal humidity measurement step of measuring the humidity (inhum.) inside the smart farm; and

An external humidity measurement step of measuring the humidity (outhum.) outside the smart farm; and

A solar radiation measurement step of measuring the solar radiation (rad.) of sunlight entering the smart farm; and

Provides a transpiration prediction method using smart farm internal environment data, where the transpiration amount of crops is calculated as (inhum.- outhum.) do.

In addition, in predicting the leaf temperature, if (inhum.-outhum.) > 0, transpiration is active, so the leaf temperature is predicted using Equation (1), and (inhum.-outhum.) <= 0 In this case, transpiration is not active and the leaf temperature is not cooled by transpiration, so the leaf temperature (canopy temp.) is intemp.×A + outtemp.×B + rad.×C + D + t×E (equation (2) It provides a leaf temperature prediction method using smart farm internal environment data, which is calculated as )). (A, B, C, D, ㄸ are constants depending on the degree to which they contribute to the increase in leaf temperature for the measured value being multiplied, and t refers to the time during which (inhum.-outhum.) <= 0 (1 hour unit))

Claims (3)

An internal temperature measurement step of measuring the temperature (intemp.) inside the smart farm; and
An external temperature measurement step of measuring the temperature (outtemp.) outside the smart farm; and
A solar radiation measurement step of measuring the solar radiation (rad.) of sunlight entering the smart farm; and
Canopy temperature prediction method using smart farm internal environment data, characterized in that canopy temperature is calculated as intemp.×A + outtemp.×B + rad.×C + D (Equation (1)).
(A, B, C, and D are constant values depending on the degree to which they contribute to the increase in leaf temperature relative to the measured value being multiplied) According to paragraph 1,
An internal humidity measurement step of measuring the humidity (inhum.) inside the smart farm; and
An external humidity measurement step of measuring the humidity (outhum.) outside the smart farm; and
A solar radiation measurement step of measuring the solar radiation (rad.) of sunlight entering the smart farm; And the transpiration amount of the crop is calculated as (inhum.-outhum.) × V when (inhum.-outhum.) > 0 under conditions of solar radiation of 700 w/m2 or more. Prediction method. According to paragraph 2,
In predicting the leaf temperature, if (inhum.-outhum.) > 0, transpiration is active and the leaf temperature is predicted using Equation (1), and if (inhum.-outhum.) <= 0 Since transpiration is not active and the leaf temperature is not cooled by transpiration, the canopy temperature is intemp.×A + outtemp.×B + rad.×C + D + t×E (Equation (2)) A leaf temperature prediction method using smart farm internal environment data, which is calculated as . (A, B, C, D, ㄸ are constants depending on the degree to which they contribute to the increase in leaf temperature for the measured value being multiplied, and t refers to the time during which (inhum.-outhum.) <= 0 (1 hour unit))