论文部分内容阅读
Abstract In this study， artificial leaf resistance was used to simulate leaf wetness. Specific to the solar greenhouse environment in Tianjin， microclimate monitoring equipment was installed for the collection of temperature group and humidity group data， as well as solar radiation and leaf wetness in the greenhouse. In order to reduce the complexity of multivariate factor prediction and ensure the richness of selected data types， correlation analysis was made to the 2 groups of data， screening 5 000 groups of data， including the humidity group data RH， RH20， RH40， temperature group data T， T20， T40， and solar radiation W. The data were then analyzed by principal component analysis， screening out 4 groups of principal components to show the leaf wetness index.
Key words Cucumber leaf wetness; Principal component analysis; Multiparameter factors
At present， the facility agriculture covers an area of 40 000 hm2 in Tianjin， which has become an important direction for the development of modern urban agriculture in Tianjin. High temperature and high wetness have certain influences on the growth and development of crops as well as the occurrences， physiological and biochemical processes of diseases under greenhouse conditions. As an important vegetative organ of plants， leaves are mainly to carry out photosynthesis and transpiration. When the leaf surface accumulates excessive water vapor for a long time， the leaf surface is also susceptible to infection by fungi and bacteria. Leaf wetness has become an important indicator for the evaluation of plant growth. In recent years， research has developed rapidly on technologies such as early warning and monitoring of facility greenhouse microclimate， but there are relatively few studies on the impact of facility environment microclimate on plant leaf wetness. Generally， the leaf surface wetness sensor， which uses artificial leaf resistance， can determine the existence of leaf wetness and its lasting time by placing near the leaves， and the leaf wetness can be reflected through the variation of simulated resistance changes. In this way， researchers or the cultivation managers can predict the occurrence of diseases， so as to take relevant measures to protect plants or crops.
Test Design
Basic situation of the test
The test was carried out in the solar greenhouse of the experimental base of Beichen District， Tianjin， from March 30， 2018 to June 18， 2018. The greenhouse was 70 m long from east to west and 12 m wide from north to south. The test crop was cucumber. During the test， the U30NRC microclimate environmental monitoring system produced by American HOBO Company was installed in the eastern， central and western regions of the greenhouse. The equipment performance indicators are shown in Table 1， and the installation position and measurement parameters were arranged as shown in Table 2. The wetness group data was collected at the time interval of 30 min， including air humidity RH， soil moisture at the depth of 10-40 cm （namely， RH10， RH20， RH30， RH40）， leaf wetness RHL， and the temperature group data included air temperature T， soil temperature at the depth of 10-40 cm （namely， T10， T20， T30， T40）， and solar radiation W， a total of 12 variables. The leaf wetness sensor was fixed horizontally on the stand， and the measurement was done from the fifth leaf from top to bottom. Temperature and wetness data correlation analysis
Data analysis was performed by using Excel 2010， SPSS20.0， Matlab2012. The correlation analysis was made between the humidity group data of air humidity， soil moisture at 10， 20， 30， 40 cm， and the temperature group data of air temperature， soil temperature at 10， 20， 30， 40 cm. The number of samples for both groups were 5 000.
Using Pearson correlation coefficient for two-tailed test， all data had statistical significance at significance Sig.<0.05， and all data were significantly correlated at the level of 0.01 （two tailed）. The correlation between air humidity index and soil moisture at 10-40 was weak， and the correlation coefficients were the greatest between the soil moisture at the layer of 20 cm and all other 3 layers， namely， 0.972， 0.931， and 0.886. The current soil layer was more correlated with the adjacent upper and middle layers after irrigation.
Using Pearson correlation coefficient for twotailed test， all data had statistical significance at significance Sig.<0.05， and all data were significantly correlated at the level of 0.01 （two tailed）. The correlation between the air temperature index and soil temperature at the layers of 10-40 cm was gradually reduced， and the correlation coefficients of soil temperate at the layer of 20 cm were the greatest with the other 3 layers， 0.899， 0.948， and 0.851， respectively.
Multiparameter Environmental Factors Principal Component Analysis and Results
It could effectively simplify the model dimension and improve the calculation speed of the model by picking up the group data with typical characteristics and high correlation coefficients with the other parameters in the group based on the correlation analysis on samples in the humidity group and temperature group which affected the intelligent irrigation control system of the solar greenhouse. Through the above analysis， the air humidity RH and the soil moisture RH20 at the soil layer of 20 cm were selected from the humidity group data， and the air temperature T and the soil temperature T20 at the soil depth of 20 cm were selected from the humidity group as the data with strong characteristics. The other data were simplified according to the independence or correlation with the selected data. The roots of the test cucumber grew to the soil depth of 15-30 cm， and deep water leakage occurred at the depth of 40 cm， which affected the irrigation water consumption and water resorption of plants. Therefore， the data related with soil moisture RH40 and soil temperature T40 at soil depth of 40 cm were kept. Princomp function was used for principal component analysis to preprocess the input samples， and multiple variables affecting irrigation water were converted into fewer independent variables that were not related to each other.
The source data samples and Xn≠m were standardized to eliminate the effects of dimensions and different orders of magnitudes among the variables， namely：
A sample matrix R was established as follows：
The eigenvalue ┡i and eigenvectors lj of the matrix R were calculated for the following calculation of the variance contribution rates and the cumulative variance contribution rates of the principal components.
The contribution rates of the principal components were arranged in descending order. When the contribution rate and cumulative contribution rate of the current principal components met the requirements of the reaction accuracy， the principal components was taken as the evaluation index.
In the meantime， the principal component scores and principal component loads were calculated by establishing the principal component equation.
From the result of result1， the contribution rate of the third principal component reached 88.647 2%， and the cumulative contribution rate of the first four principal components reached 95.034 9%.
The principal component load is the correlation coefficient between the principal component and the variable factor. As shown in Table 6， X5 and X6 were negatively correlated with principal component I， and then the principal component YI could be regarded as a comprehensive index composed of RH40 and T20. X1 was negatively correlated with principal component II， X3 was positively correlated with principal component II， and thus the main component YII could be regarded as a comprehensive index composed of RH and W. X4 was positively correlated with principal component III， and then the principal component YIII could be regarded as a comprehensive index composed of RH20. X1 and X3 were positively correlated with the principal component IV， and the principal component YIV could be regarded as a comprehensive index composed of RH and W， which reflected the comprehensive influences of the above factors on the leaf surface wetness RHL.
Conclusion
In this study， artificial leaf resistance was used to simulate leaf wetness， and the group data related with leaf wetness were collected. Then， correlation analysis was performed to screen the humidity group data and temperature group data， and the principal component analysis method was used to reduce the dimensions of the system， obtaining the comprehensive indicators related with leaf wetness. References
[1]WANG H， LI ML， XU JP， et al. An early warning method of cucumber downy mildew in solar greenhouse based on canopy temperature and humidity[J]. Chinese Journal of Applied Ecology， 2015（10）： 3027-3034.
[2]WANG HS. Controlling technologies for cucumber downy mildew through solar greenhouse temperature and humidity control[J]. China Agricultural Technology Extension， 2013 （8）： 45-46.
[3]WANG H， CHEN MX， LI WY， et al. Research on spatial distribution of daylight temperature and humidity of cucumber canopy in solar greenhouses[J]. Northern Horticulture，2015（17）：41 46.
[4]LIU X， BI HG， LI QM， et al. Effects of soil moisture on photosynthesis and antioxidant enzyme activities of cucumber seedlings under low temperature[J]. Chinese Journal of Plant Physiology， 2015（12）： 2247 2254.
[5]BAI QH， WANG WC. Modeling of the lowest temperature forecast in the sunlight greenhouse in Zhangye based on principal component regression[J]. Chinese Agricultural Science Bulletin， 2015， 31（32）： 223-228.
[6]LI N， SHEN SH， LI ZF， et al. Forecast model of minimum temperature inside greenhouse based on principal component regression[J]. Chinese Journal of Agrometeorology， 2013， 34（3）： 306-311.
Key words Cucumber leaf wetness; Principal component analysis; Multiparameter factors
At present， the facility agriculture covers an area of 40 000 hm2 in Tianjin， which has become an important direction for the development of modern urban agriculture in Tianjin. High temperature and high wetness have certain influences on the growth and development of crops as well as the occurrences， physiological and biochemical processes of diseases under greenhouse conditions. As an important vegetative organ of plants， leaves are mainly to carry out photosynthesis and transpiration. When the leaf surface accumulates excessive water vapor for a long time， the leaf surface is also susceptible to infection by fungi and bacteria. Leaf wetness has become an important indicator for the evaluation of plant growth. In recent years， research has developed rapidly on technologies such as early warning and monitoring of facility greenhouse microclimate， but there are relatively few studies on the impact of facility environment microclimate on plant leaf wetness. Generally， the leaf surface wetness sensor， which uses artificial leaf resistance， can determine the existence of leaf wetness and its lasting time by placing near the leaves， and the leaf wetness can be reflected through the variation of simulated resistance changes. In this way， researchers or the cultivation managers can predict the occurrence of diseases， so as to take relevant measures to protect plants or crops.
Test Design
Basic situation of the test
The test was carried out in the solar greenhouse of the experimental base of Beichen District， Tianjin， from March 30， 2018 to June 18， 2018. The greenhouse was 70 m long from east to west and 12 m wide from north to south. The test crop was cucumber. During the test， the U30NRC microclimate environmental monitoring system produced by American HOBO Company was installed in the eastern， central and western regions of the greenhouse. The equipment performance indicators are shown in Table 1， and the installation position and measurement parameters were arranged as shown in Table 2. The wetness group data was collected at the time interval of 30 min， including air humidity RH， soil moisture at the depth of 10-40 cm （namely， RH10， RH20， RH30， RH40）， leaf wetness RHL， and the temperature group data included air temperature T， soil temperature at the depth of 10-40 cm （namely， T10， T20， T30， T40）， and solar radiation W， a total of 12 variables. The leaf wetness sensor was fixed horizontally on the stand， and the measurement was done from the fifth leaf from top to bottom. Temperature and wetness data correlation analysis
Data analysis was performed by using Excel 2010， SPSS20.0， Matlab2012. The correlation analysis was made between the humidity group data of air humidity， soil moisture at 10， 20， 30， 40 cm， and the temperature group data of air temperature， soil temperature at 10， 20， 30， 40 cm. The number of samples for both groups were 5 000.
Using Pearson correlation coefficient for two-tailed test， all data had statistical significance at significance Sig.<0.05， and all data were significantly correlated at the level of 0.01 （two tailed）. The correlation between air humidity index and soil moisture at 10-40 was weak， and the correlation coefficients were the greatest between the soil moisture at the layer of 20 cm and all other 3 layers， namely， 0.972， 0.931， and 0.886. The current soil layer was more correlated with the adjacent upper and middle layers after irrigation.
Using Pearson correlation coefficient for twotailed test， all data had statistical significance at significance Sig.<0.05， and all data were significantly correlated at the level of 0.01 （two tailed）. The correlation between the air temperature index and soil temperature at the layers of 10-40 cm was gradually reduced， and the correlation coefficients of soil temperate at the layer of 20 cm were the greatest with the other 3 layers， 0.899， 0.948， and 0.851， respectively.
Multiparameter Environmental Factors Principal Component Analysis and Results
It could effectively simplify the model dimension and improve the calculation speed of the model by picking up the group data with typical characteristics and high correlation coefficients with the other parameters in the group based on the correlation analysis on samples in the humidity group and temperature group which affected the intelligent irrigation control system of the solar greenhouse. Through the above analysis， the air humidity RH and the soil moisture RH20 at the soil layer of 20 cm were selected from the humidity group data， and the air temperature T and the soil temperature T20 at the soil depth of 20 cm were selected from the humidity group as the data with strong characteristics. The other data were simplified according to the independence or correlation with the selected data. The roots of the test cucumber grew to the soil depth of 15-30 cm， and deep water leakage occurred at the depth of 40 cm， which affected the irrigation water consumption and water resorption of plants. Therefore， the data related with soil moisture RH40 and soil temperature T40 at soil depth of 40 cm were kept. Princomp function was used for principal component analysis to preprocess the input samples， and multiple variables affecting irrigation water were converted into fewer independent variables that were not related to each other.
The source data samples and Xn≠m were standardized to eliminate the effects of dimensions and different orders of magnitudes among the variables， namely：
A sample matrix R was established as follows：
The eigenvalue ┡i and eigenvectors lj of the matrix R were calculated for the following calculation of the variance contribution rates and the cumulative variance contribution rates of the principal components.
The contribution rates of the principal components were arranged in descending order. When the contribution rate and cumulative contribution rate of the current principal components met the requirements of the reaction accuracy， the principal components was taken as the evaluation index.
In the meantime， the principal component scores and principal component loads were calculated by establishing the principal component equation.
From the result of result1， the contribution rate of the third principal component reached 88.647 2%， and the cumulative contribution rate of the first four principal components reached 95.034 9%.
The principal component load is the correlation coefficient between the principal component and the variable factor. As shown in Table 6， X5 and X6 were negatively correlated with principal component I， and then the principal component YI could be regarded as a comprehensive index composed of RH40 and T20. X1 was negatively correlated with principal component II， X3 was positively correlated with principal component II， and thus the main component YII could be regarded as a comprehensive index composed of RH and W. X4 was positively correlated with principal component III， and then the principal component YIII could be regarded as a comprehensive index composed of RH20. X1 and X3 were positively correlated with the principal component IV， and the principal component YIV could be regarded as a comprehensive index composed of RH and W， which reflected the comprehensive influences of the above factors on the leaf surface wetness RHL.
Conclusion
In this study， artificial leaf resistance was used to simulate leaf wetness， and the group data related with leaf wetness were collected. Then， correlation analysis was performed to screen the humidity group data and temperature group data， and the principal component analysis method was used to reduce the dimensions of the system， obtaining the comprehensive indicators related with leaf wetness. References
[1]WANG H， LI ML， XU JP， et al. An early warning method of cucumber downy mildew in solar greenhouse based on canopy temperature and humidity[J]. Chinese Journal of Applied Ecology， 2015（10）： 3027-3034.
[2]WANG HS. Controlling technologies for cucumber downy mildew through solar greenhouse temperature and humidity control[J]. China Agricultural Technology Extension， 2013 （8）： 45-46.
[3]WANG H， CHEN MX， LI WY， et al. Research on spatial distribution of daylight temperature and humidity of cucumber canopy in solar greenhouses[J]. Northern Horticulture，2015（17）：41 46.
[4]LIU X， BI HG， LI QM， et al. Effects of soil moisture on photosynthesis and antioxidant enzyme activities of cucumber seedlings under low temperature[J]. Chinese Journal of Plant Physiology， 2015（12）： 2247 2254.
[5]BAI QH， WANG WC. Modeling of the lowest temperature forecast in the sunlight greenhouse in Zhangye based on principal component regression[J]. Chinese Agricultural Science Bulletin， 2015， 31（32）： 223-228.
[6]LI N， SHEN SH， LI ZF， et al. Forecast model of minimum temperature inside greenhouse based on principal component regression[J]. Chinese Journal of Agrometeorology， 2013， 34（3）： 306-311.